diff --git a/source/devices/AM62PX/linux/Linux_Performance_Guide.rst b/source/devices/AM62PX/linux/Linux_Performance_Guide.rst index 3f066de8c..bd75f0976 100644 --- a/source/devices/AM62PX/linux/Linux_Performance_Guide.rst +++ b/source/devices/AM62PX/linux/Linux_Performance_Guide.rst @@ -1,7 +1,7 @@ -====================================== - Linux 11.00 Performance Guide -====================================== +============================== + Linux 11.01 Performance Guide +============================== .. rubric:: **Read This First** :name: read-this-first-kernel-perf-guide @@ -33,15 +33,15 @@ For further information or to report any problems, contact https://e2e.ti.com/ or https://support.ti.com/ System Benchmarks -------------------- +----------------- LMBench -^^^^^^^^^^^^^^^^^^^^^^^^^^^ +^^^^^^^ LMBench is a collection of microbenchmarks of which the memory bandwidth and latency related ones are typically used to estimate processor memory system performance. More information about lmbench at -https://lmbench.sourceforge.net/whatis_lmbench.html and -https://lmbench.sourceforge.net/man/lmbench.8.html +http://lmbench.sourceforge.net/whatis_lmbench.html and +http://lmbench.sourceforge.net/man/lmbench.8.html **Latency**: lat_mem_rd-stride128-szN, where N is equal to or smaller than the cache size at given level measures the cache miss penalty. N that is at least @@ -63,75 +63,75 @@ Execute the LMBench with the following: .. csv-table:: LMBench Benchmarks :header: "Benchmarks","am62pxx_sk-fs: perf" - "af_unix_sock_stream_latency (microsec)","29.90" - "af_unix_socket_stream_bandwidth (MBs)","1187.22" - "bw_file_rd-io-1mb (MB/s)","1467.89" - "bw_file_rd-o2c-1mb (MB/s)","724.77" - "bw_mem-bcopy-16mb (MB/s)","1973.60" - "bw_mem-bcopy-1mb (MB/s)","2179.08" - "bw_mem-bcopy-2mb (MB/s)","1655.63" - "bw_mem-bcopy-4mb (MB/s)","1828.71" - "bw_mem-bcopy-8mb (MB/s)","1966.09" - "bw_mem-bzero-16mb (MB/s)","8108.11" - "bw_mem-bzero-1mb (MB/s)","5143.60 (min 2179.08, max 8108.11)" - "bw_mem-bzero-2mb (MB/s)","4875.69 (min 1655.63, max 8095.74)" - "bw_mem-bzero-4mb (MB/s)","4965.67 (min 1828.71, max 8102.63)" - "bw_mem-bzero-8mb (MB/s)","5036.42 (min 1966.09, max 8106.74)" - "bw_mem-cp-16mb (MB/s)","989.73" - "bw_mem-cp-1mb (MB/s)","4676.66 (min 960.92, max 8392.39)" - "bw_mem-cp-2mb (MB/s)","4599.93 (min 954.65, max 8245.21)" - "bw_mem-cp-4mb (MB/s)","4589.01 (min 1006.42, max 8171.60)" - "bw_mem-cp-8mb (MB/s)","4589.65 (min 1032.66, max 8146.64)" - "bw_mem-fcp-16mb (MB/s)","1835.92" - "bw_mem-fcp-1mb (MB/s)","4934.80 (min 1761.49, max 8108.11)" - "bw_mem-fcp-2mb (MB/s)","4899.95 (min 1704.16, max 8095.74)" - "bw_mem-fcp-4mb (MB/s)","4942.98 (min 1783.33, max 8102.63)" - "bw_mem-fcp-8mb (MB/s)","4963.91 (min 1821.08, max 8106.74)" - "bw_mem-frd-16mb (MB/s)","2015.11" - "bw_mem-frd-1mb (MB/s)","1890.85 (min 1761.49, max 2020.20)" - "bw_mem-frd-2mb (MB/s)","1766.66 (min 1704.16, max 1829.16)" - "bw_mem-frd-4mb (MB/s)","1880.62 (min 1783.33, max 1977.91)" - "bw_mem-frd-8mb (MB/s)","1910.54 (min 1821.08, max 2000.00)" - "bw_mem-fwr-16mb (MB/s)","8123.20" - "bw_mem-fwr-1mb (MB/s)","5206.30 (min 2020.20, max 8392.39)" - "bw_mem-fwr-2mb (MB/s)","5037.19 (min 1829.16, max 8245.21)" - "bw_mem-fwr-4mb (MB/s)","5074.76 (min 1977.91, max 8171.60)" - "bw_mem-fwr-8mb (MB/s)","5073.32 (min 2000.00, max 8146.64)" - "bw_mem-rd-16mb (MB/s)","2074.69" - "bw_mem-rd-1mb (MB/s)","2086.73 (min 1837.01, max 2336.45)" - "bw_mem-rd-2mb (MB/s)","1844.74 (min 1649.35, max 2040.12)" - "bw_mem-rd-4mb (MB/s)","1931.26 (min 1805.60, max 2056.91)" - "bw_mem-rd-8mb (MB/s)","1950.34 (min 1834.02, max 2066.65)" - "bw_mem-rdwr-16mb (MB/s)","1982.65" - "bw_mem-rdwr-1mb (MB/s)","1365.42 (min 960.92, max 1769.91)" - "bw_mem-rdwr-2mb (MB/s)","1249.89 (min 954.65, max 1545.12)" - "bw_mem-rdwr-4mb (MB/s)","1447.05 (min 1006.42, max 1887.68)" - "bw_mem-rdwr-8mb (MB/s)","1478.10 (min 1032.66, max 1923.54)" - "bw_mem-wr-16mb (MB/s)","1977.51" - "bw_mem-wr-1mb (MB/s)","1803.46 (min 1769.91, max 1837.01)" - "bw_mem-wr-2mb (MB/s)","1597.24 (min 1545.12, max 1649.35)" - "bw_mem-wr-4mb (MB/s)","1846.64 (min 1805.60, max 1887.68)" - "bw_mem-wr-8mb (MB/s)","1878.78 (min 1834.02, max 1923.54)" - "bw_mmap_rd-mo-1mb (MB/s)","2253.42" - "bw_mmap_rd-o2c-1mb (MB/s)","760.31" - "bw_pipe (MB/s)","793.54" - "bw_unix (MB/s)","1187.22" - "lat_connect (us)","56.55" - "lat_ctx-2-128k (us)","7.66" - "lat_ctx-2-256k (us)","6.67" + "af_unix_sock_stream_latency (microsec)","30.16" + "af_unix_socket_stream_bandwidth (MBs)","1195.29" + "bw_file_rd-io-1mb (MB/s)","1441.44" + "bw_file_rd-o2c-1mb (MB/s)","802.95" + "bw_mem-bcopy-16mb (MB/s)","1962.71" + "bw_mem-bcopy-1mb (MB/s)","2096.88" + "bw_mem-bcopy-2mb (MB/s)","1846.15" + "bw_mem-bcopy-4mb (MB/s)","1671.54" + "bw_mem-bcopy-8mb (MB/s)","1938.45" + "bw_mem-bzero-16mb (MB/s)","8114.96" + "bw_mem-bzero-1mb (MB/s)","5095.88 (min 2096.88, max 8094.88)" + "bw_mem-bzero-2mb (MB/s)","4970.95 (min 1846.15, max 8095.74)" + "bw_mem-bzero-4mb (MB/s)","4891.20 (min 1671.54, max 8110.85)" + "bw_mem-bzero-8mb (MB/s)","5023.97 (min 1938.45, max 8109.48)" + "bw_mem-cp-16mb (MB/s)","981.17" + "bw_mem-cp-1mb (MB/s)","4660.98 (min 968.52, max 8353.44)" + "bw_mem-cp-2mb (MB/s)","4599.50 (min 955.26, max 8243.73)" + "bw_mem-cp-4mb (MB/s)","4596.81 (min 1006.67, max 8186.94)" + "bw_mem-cp-8mb (MB/s)","4586.70 (min 1025.38, max 8148.02)" + "bw_mem-fcp-16mb (MB/s)","1831.08" + "bw_mem-fcp-1mb (MB/s)","4939.98 (min 1785.08, max 8094.88)" + "bw_mem-fcp-2mb (MB/s)","4916.98 (min 1738.22, max 8095.74)" + "bw_mem-fcp-4mb (MB/s)","4926.51 (min 1742.16, max 8110.85)" + "bw_mem-fcp-8mb (MB/s)","4962.49 (min 1815.50, max 8109.48)" + "bw_mem-frd-16mb (MB/s)","1847.79" + "bw_mem-frd-1mb (MB/s)","1909.49 (min 1785.08, max 2033.90)" + "bw_mem-frd-2mb (MB/s)","1869.31 (min 1738.22, max 2000.40)" + "bw_mem-frd-4mb (MB/s)","1831.85 (min 1742.16, max 1921.54)" + "bw_mem-frd-8mb (MB/s)","1872.19 (min 1815.50, max 1928.87)" + "bw_mem-fwr-16mb (MB/s)","8128.70" + "bw_mem-fwr-1mb (MB/s)","5193.67 (min 2033.90, max 8353.44)" + "bw_mem-fwr-2mb (MB/s)","5122.07 (min 2000.40, max 8243.73)" + "bw_mem-fwr-4mb (MB/s)","5054.24 (min 1921.54, max 8186.94)" + "bw_mem-fwr-8mb (MB/s)","5038.45 (min 1928.87, max 8148.02)" + "bw_mem-rd-16mb (MB/s)","2065.05" + "bw_mem-rd-1mb (MB/s)","2124.95 (min 1916.38, max 2333.51)" + "bw_mem-rd-2mb (MB/s)","1874.93 (min 1704.16, max 2045.69)" + "bw_mem-rd-4mb (MB/s)","1944.50 (min 1829.27, max 2059.73)" + "bw_mem-rd-8mb (MB/s)","2000.01 (min 1930.97, max 2069.05)" + "bw_mem-rdwr-16mb (MB/s)","1983.88" + "bw_mem-rdwr-1mb (MB/s)","1404.23 (min 968.52, max 1839.93)" + "bw_mem-rdwr-2mb (MB/s)","1272.39 (min 955.26, max 1589.51)" + "bw_mem-rdwr-4mb (MB/s)","1417.41 (min 1006.67, max 1828.15)" + "bw_mem-rdwr-8mb (MB/s)","1493.08 (min 1025.38, max 1960.78)" + "bw_mem-wr-16mb (MB/s)","1972.87" + "bw_mem-wr-1mb (MB/s)","1878.16 (min 1839.93, max 1916.38)" + "bw_mem-wr-2mb (MB/s)","1646.84 (min 1589.51, max 1704.16)" + "bw_mem-wr-4mb (MB/s)","1828.71 (min 1828.15, max 1829.27)" + "bw_mem-wr-8mb (MB/s)","1945.88 (min 1930.97, max 1960.78)" + "bw_mmap_rd-mo-1mb (MB/s)","2246.03" + "bw_mmap_rd-o2c-1mb (MB/s)","796.31" + "bw_pipe (MB/s)","780.02" + "bw_unix (MB/s)","1195.29" + "lat_connect (us)","56.93" + "lat_ctx-2-128k (us)","7.60" + "lat_ctx-2-256k (us)","7.56" "lat_ctx-4-128k (us)","7.18" - "lat_ctx-4-256k (us)","6.67" - "lat_fs-0k (num_files)","220.00" - "lat_fs-10k (num_files)","115.00" - "lat_fs-1k (num_files)","168.00" - "lat_fs-4k (num_files)","161.00" - "lat_mem_rd-stride128-sz1000k (ns)","29.37" - "lat_mem_rd-stride128-sz125k (ns)","5.55" - "lat_mem_rd-stride128-sz250k (ns)","5.84" + "lat_ctx-4-256k (us)","6.51" + "lat_fs-0k (num_files)","248.00" + "lat_fs-10k (num_files)","108.00" + "lat_fs-1k (num_files)","178.00" + "lat_fs-4k (num_files)","175.00" + "lat_mem_rd-stride128-sz1000k (ns)","29.18" + "lat_mem_rd-stride128-sz125k (ns)","5.62" + "lat_mem_rd-stride128-sz250k (ns)","5.83" "lat_mem_rd-stride128-sz31k (ns)","3.65" "lat_mem_rd-stride128-sz50 (ns)","2.15" - "lat_mem_rd-stride128-sz500k (ns)","11.25" - "lat_mem_rd-stride128-sz62k (ns)","5.24" + "lat_mem_rd-stride128-sz500k (ns)","12.89" + "lat_mem_rd-stride128-sz62k (ns)","5.22" "lat_mmap-1m (us)","56.00" "lat_ops-double-add (ns)","2.86" "lat_ops-double-div (ns)","15.74" @@ -143,44 +143,46 @@ Execute the LMBench with the following: "lat_ops-int-bit (ns)","0.48" "lat_ops-int-div (ns)","4.29" "lat_ops-int-mod (ns)","4.53" - "lat_ops-int-mul (ns)","3.07" + "lat_ops-int-mul (ns)","3.08" "lat_ops-int64-add (ns)","0.72" "lat_ops-int64-bit (ns)","0.48" - "lat_ops-int64-div (ns)","6.80" + "lat_ops-int64-div (ns)","6.79" "lat_ops-int64-mod (ns)","5.25" "lat_ops-int64-mul (ns)","3.55" - "lat_pagefault (us)","0.51" - "lat_pipe (us)","25.28" - "lat_proc-exec (us)","687.71" - "lat_proc-fork (us)","577.30" + "lat_pagefault (us)","0.50" + "lat_pipe (us)","24.46" + "lat_proc-exec (us)","682.50" + "lat_proc-fork (us)","590.20" "lat_proc-proccall (us)","0.01" - "lat_select (us)","33.98" - "lat_sem (us)","2.72" - "lat_sig-catch (us)","5.57" - "lat_sig-install (us)","0.68" - "lat_sig-prot (us)","0.86" - "lat_syscall-fstat (us)","1.96" + "lat_select (us)","34.04" + "lat_sem (us)","3.53" + "lat_sig-catch (us)","5.30" + "lat_sig-install (us)","0.65" + "lat_sig-prot (us)","0.64" + "lat_syscall-fstat (us)","1.99" "lat_syscall-null (us)","0.46" - "lat_syscall-open (us)","151.08" - "lat_syscall-read (us)","0.79" - "lat_syscall-stat (us)","4.76" - "lat_syscall-write (us)","0.76" - "lat_tcp (us)","0.91" - "lat_unix (us)","29.90" - "latency_for_0.50_mb_block_size (nanosec)","11.25" - "latency_for_1.00_mb_block_size (nanosec)","14.68 (min 0.00, max 29.37)" - "pipe_bandwidth (MBs)","793.54" - "pipe_latency (microsec)","25.28" + "lat_syscall-open (us)","147.43" + "lat_syscall-read (us)","0.80" + "lat_syscall-stat (us)","4.85" + "lat_syscall-write (us)","0.82" + "lat_tcp (us)","0.93" + "lat_unix (us)","30.16" + "latency_for_0.50_mb_block_size (nanosec)","12.89" + "latency_for_1.00_mb_block_size (nanosec)","14.59 (min 0.00, max 29.18)" + "pipe_bandwidth (MBs)","780.02" + "pipe_latency (microsec)","24.46" "procedure_call (microsec)","0.01" - "select_on_200_tcp_fds (microsec)","33.98" - "semaphore_latency (microsec)","2.72" - "signal_handler_latency (microsec)","0.68" - "signal_handler_overhead (microsec)","5.57" - "tcp_ip_connection_cost_to_localhost (microsec)","56.55" - "tcp_latency_using_localhost (microsec)","0.91" + "select_on_200_tcp_fds (microsec)","34.04" + "semaphore_latency (microsec)","3.53" + "signal_handler_latency (microsec)","0.65" + "signal_handler_overhead (microsec)","5.30" + "tcp_ip_connection_cost_to_localhost (microsec)","56.93" + "tcp_latency_using_localhost (microsec)","0.93" + + Dhrystone -^^^^^^^^^^^^^^^^^^^^^^^^^^^ +^^^^^^^^^ Dhrystone is a core only benchmark that runs from warm L1 caches in all modern processors. It scales linearly with clock speed. @@ -188,6 +190,7 @@ Please take note, different run may produce different slightly results. This is advised to run this test multiple times in order to get maximum performance numbers. + Execute the benchmark with the following: :: @@ -201,8 +204,10 @@ Execute the benchmark with the following: "dhrystone_per_mhz (DMIPS/MHz)","2.90" "dhrystone_per_second (DhrystoneP)","7142857.00" + + Whetstone -^^^^^^^^^^^^^^^^^^^^^^^^^^^ +^^^^^^^^^ Whetstone is a benchmark primarily measuring floating-point arithmetic performance. Execute the benchmark with the following: @@ -216,18 +221,22 @@ Execute the benchmark with the following: "whetstone (MIPS)","10000.00" + + Linpack -^^^^^^^^^^^^^^^^^^^^^^^^^^^ +^^^^^^^ Linpack measures peak double precision (64 bit) floating point performance in solving a dense linear system. .. csv-table:: Linpack Benchmarks :header: "Benchmarks","am62pxx_sk-fs: perf" - "linpack (Kflops)","577371.00" + "linpack (Kflops)","329621.00" + + NBench -^^^^^^^^^^^^^^^^^^^^^^^^^^^ +^^^^^^ NBench which stands for Native Benchmark is used to measure macro benchmarks for commonly used operations such as sorting and analysis algorithms. More information about NBench at @@ -237,18 +246,20 @@ https://nbench.io/articles/index.html .. csv-table:: NBench Benchmarks :header: "Benchmarks","am62pxx_sk-fs: perf" - "assignment (Iterations)","14.49" - "fourier (Iterations)","22828.00" - "fp_emulation (Iterations)","215.61" - "huffman (Iterations)","1184.20" - "idea (Iterations)","3444.20" - "lu_decomposition (Iterations)","526.96" - "neural_net (Iterations)","8.66" - "numeric_sort (Iterations)","628.16" - "string_sort (Iterations)","163.90" + "assignment (Iterations)","8.26" + "fourier (Iterations)","13044.00" + "fp_emulation (Iterations)","123.21" + "huffman (Iterations)","676.56" + "idea (Iterations)","1968.20" + "lu_decomposition (Iterations)","302.85" + "neural_net (Iterations)","4.95" + "numeric_sort (Iterations)","358.59" + "string_sort (Iterations)","93.66" + + Stream -^^^^^^^^^^^^^^^^^^^^^^^^^^^ +^^^^^^ STREAM is a microbenchmark for measuring data memory system performance without any data reuse. It is designed to miss on caches and exercise data prefetcher and speculative accesses. @@ -268,49 +279,58 @@ Execute the benchmark with the following: .. csv-table:: Stream Benchmarks :header: "Benchmarks","am62pxx_sk-fs: perf" - "add (MB/s)","2887.70" - "copy (MB/s)","4034.70" - "scale (MB/s)","3712.80" - "triad (MB/s)","2539.20" + "add (MB/s)","2719.00" + "copy (MB/s)","3791.20" + "scale (MB/s)","3426.70" + "triad (MB/s)","2418.50" + + CoreMarkPro -^^^^^^^^^^^^^^^^^^^^^^^^^^^ +^^^^^^^^^^^ CoreMark®-Pro is a comprehensive, advanced processor benchmark that works with and enhances the market-proven industry-standard EEMBC CoreMark® benchmark. While CoreMark stresses the CPU pipeline, CoreMark-Pro tests the entire processor, adding comprehensive support for multicore technology, a combination of integer and floating-point workloads, and data sets for utilizing larger memory subsystems. + .. csv-table:: CoreMarkPro Benchmarks :header: "Benchmarks","am62pxx_sk-fs: perf" - "cjpeg-rose7-preset (workloads/)","23.98" - "core (workloads/)","0.17" - "coremark-pro ()","544.04" - "linear_alg-mid-100x100-sp (workloads/)","8.37" - "loops-all-mid-10k-sp (workloads/)","0.42" - "nnet_test (workloads/)","0.62" - "parser-125k (workloads/)","5.13" - "radix2-big-64k (workloads/)","43.30" - "sha-test (workloads/)","46.73" - "zip-test (workloads/)","12.82" + "cjpeg-rose7-preset (workloads/)","41.84" + "core (workloads/)","0.30" + "coremark-pro ()","918.00" + "linear_alg-mid-100x100-sp (workloads/)","14.67" + "loops-all-mid-10k-sp (workloads/)","0.71" + "nnet_test (workloads/)","1.09" + "parser-125k (workloads/)","8.77" + "radix2-big-64k (workloads/)","59.36" + "sha-test (workloads/)","81.30" + "zip-test (workloads/)","21.74" + + + .. csv-table:: CoreMarkProTwoCore Benchmarks :header: "Benchmarks","am62pxx_sk-fs: perf" "cjpeg-rose7-preset (workloads/)","84.03" "core (workloads/)","0.60" - "coremark-pro ()","1667.13" - "linear_alg-mid-100x100-sp (workloads/)","29.34" + "coremark-pro ()","1676.82" + "linear_alg-mid-100x100-sp (workloads/)","29.36" "loops-all-mid-10k-sp (workloads/)","1.30" "nnet_test (workloads/)","2.17" - "parser-125k (workloads/)","13.70" - "radix2-big-64k (workloads/)","71.80" + "parser-125k (workloads/)","14.08" + "radix2-big-64k (workloads/)","71.98" "sha-test (workloads/)","161.29" - "zip-test (workloads/)","42.55" + "zip-test (workloads/)","43.48" + + + MultiBench -^^^^^^^^^^^^^^^^^^^^^^^^^^^ +^^^^^^^^^^ MultiBench™ is a suite of benchmarks that allows processor and system designers to analyze, test, and improve multicore processors. It uses three forms of concurrency: Data decomposition: multiple threads cooperating on achieving a unified goal and @@ -329,91 +349,118 @@ thread-enabled workloads to be tested. .. csv-table:: Multibench Benchmarks :header: "Benchmarks","am62pxx_sk-fs: perf" - "4m-check (workloads/)","431.11" - "4m-check-reassembly (workloads/)","134.59" + "4m-check (workloads/)","428.45" + "4m-check-reassembly (workloads/)","126.90" "4m-check-reassembly-tcp (workloads/)","61.12" - "4m-check-reassembly-tcp-cmykw2-rotatew2 (workloads/)","34.56" - "4m-check-reassembly-tcp-x264w2 (workloads/)","1.90" - "4m-cmykw2 (workloads/)","244.80" + "4m-check-reassembly-tcp-cmykw2-rotatew2 (workloads/)","34.48" + "4m-check-reassembly-tcp-x264w2 (workloads/)","1.91" + "4m-cmykw2 (workloads/)","243.90" "4m-cmykw2-rotatew2 (workloads/)","51.64" - "4m-reassembly (workloads/)","86.81" - "4m-rotatew2 (workloads/)","53.91" + "4m-reassembly (workloads/)","87.11" + "4m-rotatew2 (workloads/)","54.23" "4m-tcp-mixed (workloads/)","120.30" - "4m-x264w2 (workloads/)","2.00" + "4m-x264w2 (workloads/)","1.94" "empty-wld (workloads/)","1.00" - "idct-4m (workloads/)","19.26" - "idct-4mw1 (workloads/)","19.28" - "ippktcheck-4m (workloads/)","432.45" - "ippktcheck-4mw1 (workloads/)","432.30" - "ipres-4m (workloads/)","117.93" - "ipres-4mw1 (workloads/)","116.91" - "md5-4m (workloads/)","28.52" - "md5-4mw1 (workloads/)","28.42" + "idct-4m (workloads/)","19.28" + "idct-4mw1 (workloads/)","19.29" + "ippktcheck-4m (workloads/)","431.93" + "ippktcheck-4mw1 (workloads/)","430.66" + "ipres-4m (workloads/)","118.02" + "ipres-4mw1 (workloads/)","117.46" + "md5-4m (workloads/)","27.98" + "md5-4mw1 (workloads/)","28.02" "rgbcmyk-4m (workloads/)","62.75" - "rgbcmyk-4mw1 (workloads/)","62.89" - "rotate-4ms1 (workloads/)","24.06" - "rotate-4ms1w1 (workloads/)","24.10" - "rotate-4ms64 (workloads/)","24.34" - "rotate-4ms64w1 (workloads/)","24.38" + "rgbcmyk-4mw1 (workloads/)","62.72" + "rotate-4ms1 (workloads/)","24.08" + "rotate-4ms1w1 (workloads/)","24.15" + "rotate-4ms64 (workloads/)","24.35" + "rotate-4ms64w1 (workloads/)","24.32" "x264-4mq (workloads/)","0.58" "x264-4mqw1 (workloads/)","0.58" + + + Boot-time Measurement --------------------- + Boot media: MMCSD ^^^^^^^^^^^^^^^^^ .. csv-table:: Linux boot time MMCSD :header: "Boot Configuration","am62pxx_sk-fs: Boot time in seconds: avg(min,max)" - "Linux boot time from SD with default rootfs (20 boot cycles)","41.19 (min 13.87, max 122.38)" + "Linux boot time from SD with default rootfs (20 boot cycles)","13.63 (min 13.22, max 14.01)" + + Boot time numbers [avg, min, max] are measured from "Starting kernel" to Linux prompt across 20 boot cycles. + + | ALSA SoC Audio Driver -------------------------- +--------------------- #. Access type - RW\_INTERLEAVED #. Channels - 2 #. Format - S16\_LE #. Period size - 64 + .. csv-table:: Audio Capture :header: "Sampling Rate (Hz)","am62pxx_sk-fs: Throughput (bits/sec)","am62pxx_sk-fs: CPU Load (%)" - "8000","255998.00","0.22" - "11025","352798.00","0.17" - "16000","511997.00","0.11" - "22050","705596.00","0.28" - "24000","705594.00","0.28" - "32000","1023987.00","0.18" - "44100","1411184.00","0.46" - "48000","1535984.00","0.24" - "88200","2822369.00","0.86" - "96000","3071967.00","0.40" + "11025","352798.00","0.10" + "16000","511998.00","0.09" + "22050","705597.00","0.13" + "24000","705597.00","0.13" + "32000","1023995.00","0.10" + "44100","1411193.00","0.21" + "48000","1535992.00","0.16" + "88200","2822382.00","0.35" + "96000","3071979.00","0.18" + + + + +.. csv-table:: Audio Playback + :header: "Sampling Rate (Hz)","am62pxx_sk-fs: Throughput (bits/sec)","am62pxx_sk-fs: CPU Load (%)" + + "11025","352944.00","0.14" + "16000","512210.00","0.26" + "22050","705885.00","0.17" + "24000","705887.00","0.19" + "32000","1024418.00","0.16" + "44100","1411775.00","0.28" + "48000","1536626.00","0.23" + "88200","2823541.00","0.46" + "96000","3073248.00","0.31" + | Graphics SGX/RGX Driver -------------------------- +----------------------- + GFXBench -^^^^^^^^^^^^^^^^^^^^^^^^^^^ +^^^^^^^^ Run GFXBench and capture performance reported (Score and Display rate in fps). All display outputs (HDMI, Displayport and/or LCD) are connected when running these tests .. csv-table:: GFXBench Performance :header: "Benchmark","am62pxx_sk-fs: Score","am62pxx_sk-fs: Fps" - " GFXBench 3.x gl_manhattan_off","912.50","14.72" - " GFXBench 3.x gl_trex_off","1596.34","28.51" - " GFXBench 4.x gl_4_off","260.30","4.40" - " GFXBench 5.x gl_5_high_off","114.48","1.78" + " GFXBench 3.x gl_manhattan_off","907.99","14.65" + " GFXBench 3.x gl_trex_off","1595.80","28.50" + " GFXBench 4.x gl_4_off","260.27","4.40" + " GFXBench 5.x gl_5_high_off","113.55","1.77" + Glmark2 -^^^^^^^^^^^^^^^^^^^^^^^^^^^ +^^^^^^^ Run Glmark2 and capture performance reported (Score). All display outputs (HDMI, Displayport and/or LCD) are connected when running these tests @@ -421,12 +468,13 @@ Run Glmark2 and capture performance reported (Score). All display outputs (HDMI, :header: "Benchmark","am62pxx_sk-fs: Score" "Glmark2-DRM","355.00" - "Glmark2-Wayland","737.00" + "Glmark2-Wayland","734.00" + | Ethernet ------------------ +-------- Ethernet performance benchmarks were measured using Netperf 2.7.1 https://hewlettpackard.github.io/netperf/doc/netperf.html Test procedures were modeled after those defined in RFC-2544: https://tools.ietf.org/html/rfc2544, where the DUT is the TI device @@ -492,10 +540,12 @@ Running the following commands will trigger netperf clients to measure UDP burst netperf -H -j -C -l 60 -t UDP_STREAM -b -w -- -m -k DIRECTION,THROUGHPUT,MEAN_LATENCY,LOCAL_CPU_UTIL,REMOTE_CPU_UTIL,LOCAL_BYTES_SENT,REMOTE_BYTES_RECVD,LOCAL_SEND_SIZE + CPSW/CPSW2g/CPSW3g Ethernet Driver ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ -- CPSW3g: AM62px +- CPSW3g: AM64x, AM62x, AM62ax, AM62px + .. rubric:: TCP Bidirectional Throughput :name: CPSW2g-tcp-bidirectional-throughput @@ -503,7 +553,10 @@ CPSW/CPSW2g/CPSW3g Ethernet Driver .. csv-table:: CPSW2g TCP Bidirectional Throughput :header: "Command Used","am62pxx_sk-fs: THROUGHPUT (Mbits/sec)","am62pxx_sk-fs: CPU Load % (LOCAL_CPU_UTIL)" - "netperf -H 192.168.0.1 -j -c -C -l 60 -t TCP_STREAM; netperf -H 192.168.0.1 -j -c -C -l 60 -t TCP_MAERTS","1580.77","62.51" + "netperf -H 192.168.0.1 -j -c -C -l 60 -t TCP_STREAM; netperf -H 192.168.0.1 -j -c -C -l 60 -t TCP_MAERTS","1818.42","64.41" + + + .. rubric:: TCP Bidirectional Throughput Interrupt Pacing :name: CPSW2g-tcp-bidirectional-throughput-interrupt-pacing @@ -511,7 +564,10 @@ CPSW/CPSW2g/CPSW3g Ethernet Driver .. csv-table:: CPSW2g TCP Bidirectional Throughput Interrupt Pacing :header: "Command Used","am62pxx_sk-fs: THROUGHPUT (Mbits/sec)","am62pxx_sk-fs: CPU Load % (LOCAL_CPU_UTIL)" - "netperf -H 192.168.0.1 -j -c -C -l 60 -t TCP_STREAM; netperf -H 192.168.0.1 -j -c -C -l 60 -t TCP_MAERTS","1645.43","33.49" + "netperf -H 192.168.0.1 -j -c -C -l 60 -t TCP_STREAM; netperf -H 192.168.0.1 -j -c -C -l 60 -t TCP_MAERTS","1812.61","29.43" + + + .. rubric:: UDP Throughput :name: CPSW2g-udp-throughput-0-loss @@ -519,248 +575,375 @@ CPSW/CPSW2g/CPSW3g Ethernet Driver .. csv-table:: CPSW2g UDP Egress Throughput 0 loss :header: "Frame Size(bytes)","am62pxx_sk-fs: UDP Datagram Size(bytes) (LOCAL_SEND_SIZE)","am62pxx_sk-fs: THROUGHPUT (Mbits/sec)","am62pxx_sk-fs: Packets Per Second (kPPS)","am62pxx_sk-fs: CPU Load % (LOCAL_CPU_UTIL)" - "64","18.00","15.72","109.00","40.18" - "128","82.00","69.74","106.00","39.87" - "256","210.00","178.63","106.00","39.75" - "1024","978.00","804.42","103.00","39.63" - "1518","1472.00","956.34","81.00","35.64" + "64","18.00","55.07","108.00","39.97" + "128","82.00","108.69","106.00","39.97" + "256","210.00","218.80","107.00","39.98" + "1024","978.00","839.58","102.00","39.48" + "1518","1472.00","856.67","71.00","37.44" + + + | Linux OSPI Flash Driver -------------------------- +----------------------- + + + AM62PXX-SK -^^^^^^^^^^^^^^^^^^^^^^^^^^^ +^^^^^^^^^^ + UBIFS -""""""""""""""""""""""""""" +""""" .. csv-table:: OSPI Flash Driver :header: "Buffer size (bytes)","am62pxx_sk-fs: Write UBIFS Throughput (Mbytes/sec)","am62pxx_sk-fs: Write UBIFS CPU Load (%)","am62pxx_sk-fs: Read UBIFS Throughput (Mbytes/sec)","am62pxx_sk-fs: Read UBIFS CPU Load (%)" - "102400","0.18 (min 0.13, max 0.28)","28.87 (min 24.74, max 31.69)","28.60","9.68" - "262144","0.14 (min 0.11, max 0.18)","28.98 (min 25.40, max 31.51)","28.45","6.67" - "524288","0.14 (min 0.11, max 0.18)","29.47 (min 27.53, max 31.76)","28.47","3.45" - "1048576","0.14 (min 0.11, max 0.18)","27.99 (min 26.58, max 28.91)","28.00","3.45" + "102400","0.18 (min 0.13, max 0.29)","27.43 (min 24.77, max 30.04)","28.58","12.90" + "262144","0.15 (min 0.11, max 0.19)","28.31 (min 26.75, max 30.10)","28.55","6.90" + "524288","0.15 (min 0.11, max 0.19)","28.51 (min 27.57, max 30.32)","28.26","12.90" + "1048576","0.15 (min 0.11, max 0.19)","28.08 (min 25.48, max 30.41)","28.21","9.68" + + RAW -""""""""""""""""""""""""""" +""" .. csv-table:: OSPI Raw Flash Driver :header: "File size (Mbytes)","am62pxx_sk-fs: Raw Read Throughput (Mbytes/sec)" "50","37.88" + + + EMMC Driver ----------- -.. important:: +.. warning:: - The performance numbers can be severely affected if the media is + **IMPORTANT**: The performance numbers can be severely affected if the media is mounted in sync mode. Hot plug scripts in the filesystem mount removable media in sync mode to ensure data integrity. For performance sensitive applications, umount the auto-mounted filesystem and re-mount in async mode. + + EMMC EXT4 FIO 1G ^^^^^^^^^^^^^^^^ + .. csv-table:: EMMC EXT4 FIO 1G :header: "Buffer size (bytes)","am62pxx_sk-fs: Write EXT4 Throughput (Mbytes/sec)","am62pxx_sk-fs: Write EXT4 CPU Load (%)","am62pxx_sk-fs: Read EXT4 Throughput (Mbytes/sec)","am62pxx_sk-fs: Read EXT4 CPU Load (%)" - "1m","91.50","2.05","172.00","3.03" - "4m","97.10","1.57","169.00","2.17" - "4k","64.50","31.80","88.50","30.07" - "256k","91.30","3.13","171.00","4.34" + "1m","90.50","1.61","171.00","1.83" + "4m","95.60","0.99","172.00","1.41" + "4k","64.30","21.67","93.40","22.26" + "256k","90.50","2.01","171.00","2.62" + + + EMMC RAW FIO 1G ^^^^^^^^^^^^^^^ + + .. csv-table:: EMMC RAW FIO 1G :header: "Buffer size (bytes)","am62pxx_sk-fs: Write Raw Throughput (Mbytes/sec)","am62pxx_sk-fs: Write Raw CPU Load (%)","am62pxx_sk-fs: Read Raw Throughput (Mbytes/sec)","am62pxx_sk-fs: Read Raw CPU Load (%)" - "1m","91.20","1.35","174.00","1.77" - "4m","97.60","0.93","171.00","1.27" - "4k","64.90","15.82","92.70","20.34" - "256k","91.20","1.63","173.00","2.40" + "1m","90.10","1.38","174.00","1.76" + "4m","97.00","1.01","138.00","1.20" + "4k","64.70","16.10","92.40","20.30" + "256k","90.20","1.85","173.00","2.45" + + + EMMC EXT4 ^^^^^^^^^ + + .. csv-table:: EMMC EXT4 :header: "Buffer size (bytes)","am62pxx_sk-fs: Write EXT4 Throughput (Mbytes/sec)","am62pxx_sk-fs: Write EXT4 CPU Load (%)","am62pxx_sk-fs: Read EXT4 Throughput (Mbytes/sec)","am62pxx_sk-fs: Read EXT4 CPU Load (%)" - "102400","85.52 (min 79.45, max 87.92)","6.24 (min 5.36, max 8.32)","178.28","9.32" - "262144","78.27 (min 52.51, max 87.84)","5.68 (min 3.75, max 7.72)","181.22","10.73" - "524288","74.00 (min 51.70, max 86.90)","5.55 (min 3.33, max 7.84)","181.93","9.52" - "1048576","73.58 (min 54.96, max 88.20)","5.39 (min 3.79, max 7.49)","181.98","9.17" - "5242880","74.34 (min 58.42, max 86.76)","5.39 (min 3.82, max 7.92)","181.95","8.33" + "102400","83.74 (min 81.64, max 85.23)","5.78 (min 4.97, max 7.62)","178.99","8.97" + "262144","82.29 (min 61.99, max 90.58)","5.84 (min 3.99, max 8.08)","181.07","9.48" + "524288","79.81 (min 51.54, max 89.29)","5.46 (min 3.20, max 8.02)","181.88","9.52" + "1048576","73.44 (min 49.22, max 89.90)","5.08 (min 2.94, max 8.17)","181.74","8.66" + "5242880","74.59 (min 50.87, max 87.83)","5.30 (min 3.27, max 8.33)","181.91","8.30" + + + +EMMC EXT2 +^^^^^^^^^ + + + +.. csv-table:: EMMC EXT2 + :header: "Buffer size (bytes)","am62pxx_sk-fs: Write EXT2 Throughput (Mbytes/sec)","am62pxx_sk-fs: Write EXT2 CPU Load (%)","am62pxx_sk-fs: Read EXT2 Throughput (Mbytes/sec)","am62pxx_sk-fs: Read EXT2 CPU Load (%)" + + "102400","36.82 (min 7.02, max 46.19)","4.08 (min 2.94, max 4.65)","140.53","8.00" + "262144","50.11 (min 20.89, max 59.79)","5.67 (min 2.87, max 11.52)","177.12","14.41" + "524288","51.44 (min 7.44, max 67.81)","4.78 (min 2.24, max 6.37)","137.86","7.79" + "1048576","52.32 (min 7.48, max 68.81)","4.22 (min 2.35, max 5.42)","152.16","7.97" + "5242880","54.76 (min 7.50, max 70.98)","4.21 (min 2.39, max 4.92)","152.90","6.32" + EMMC VFAT ^^^^^^^^^ + + .. csv-table:: EMMC VFAT :header: "Buffer size (bytes)","am62pxx_sk-fs: Write VFAT Throughput (Mbytes/sec)","am62pxx_sk-fs: Write VFAT CPU Load (%)","am62pxx_sk-fs: Read VFAT Throughput (Mbytes/sec)","am62pxx_sk-fs: Read VFAT CPU Load (%)" - "102400","36.69 (min 11.54, max 45.34)","4.74 (min 3.72, max 5.55)","137.56","7.82" - "262144","43.58 (min 10.99, max 61.71)","4.96 (min 4.07, max 6.86)","176.50","10.08" - "524288","50.97 (min 12.62, max 65.36)","5.65 (min 4.45, max 6.49)","167.18","8.84" - "1048576","46.55 (min 12.36, max 66.45)","4.51 (min 3.14, max 5.96)","166.91","8.76" - "5242880","48.35 (min 12.52, max 69.13)","4.40 (min 3.32, max 5.51)","167.83","8.80" + "102400","35.84 (min 11.13, max 50.42)","4.93 (min 3.85, max 6.36)","173.89","8.40" + "262144","45.10 (min 11.46, max 58.09)","6.41 (min 4.90, max 7.20)","175.12","14.23" + "524288","50.40 (min 12.70, max 72.57)","4.73 (min 3.83, max 5.93)","176.19","8.82" + "1048576","57.99 (min 12.88, max 75.35)","5.09 (min 4.03, max 5.97)","176.20","9.21" + "5242880","56.58 (min 13.07, max 82.67)","5.09 (min 3.91, max 6.71)","176.26","9.62" + + UBoot EMMC Driver ----------------- + + + .. csv-table:: UBOOT EMMC RAW :header: "File size (bytes in hex)","am62pxx_sk-fs: Write Throughput (Kbytes/sec)","am62pxx_sk-fs: Read Throughput (Kbytes/sec)" - "2000000","97234.42","143719.30" - "4000000","98550.38","174762.67" + "2000000","98402.40","172463.16" + "4000000","96946.75","174762.67" + + + + MMCSD ----- -.. important:: +.. warning:: - The performance numbers can be severely affected if the media is + **IMPORTANT**: The performance numbers can be severely affected if the media is mounted in sync mode. Hot plug scripts in the filesystem mount removable media in sync mode to ensure data integrity. For performance sensitive applications, umount the auto-mounted filesystem and re-mount in async mode. + MMC EXT4 FIO 1G ^^^^^^^^^^^^^^^ + + + .. csv-table:: MMC EXT4 FIO 1G :header: "Buffer size (bytes)","am62pxx_sk-fs: Write EXT4 Throughput (Mbytes/sec)","am62pxx_sk-fs: Write EXT4 CPU Load (%)","am62pxx_sk-fs: Read EXT4 Throughput (Mbytes/sec)","am62pxx_sk-fs: Read EXT4 CPU Load (%)" - "1m","42.40","0.95","87.40","1.26" - "4m","42.00","0.62","87.40","0.87" - "4k","2.80","1.59","12.90","4.33" - "256k","37.10","1.09","83.90","1.49" + "1m","42.10","1.04","87.20","1.28" + "4m","42.90","0.79","87.30","0.97" + "4k","2.78","1.69","12.80","4.25" + "256k","38.10","1.22","83.70","1.55" + + + + MMC RAW FIO 1G ^^^^^^^^^^^^^^ + + + .. csv-table:: MMC RAW FIO 1G :header: "Buffer size (bytes)","am62pxx_sk-fs: Write Raw Throughput (Mbytes/sec)","am62pxx_sk-fs: Write Raw CPU Load (%)","am62pxx_sk-fs: Read Raw Throughput (Mbytes/sec)","am62pxx_sk-fs: Read Raw CPU Load (%)" - "1m","42.70","0.81","88.30","1.10" - "4m","42.70","0.66","88.30","0.81" - "4k","2.83","1.33","13.10","3.82" - "256k","36.50","0.94","84.50","1.41" + "1m","43.10","1.40","88.00","1.87" + "4m","43.60","1.02","88.00","1.40" + "4k","2.77","2.21","13.00","6.68" + "256k","36.80","1.56","84.00","2.38" + + + MMC EXT4 ^^^^^^^^ + + .. csv-table:: MMC EXT4 :header: "Buffer size (bytes)","am62pxx_sk-fs: Write Raw Throughput (Mbytes/sec)","am62pxx_sk-fs: Write Raw CPU Load (%)","am62pxx_sk-fs: Read Raw Throughput (Mbytes/sec)","am62pxx_sk-fs: Read Raw CPU Load (%)" - "102400","10.62 (min 10.41, max 10.73)","0.82 (min 0.72, max 1.04)","10.71","0.68" - "262144","10.52 (min 10.30, max 10.84)","0.85 (min 0.73, max 1.26)","11.14","0.69" - "524288","10.63 (min 10.60, max 10.67)","0.85 (min 0.76, max 1.09)","11.08","0.63" - "1048576","10.43 (min 10.36, max 10.57)","0.86 (min 0.72, max 1.06)","10.87","0.62" - "5242880","11.13 (min 11.06, max 11.17)","0.84 (min 0.72, max 1.21)","12.01","0.69" + "102400","10.37 (min 10.30, max 10.63)","0.80 (min 0.69, max 1.12)","10.68","0.76" + "262144","10.36 (min 10.28, max 10.49)","0.85 (min 0.63, max 1.32)","11.22","0.67" + "524288","10.40 (min 10.33, max 10.47)","0.83 (min 0.72, max 1.05)","11.50","0.63" + "1048576","10.45 (min 10.40, max 10.55)","0.86 (min 0.80, max 1.03)","11.05","0.66" + "5242880","10.65 (min 10.31, max 10.87)","0.83 (min 0.70, max 1.08)","12.03","0.66" + + + +MMC EXT3 +^^^^^^^^ + + + +.. csv-table:: MMC EXT3 + :header: "Buffer size (bytes)","am62pxx_sk-fs: Write Raw Throughput (Mbytes/sec)","am62pxx_sk-fs: Write Raw CPU Load (%)","am62pxx_sk-fs: Read Raw Throughput (Mbytes/sec)","am62pxx_sk-fs: Read Raw CPU Load (%)" + + "102400","40.74 (min 35.43, max 45.16)","3.67 (min 2.97, max 6.09)","80.08","4.23" + "262144","40.14 (min 35.43, max 41.78)","3.57 (min 2.82, max 5.91)","82.32","5.14" + "524288","36.55 (min 32.31, max 38.14)","3.16 (min 2.55, max 5.39)","90.35","4.75" + "1048576","36.11 (min 31.85, max 37.76)","3.26 (min 2.62, max 5.39)","90.20","4.95" + "5242880","39.66 (min 34.22, max 42.15)","3.30 (min 2.58, max 5.32)","90.47","4.98" + + + + +MMC EXT2 +^^^^^^^^ + + + + +.. csv-table:: MMC EXT2 + :header: "Buffer size (bytes)","am62pxx_sk-fs: Write Raw Throughput (Mbytes/sec)","am62pxx_sk-fs: Write Raw CPU Load (%)","am62pxx_sk-fs: Read Raw Throughput (Mbytes/sec)","am62pxx_sk-fs: Read Raw CPU Load (%)" + + "102400","17.47 (min 3.49, max 21.09)","2.33 (min 2.14, max 2.47)","76.97","4.41" + "262144","31.21 (min 3.88, max 38.66)","3.17 (min 2.06, max 3.75)","81.53","4.50" + "524288","34.84 (min 3.99, max 42.82)","3.34 (min 2.05, max 4.00)","90.76","4.76" + "1048576","33.57 (min 4.02, max 41.12)","2.83 (min 2.09, max 3.23)","90.54","4.35" + "5242880","32.54 (min 4.03, max 39.87)","2.62 (min 2.03, max 2.86)","90.56","4.55" + + + + The performance numbers were captured using the following: - SanDisk Max Endurance SD card (SDSQQVR-032G-GN6IA) - Partition was mounted with async option + UBoot MMCSD ----------- + UBOOT MMCSD FAT ^^^^^^^^^^^^^^^ + + + .. csv-table:: UBOOT MMCSD FAT :header: "File size (bytes in hex)","am62pxx_sk-fs: Write Throughput (Kbytes/sec)","am62pxx_sk-fs: Read Throughput (Kbytes/sec)" - "400000","40156.86","81920.00" - "800000","45765.36","87148.94" - "1000000","47627.91","90021.98" + "400000","38280.37","81920.00" + "800000","40554.46","87148.94" + "1000000","42666.67","90021.98" + + The performance numbers were captured using the following: - SanDisk Max Endurance SD card (SDSQQVR-032G-GN6IA) + + | USB Driver -------------------------- +---------- + + USB Device Controller -^^^^^^^^^^^^^^^^^^^^^^^^^^^ +^^^^^^^^^^^^^^^^^^^^^ + + .. csv-table:: USBDEVICE HIGHSPEED SLAVE_READ_THROUGHPUT :header: "Number of Blocks","am62pxx_sk-fs: Throughput (MB/sec)" - "150","31.60" + "150","44.00" + + .. csv-table:: USBDEVICE HIGHSPEED SLAVE_WRITE_THROUGHPUT :header: "Number of Blocks","am62pxx_sk-fs: Throughput (MB/sec)" - "150","27.70" + "150","37.30" + + | CRYPTO Driver -------------------------- +------------- + OpenSSL Performance -^^^^^^^^^^^^^^^^^^^^^^^^^^^ +^^^^^^^^^^^^^^^^^^^ .. csv-table:: OpenSSL Performance :header: "Algorithm","Buffer Size (in bytes)","am62pxx_sk-fs: throughput (KBytes/Sec)" - "aes-128-cbc","1024","23437.31" - "aes-128-cbc","16","418.09" - "aes-128-cbc","16384","85475.33" - "aes-128-cbc","256","6860.89" - "aes-128-cbc","64","1826.92" - "aes-128-cbc","8192","71669.08" - "aes-128-ecb","1024","23842.13" - "aes-128-ecb","16","431.57" - "aes-128-ecb","16384","87714.47" - "aes-128-ecb","256","7033.69" - "aes-128-ecb","64","1867.22" - "aes-128-ecb","8192","73673.39" - "aes-192-cbc","1024","22849.19" - "aes-192-cbc","16","419.90" - "aes-192-cbc","16384","77130.41" - "aes-192-cbc","256","6825.47" - "aes-192-cbc","64","1823.79" - "aes-192-cbc","8192","65653.42" - "aes-192-ecb","1024","23388.50" - "aes-192-ecb","16","430.66" - "aes-192-ecb","16384","79205.72" - "aes-192-ecb","256","6999.81" - "aes-192-ecb","64","1872.60" + "aes-128-cbc","1024","23455.06" + "aes-128-cbc","16","421.01" + "aes-128-cbc","16384","85224.11" + "aes-128-cbc","256","6907.22" + "aes-128-cbc","64","1825.54" + "aes-128-cbc","8192","71923.03" + "aes-128-ecb","1024","23801.51" + "aes-128-ecb","16","430.91" + "aes-128-ecb","16384","87780.01" + "aes-128-ecb","256","7053.99" + "aes-128-ecb","64","1870.70" + "aes-128-ecb","8192","73659.73" + "aes-192-cbc","1024","22887.42" + "aes-192-cbc","16","423.09" + "aes-192-cbc","16384","76808.19" + "aes-192-cbc","256","6910.29" + "aes-192-cbc","64","1825.79" + "aes-192-cbc","8192","65912.83" + "aes-192-ecb","1024","23310.34" + "aes-192-ecb","16","429.07" + "aes-192-ecb","16384","79123.80" + "aes-192-ecb","256","6940.50" + "aes-192-ecb","64","1837.50" "aes-192-ecb","8192","67548.50" - "aes-256-cbc","1024","22113.62" - "aes-256-cbc","16","418.55" - "aes-256-cbc","16384","70593.19" - "aes-256-cbc","256","6764.97" - "aes-256-cbc","64","1822.06" - "aes-256-cbc","8192","61227.01" - "aes-256-ecb","1024","22848.51" - "aes-256-ecb","16","429.44" - "aes-256-ecb","16384","72406.36" - "aes-256-ecb","256","6989.06" - "aes-256-ecb","64","1867.35" - "aes-256-ecb","8192","62788.95" - "sha256","1024","35276.46" - "sha256","16","584.42" - "sha256","16384","290111.49" - "sha256","256","9243.48" - "sha256","64","2319.10" - "sha256","8192","191660.03" - "sha512","1024","24752.81" - "sha512","16","563.81" - "sha512","16384","67693.23" - "sha512","256","8133.46" - "sha512","64","2253.44" - "sha512","8192","60473.34" + "aes-256-cbc","1024","22148.10" + "aes-256-cbc","16","421.00" + "aes-256-cbc","16384","69435.39" + "aes-256-cbc","256","6811.73" + "aes-256-cbc","64","1828.39" + "aes-256-cbc","8192","60792.83" + "aes-256-ecb","1024","22760.45" + "aes-256-ecb","16","432.77" + "aes-256-ecb","16384","72286.21" + "aes-256-ecb","256","6953.05" + "aes-256-ecb","64","1869.91" + "aes-256-ecb","8192","62401.19" + "sha256","1024","34637.82" + "sha256","16","574.17" + "sha256","16384","284317.01" + "sha256","256","8920.66" + "sha256","64","2263.98" + "sha256","8192","190515.88" + "sha512","1024","24553.81" + "sha512","16","560.17" + "sha512","16384","67567.62" + "sha512","256","8053.59" + "sha512","64","2252.35" + "sha512","8192","60293.12" + + .. csv-table:: OpenSSL CPU Load :header: "Algorithm","am62pxx_sk-fs: CPU Load" @@ -768,22 +951,24 @@ OpenSSL Performance "aes-128-cbc","31.00" "aes-128-ecb","32.00" "aes-192-cbc","31.00" - "aes-192-ecb","31.00" + "aes-192-ecb","32.00" "aes-256-cbc","30.00" "aes-256-ecb","31.00" "sha256","96.00" "sha512","96.00" -Listed for each algorithm are the code snippets used to run each benchmark test. + + +Listed for each algorithm are the code snippets used to run each + benchmark test. :: time -v openssl speed -elapsed -evp aes-128-cbc -| - + Low Power Performance -------------------------- +--------------------- Table: **Deep sleep** diff --git a/source/devices/AM62PX/linux/RT_Linux_Performance_Guide.rst b/source/devices/AM62PX/linux/RT_Linux_Performance_Guide.rst index 6cfe55eb7..780c34920 100644 --- a/source/devices/AM62PX/linux/RT_Linux_Performance_Guide.rst +++ b/source/devices/AM62PX/linux/RT_Linux_Performance_Guide.rst @@ -1,6 +1,6 @@ -====================================== - RT-linux 11.00 Performance Guide -====================================== +================================= + RT-linux 11.01 Performance Guide +================================= .. rubric:: **Read This First** :name: read-this-first-kernel-perf-guide @@ -34,7 +34,7 @@ https://e2e.ti.com/ or https://support.ti.com/ | System Benchmarks -------------------------- +----------------- Stress-ng and Cyclic Test diff --git a/source/devices/AM62PX/linux/Release_Specific_Release_Notes.rst b/source/devices/AM62PX/linux/Release_Specific_Release_Notes.rst index dad5487f1..a2a87a38a 100644 --- a/source/devices/AM62PX/linux/Release_Specific_Release_Notes.rst +++ b/source/devices/AM62PX/linux/Release_Specific_Release_Notes.rst @@ -36,18 +36,18 @@ found on the SDK download page or in the installed directory as indicated below. - Linux Manifest: :file:`/manifest/software_manifest.htm` -Release 11.00.09.04 +Release 11.01.05.03 =================== -Released on April 2025 +Released on July 2025 What's new ---------- **Processor SDK Linux AM62PX Release has following new features:** -- Second 2025 LTS Reference Release Including RT combined branch model - - Falcon mode through R5 SPL. + - Second 2025 LTS Reference Release Including RT combined branch model + - Falcon mode through R5 SPL :ref:`U-Boot Falcon Mode ` - Important Bug Fixes on top of Processor SDK 11.00.09.04 Release. - Review Issue Tracker Section for the new fixes. diff --git a/source/devices/AM62X/linux/Linux_Performance_Guide.rst b/source/devices/AM62X/linux/Linux_Performance_Guide.rst index ec427377d..18aafdd1a 100644 --- a/source/devices/AM62X/linux/Linux_Performance_Guide.rst +++ b/source/devices/AM62X/linux/Linux_Performance_Guide.rst @@ -1,7 +1,7 @@ -====================================== - Linux 11.00 Performance Guide -====================================== +============================== + Linux 11.01 Performance Guide +============================== .. rubric:: **Read This First** :name: read-this-first-kernel-perf-guide @@ -34,18 +34,20 @@ with drivers included in a particular release. :name: if-you-need-assistance-kernel-perf-guide For further information or to report any problems, contact -https://e2e.ti.com/ or https://support.ti.com/ +http://e2e.ti.com/ or http://support.ti.com/ + System Benchmarks -------------------- +----------------- + LMBench -^^^^^^^^^^^^^^^^^^^^^^^^^^^ -LMBench is a collection of microbenchmarks of which the memory bandwidth -and latency related ones are typically used to estimate processor +^^^^^^^ +LMBench is a collection of microbenchmarks of which the memory bandwidth +and latency related ones are typically used to estimate processor memory system performance. More information about lmbench at -https://lmbench.sourceforge.net/whatis_lmbench.html and -https://lmbench.sourceforge.net/man/lmbench.8.html +http://lmbench.sourceforge.net/whatis_lmbench.html and +http://lmbench.sourceforge.net/man/lmbench.8.html **Latency**: lat_mem_rd-stride128-szN, where N is equal to or smaller than the cache size at given level measures the cache miss penalty. N that is at least @@ -65,133 +67,136 @@ Execute the LMBench with the following: ./runltp -P j721e-idk-gw -f ddt/lmbench -s LMBENCH_L_PERF_0001 .. csv-table:: LMBench Benchmarks - :header: "Benchmarks","am62xx_lp_sk-fs: perf","am62xx_sk-fs: perf" - - "af_unix_sock_stream_latency (microsec)","32.14","29.59" - "af_unix_socket_stream_bandwidth (MBs)","540.61","600.15" - "bw_file_rd-io-1mb (MB/s)","880.13","954.50" - "bw_file_rd-o2c-1mb (MB/s)","455.10","532.86" - "bw_mem-bcopy-16mb (MB/s)","672.10","758.26" - "bw_mem-bcopy-1mb (MB/s)","700.89","801.03" - "bw_mem-bcopy-2mb (MB/s)","669.01","769.97" - "bw_mem-bcopy-4mb (MB/s)","695.05","799.28" - "bw_mem-bcopy-8mb (MB/s)","719.62","803.29" - "bw_mem-bzero-16mb (MB/s)","1730.10","1760.56" - "bw_mem-bzero-1mb (MB/s)","1208.67 (min 700.89, max 1716.44)","1265.57 (min 801.03, max 1730.10)" - "bw_mem-bzero-2mb (MB/s)","1194.20 (min 669.01, max 1719.39)","1251.84 (min 769.97, max 1733.70)" - "bw_mem-bzero-4mb (MB/s)","1206.88 (min 695.05, max 1718.71)","1268.33 (min 799.28, max 1737.37)" - "bw_mem-bzero-8mb (MB/s)","1218.00 (min 719.62, max 1716.37)","1276.44 (min 803.29, max 1749.59)" - "bw_mem-cp-16mb (MB/s)","393.28","455.17" - "bw_mem-cp-1mb (MB/s)","1124.57 (min 396.25, max 1852.88)","1161.67 (min 452.08, max 1871.26)" - "bw_mem-cp-2mb (MB/s)","1090.58 (min 391.93, max 1789.23)","1138.54 (min 459.56, max 1817.52)" - "bw_mem-cp-4mb (MB/s)","1091.43 (min 414.25, max 1768.61)","1133.29 (min 474.72, max 1791.85)" - "bw_mem-cp-8mb (MB/s)","1096.93 (min 433.67, max 1760.18)","1138.86 (min 501.13, max 1776.59)" - "bw_mem-fcp-16mb (MB/s)","689.03","740.02" - "bw_mem-fcp-1mb (MB/s)","1254.23 (min 792.02, max 1716.44)","1322.59 (min 915.08, max 1730.10)" - "bw_mem-fcp-2mb (MB/s)","1247.97 (min 776.55, max 1719.39)","1288.20 (min 842.70, max 1733.70)" - "bw_mem-fcp-4mb (MB/s)","1249.53 (min 780.34, max 1718.71)","1304.09 (min 870.80, max 1737.37)" - "bw_mem-fcp-8mb (MB/s)","1254.82 (min 793.26, max 1716.37)","1297.68 (min 845.76, max 1749.59)" - "bw_mem-frd-16mb (MB/s)","1056.11","1142.29" - "bw_mem-frd-1mb (MB/s)","991.15 (min 792.02, max 1190.27)","1100.11 (min 915.08, max 1285.14)" - "bw_mem-frd-2mb (MB/s)","901.98 (min 776.55, max 1027.40)","982.20 (min 842.70, max 1121.70)" - "bw_mem-frd-4mb (MB/s)","915.04 (min 780.34, max 1049.73)","986.06 (min 870.80, max 1101.32)" - "bw_mem-frd-8mb (MB/s)","925.87 (min 793.26, max 1058.48)","981.23 (min 845.76, max 1116.69)" - "bw_mem-fwr-16mb (MB/s)","1737.05","1767.96" - "bw_mem-fwr-1mb (MB/s)","1521.58 (min 1190.27, max 1852.88)","1578.20 (min 1285.14, max 1871.26)" - "bw_mem-fwr-2mb (MB/s)","1408.32 (min 1027.40, max 1789.23)","1469.61 (min 1121.70, max 1817.52)" - "bw_mem-fwr-4mb (MB/s)","1409.17 (min 1049.73, max 1768.61)","1446.59 (min 1101.32, max 1791.85)" - "bw_mem-fwr-8mb (MB/s)","1409.33 (min 1058.48, max 1760.18)","1446.64 (min 1116.69, max 1776.59)" - "bw_mem-rd-16mb (MB/s)","1055.27","1136.44" - "bw_mem-rd-1mb (MB/s)","890.15 (min 569.88, max 1210.41)","992.58 (min 663.90, max 1321.25)" - "bw_mem-rd-2mb (MB/s)","790.99 (min 522.47, max 1059.51)","873.95 (min 619.87, max 1128.03)" - "bw_mem-rd-4mb (MB/s)","822.28 (min 584.11, max 1060.45)","909.27 (min 687.64, max 1130.90)" - "bw_mem-rd-8mb (MB/s)","856.77 (min 657.57, max 1055.97)","941.08 (min 764.53, max 1117.63)" - "bw_mem-rdwr-16mb (MB/s)","657.14","757.00" - "bw_mem-rdwr-1mb (MB/s)","489.72 (min 396.25, max 583.18)","551.56 (min 452.08, max 651.04)" - "bw_mem-rdwr-2mb (MB/s)","450.81 (min 391.93, max 509.68)","534.94 (min 459.56, max 610.31)" - "bw_mem-rdwr-4mb (MB/s)","490.09 (min 414.25, max 565.93)","585.55 (min 474.72, max 696.38)" - "bw_mem-rdwr-8mb (MB/s)","517.41 (min 433.67, max 601.14)","623.01 (min 501.13, max 744.88)" - "bw_mem-wr-16mb (MB/s)","678.94","838.62" - "bw_mem-wr-1mb (MB/s)","576.53 (min 569.88, max 583.18)","657.47 (min 651.04, max 663.90)" - "bw_mem-wr-2mb (MB/s)","516.08 (min 509.68, max 522.47)","615.09 (min 610.31, max 619.87)" - "bw_mem-wr-4mb (MB/s)","575.02 (min 565.93, max 584.11)","692.01 (min 687.64, max 696.38)" - "bw_mem-wr-8mb (MB/s)","629.36 (min 601.14, max 657.57)","754.71 (min 744.88, max 764.53)" - "bw_mmap_rd-mo-1mb (MB/s)","1179.25","1290.80" - "bw_mmap_rd-o2c-1mb (MB/s)","452.42","518.76" - "bw_pipe (MB/s)","453.11","499.59" - "bw_unix (MB/s)","540.61","600.15" - "lat_connect (us)","67.32","58.96" - "lat_ctx-2-128k (us)","8.35","7.77" - "lat_ctx-2-256k (us)","14.15","10.21" - "lat_ctx-4-128k (us)","8.01","7.30" - "lat_ctx-4-256k (us)","12.53","11.14" - "lat_fs-0k (num_files)","209.00","212.00" - "lat_fs-10k (num_files)","91.00","95.00" - "lat_fs-1k (num_files)","126.00","144.00" - "lat_fs-4k (num_files)","144.00","137.00" - "lat_mem_rd-stride128-sz1000k (ns)","56.29","51.97" - "lat_mem_rd-stride128-sz125k (ns)","6.26","5.58" - "lat_mem_rd-stride128-sz250k (ns)","6.53","5.83" - "lat_mem_rd-stride128-sz31k (ns)","4.69","2.16" - "lat_mem_rd-stride128-sz50 (ns)","2.40","2.15" - "lat_mem_rd-stride128-sz500k (ns)","21.81","16.52" - "lat_mem_rd-stride128-sz62k (ns)","5.85","5.24" - "lat_mmap-1m (us)","69.00","61.00" - "lat_ops-double-add (ns)","3.21","2.86" - "lat_ops-double-div (ns)","17.63","15.80" - "lat_ops-double-mul (ns)","3.21","2.86" - "lat_ops-float-add (ns)","3.21","2.86" - "lat_ops-float-div (ns)","10.44","9.30" - "lat_ops-float-mul (ns)","3.21","2.86" - "lat_ops-int-add (ns)","0.80","0.72" - "lat_ops-int-bit (ns)","0.53","0.48" - "lat_ops-int-div (ns)","4.81","4.30" - "lat_ops-int-mod (ns)","5.08","4.53" - "lat_ops-int-mul (ns)","3.45","3.08" - "lat_ops-int64-add (ns)","0.80","0.72" - "lat_ops-int64-bit (ns)","0.53","0.48" - "lat_ops-int64-div (ns)","7.63","6.80" - "lat_ops-int64-mod (ns)","5.88","5.25" - "lat_ops-int64-mul (ns)","3.99","3.55" - "lat_pagefault (us)","0.78","0.72" - "lat_pipe (us)","28.50","25.19" - "lat_proc-exec (us)","1022.40","950.00" - "lat_proc-fork (us)","897.00","822.57" - "lat_proc-proccall (us)","0.01","0.01" - "lat_select (us)","38.10","33.92" - "lat_sem (us)","3.22","2.80" - "lat_sig-catch (us)","6.22","5.54" - "lat_sig-install (us)","0.76","0.68" - "lat_sig-prot (us)","1.75","0.89" - "lat_syscall-fstat (us)","2.22","1.97" - "lat_syscall-null (us)","0.51","0.46" - "lat_syscall-open (us)","205.22","153.76" - "lat_syscall-read (us)","0.88","0.79" - "lat_syscall-stat (us)","5.35","4.82" - "lat_syscall-write (us)","0.85","0.76" - "lat_tcp (us)","1.02","0.91" - "lat_unix (us)","32.14","29.59" - "latency_for_0.50_mb_block_size (nanosec)","21.81","16.52" - "latency_for_1.00_mb_block_size (nanosec)","28.15 (min 0.00, max 56.29)","25.98 (min 0.00, max 51.97)" - "pipe_bandwidth (MBs)","453.11","499.59" - "pipe_latency (microsec)","28.50","25.19" - "procedure_call (microsec)","0.01","0.01" - "select_on_200_tcp_fds (microsec)","38.10","33.92" - "semaphore_latency (microsec)","3.22","2.80" - "signal_handler_latency (microsec)","0.76","0.68" - "signal_handler_overhead (microsec)","6.22","5.54" - "tcp_ip_connection_cost_to_localhost (microsec)","67.32","58.96" - "tcp_latency_using_localhost (microsec)","1.02","0.91" + :header: "Benchmarks","am62xx_lp_sk-fs: perf","am62xx_sk-fs: perf","am62xxsip_sk-fs: perf" + + "af_unix_sock_stream_latency (microsec)","33.72","29.04","33.12" + "af_unix_socket_stream_bandwidth (MBs)","549.13","599.28","654.14" + "bw_file_rd-io-1mb (MB/s)","886.84","947.27","882.30" + "bw_file_rd-o2c-1mb (MB/s)","488.68","501.67","478.39" + "bw_mem-bcopy-16mb (MB/s)","702.31","758.69","750.47" + "bw_mem-bcopy-1mb (MB/s)","731.40","810.64","790.83" + "bw_mem-bcopy-2mb (MB/s)","701.26","754.34","724.51" + "bw_mem-bcopy-4mb (MB/s)","725.16","797.77","742.25" + "bw_mem-bcopy-8mb (MB/s)","737.40","788.49","726.02" + "bw_mem-bzero-16mb (MB/s)","1777.58","1757.08","1747.87" + "bw_mem-bzero-1mb (MB/s)","1249.56 (min 731.40, max 1767.72)","1273.23 (min 810.64, max 1735.81)","1259.27 (min 790.83, max 1727.71)" + "bw_mem-bzero-2mb (MB/s)","1234.34 (min 701.26, max 1767.41)","1244.32 (min 754.34, max 1734.30)","1223.29 (min 724.51, max 1722.06)" + "bw_mem-bzero-4mb (MB/s)","1247.54 (min 725.16, max 1769.91)","1274.42 (min 797.77, max 1751.06)","1240.19 (min 742.25, max 1738.12)" + "bw_mem-bzero-8mb (MB/s)","1251.41 (min 737.40, max 1765.42)","1266.66 (min 788.49, max 1744.82)","1237.62 (min 726.02, max 1749.21)" + "bw_mem-cp-16mb (MB/s)","408.49","454.88","449.99" + "bw_mem-cp-1mb (MB/s)","1146.98 (min 392.46, max 1901.50)","1153.53 (min 430.17, max 1876.88)","1156.08 (min 446.83, max 1865.32)" + "bw_mem-cp-2mb (MB/s)","1126.54 (min 410.09, max 1842.98)","1128.29 (min 438.07, max 1818.51)","1103.62 (min 412.54, max 1794.69)" + "bw_mem-cp-4mb (MB/s)","1103.66 (min 388.31, max 1819.01)","1136.35 (min 485.38, max 1787.31)","1104.23 (min 437.49, max 1770.96)" + "bw_mem-cp-8mb (MB/s)","1122.78 (min 439.68, max 1805.87)","1125.20 (min 477.73, max 1772.66)","1079.77 (min 399.36, max 1760.18)" + "bw_mem-fcp-16mb (MB/s)","696.89","736.95","808.28" + "bw_mem-fcp-1mb (MB/s)","1292.16 (min 816.59, max 1767.72)","1310.31 (min 884.80, max 1735.81)","1264.43 (min 801.15, max 1727.71)" + "bw_mem-fcp-2mb (MB/s)","1260.50 (min 753.58, max 1767.41)","1303.26 (min 872.22, max 1734.30)","1260.02 (min 797.98, max 1722.06)" + "bw_mem-fcp-4mb (MB/s)","1282.65 (min 795.39, max 1769.91)","1316.69 (min 882.32, max 1751.06)","1256.21 (min 774.29, max 1738.12)" + "bw_mem-fcp-8mb (MB/s)","1283.72 (min 802.01, max 1765.42)","1300.13 (min 855.43, max 1744.82)","1286.00 (min 822.79, max 1749.21)" + "bw_mem-frd-16mb (MB/s)","1022.56","1132.74","1084.82" + "bw_mem-frd-1mb (MB/s)","994.66 (min 816.59, max 1172.73)","1082.72 (min 884.80, max 1280.64)","1006.95 (min 801.15, max 1212.75)" + "bw_mem-frd-2mb (MB/s)","879.12 (min 753.58, max 1004.65)","997.70 (min 872.22, max 1123.17)","937.88 (min 797.98, max 1077.78)" + "bw_mem-frd-4mb (MB/s)","924.01 (min 795.39, max 1052.63)","1003.36 (min 882.32, max 1124.39)","900.76 (min 774.29, max 1027.22)" + "bw_mem-frd-8mb (MB/s)","937.56 (min 802.01, max 1073.11)","988.33 (min 855.43, max 1121.23)","946.94 (min 822.79, max 1071.09)" + "bw_mem-fwr-16mb (MB/s)","1780.94","1753.81","1753.81" + "bw_mem-fwr-1mb (MB/s)","1537.12 (min 1172.73, max 1901.50)","1578.76 (min 1280.64, max 1876.88)","1539.04 (min 1212.75, max 1865.32)" + "bw_mem-fwr-2mb (MB/s)","1423.82 (min 1004.65, max 1842.98)","1470.84 (min 1123.17, max 1818.51)","1436.24 (min 1077.78, max 1794.69)" + "bw_mem-fwr-4mb (MB/s)","1435.82 (min 1052.63, max 1819.01)","1455.85 (min 1124.39, max 1787.31)","1399.09 (min 1027.22, max 1770.96)" + "bw_mem-fwr-8mb (MB/s)","1439.49 (min 1073.11, max 1805.87)","1446.95 (min 1121.23, max 1772.66)","1415.64 (min 1071.09, max 1760.18)" + "bw_mem-rd-16mb (MB/s)","1088.51","1135.96","1081.81" + "bw_mem-rd-1mb (MB/s)","912.37 (min 579.63, max 1245.11)","976.13 (min 648.72, max 1303.54)","946.52 (min 620.54, max 1272.50)" + "bw_mem-rd-2mb (MB/s)","791.99 (min 496.83, max 1087.15)","855.46 (min 573.48, max 1137.44)","828.62 (min 554.25, max 1102.99)" + "bw_mem-rd-4mb (MB/s)","825.24 (min 586.08, max 1064.40)","899.80 (min 674.42, max 1125.18)","834.36 (min 572.08, max 1096.64)" + "bw_mem-rd-8mb (MB/s)","883.75 (min 679.35, max 1088.14)","942.36 (min 754.93, max 1129.78)","830.91 (min 573.68, max 1088.14)" + "bw_mem-rdwr-16mb (MB/s)","675.96","830.95","597.26" + "bw_mem-rdwr-1mb (MB/s)","475.87 (min 392.46, max 559.28)","538.40 (min 430.17, max 646.62)","527.09 (min 446.83, max 607.35)" + "bw_mem-rdwr-2mb (MB/s)","452.73 (min 410.09, max 495.36)","506.60 (min 438.07, max 575.13)","480.47 (min 412.54, max 548.40)" + "bw_mem-rdwr-4mb (MB/s)","485.45 (min 388.31, max 582.58)","579.51 (min 485.38, max 673.63)","505.73 (min 437.49, max 573.97)" + "bw_mem-rdwr-8mb (MB/s)","567.00 (min 439.68, max 694.32)","619.49 (min 477.73, max 761.25)","487.83 (min 399.36, max 576.29)" + "bw_mem-wr-16mb (MB/s)","745.19","766.06","609.45" + "bw_mem-wr-1mb (MB/s)","569.46 (min 559.28, max 579.63)","647.67 (min 646.62, max 648.72)","613.95 (min 607.35, max 620.54)" + "bw_mem-wr-2mb (MB/s)","496.10 (min 495.36, max 496.83)","574.31 (min 573.48, max 575.13)","551.33 (min 548.40, max 554.25)" + "bw_mem-wr-4mb (MB/s)","584.33 (min 582.58, max 586.08)","674.03 (min 673.63, max 674.42)","573.03 (min 572.08, max 573.97)" + "bw_mem-wr-8mb (MB/s)","686.84 (min 679.35, max 694.32)","758.09 (min 754.93, max 761.25)","574.99 (min 573.68, max 576.29)" + "bw_mmap_rd-mo-1mb (MB/s)","1219.51","1265.82","1244.22" + "bw_mmap_rd-o2c-1mb (MB/s)","472.52","500.92","446.10" + "bw_pipe (MB/s)","462.28","507.32","468.69" + "bw_unix (MB/s)","549.13","599.28","654.14" + "lat_connect (us)","66.58","59.62","66.43" + "lat_ctx-2-128k (us)","8.57","7.38","9.05" + "lat_ctx-2-256k (us)","10.84","10.58","12.25" + "lat_ctx-4-128k (us)","8.82","7.32","8.58" + "lat_ctx-4-256k (us)","9.82","9.93","10.74" + "lat_fs-0k (num_files)","209.00","220.00","182.00" + "lat_fs-10k (num_files)","87.00","96.00","93.00" + "lat_fs-1k (num_files)","126.00","143.00","136.00" + "lat_fs-4k (num_files)","131.00","130.00","123.00" + "lat_mem_rd-stride128-sz1000k (ns)","54.80","51.88","53.94" + "lat_mem_rd-stride128-sz125k (ns)","6.22","5.57","6.22" + "lat_mem_rd-stride128-sz250k (ns)","6.55","5.84","6.55" + "lat_mem_rd-stride128-sz31k (ns)","4.70","2.16","4.09" + "lat_mem_rd-stride128-sz50 (ns)","2.40","2.15","2.41" + "lat_mem_rd-stride128-sz500k (ns)","15.61","24.71","15.47" + "lat_mem_rd-stride128-sz62k (ns)","5.89","5.25","5.86" + "lat_mmap-1m (us)","60.00","64.00","61.00" + "lat_ops-double-add (ns)","3.21","2.86","3.21" + "lat_ops-double-div (ns)","17.63","15.75","17.63" + "lat_ops-double-mul (ns)","3.21","2.86","3.20" + "lat_ops-float-add (ns)","3.21","2.86","3.21" + "lat_ops-float-div (ns)","10.43","9.31","10.42" + "lat_ops-float-mul (ns)","3.21","2.86","3.21" + "lat_ops-int-add (ns)","0.80","0.72","0.80" + "lat_ops-int-bit (ns)","0.53","0.48","0.53" + "lat_ops-int-div (ns)","4.81","4.29","4.81" + "lat_ops-int-mod (ns)","5.08","4.53","5.08" + "lat_ops-int-mul (ns)","3.51","3.13","3.51" + "lat_ops-int64-add (ns)","0.80","0.72","0.80" + "lat_ops-int64-bit (ns)","0.53","0.48","0.53" + "lat_ops-int64-div (ns)","7.61","6.80","7.62" + "lat_ops-int64-mod (ns)","5.88","5.25","5.88" + "lat_ops-int64-mul (ns)","3.98","3.55","3.98" + "lat_pagefault (us)","0.78","0.71","1.55" + "lat_pipe (us)","27.48","25.15","28.22" + "lat_proc-exec (us)","1071.80","1023.00","1042.80" + "lat_proc-fork (us)","882.86","852.86","929.00" + "lat_proc-proccall (us)","0.01","0.01","0.01" + "lat_select (us)","38.14","34.07","37.98" + "lat_sem (us)","4.37","2.98","3.22" + "lat_sig-catch (us)","5.97","5.31","6.03" + "lat_sig-install (us)","0.73","0.65","0.73" + "lat_sig-prot (us)","0.78","0.58","0.74" + "lat_syscall-fstat (us)","2.23","1.97","2.22" + "lat_syscall-null (us)","0.52","0.46","0.52" + "lat_syscall-open (us)","185.60","155.14","224.73" + "lat_syscall-read (us)","0.90","0.80","0.90" + "lat_syscall-stat (us)","5.44","4.87","5.44" + "lat_syscall-write (us)","0.91","0.81","0.92" + "lat_tcp (us)","1.04","0.93","1.04" + "lat_unix (us)","33.72","29.04","33.12" + "latency_for_0.50_mb_block_size (nanosec)","15.61","24.71","15.47" + "latency_for_1.00_mb_block_size (nanosec)","27.40 (min 0.00, max 54.80)","25.94 (min 0.00, max 51.88)","26.97 (min 0.00, max 53.94)" + "pipe_bandwidth (MBs)","462.28","507.32","468.69" + "pipe_latency (microsec)","27.48","25.15","28.22" + "procedure_call (microsec)","0.01","0.01","0.01" + "select_on_200_tcp_fds (microsec)","38.14","34.07","37.98" + "semaphore_latency (microsec)","4.37","2.98","3.22" + "signal_handler_latency (microsec)","0.73","0.65","0.73" + "signal_handler_overhead (microsec)","5.97","5.31","6.03" + "tcp_ip_connection_cost_to_localhost (microsec)","66.58","59.62","66.43" + "tcp_latency_using_localhost (microsec)","1.04","0.93","1.04" + + Dhrystone -^^^^^^^^^^^^^^^^^^^^^^^^^^^ +^^^^^^^^^ Dhrystone is a core only benchmark that runs from warm L1 caches in all modern processors. It scales linearly with clock speed. Please take note, different run may produce different slightly results. -This is advised to run this test multiple times in order to get maximum +This is advised to run this test multiple times in order to get maximum performance numbers. + Execute the benchmark with the following: :: @@ -199,14 +204,16 @@ Execute the benchmark with the following: runDhrystone .. csv-table:: Dhrystone Benchmarks - :header: "Benchmarks","am62xx_lp_sk-fs: perf","am62xx_sk-fs: perf" + :header: "Benchmarks","am62xx_lp_sk-fs: perf","am62xx_sk-fs: perf","am62xxsip_sk-fs: perf" + + "cpu_clock (MHz)","1250.00","1400.00","1250.00" + "dhrystone_per_mhz (DMIPS/MHz)","2.90","2.90","2.90" + "dhrystone_per_second (DhrystoneP)","6451613.00","7142857.00","6451613.00" + - "cpu_clock (MHz)","1250.00","1400.00" - "dhrystone_per_mhz (DMIPS/MHz)","2.90","2.90" - "dhrystone_per_second (DhrystoneP)","6451613.00","7142857.00" Whetstone -^^^^^^^^^^^^^^^^^^^^^^^^^^^ +^^^^^^^^^ Whetstone is a benchmark primarily measuring floating-point arithmetic performance. Execute the benchmark with the following: @@ -216,22 +223,26 @@ Execute the benchmark with the following: runWhetstone .. csv-table:: Whetstone Benchmarks - :header: "Benchmarks","am62xx_lp_sk-fs: perf","am62xx_sk-fs: perf" + :header: "Benchmarks","am62xx_lp_sk-fs: perf","am62xx_sk-fs: perf","am62xxsip_sk-fs: perf" + + "whetstone (MIPS)","5000.00","5000.00","5000.00" + - "whetstone (MIPS)","5000.00","5000.00" Linpack -^^^^^^^^^^^^^^^^^^^^^^^^^^^ +^^^^^^^ Linpack measures peak double precision (64 bit) floating point performance in solving a dense linear system. .. csv-table:: Linpack Benchmarks - :header: "Benchmarks","am62xx_lp_sk-fs: perf","am62xx_sk-fs: perf" + :header: "Benchmarks","am62xx_lp_sk-fs: perf","am62xx_sk-fs: perf","am62xxsip_sk-fs: perf" + + "linpack (Kflops)","515516.00","575242.00","515670.00" + - "linpack (Kflops)","515361.00","576740.00" NBench -^^^^^^^^^^^^^^^^^^^^^^^^^^^ +^^^^^^ NBench which stands for Native Benchmark is used to measure macro benchmarks for commonly used operations such as sorting and analysis algorithms. More information about NBench at @@ -239,20 +250,22 @@ https://en.wikipedia.org/wiki/NBench and https://nbench.io/articles/index.html .. csv-table:: NBench Benchmarks - :header: "Benchmarks","am62xx_lp_sk-fs: perf","am62xx_sk-fs: perf" - - "assignment (Iterations)","12.93","14.52" - "fourier (Iterations)","20384.00","22831.00" - "fp_emulation (Iterations)","192.52","215.63" - "huffman (Iterations)","1057.30","1184.20" - "idea (Iterations)","3075.50","3444.70" - "lu_decomposition (Iterations)","473.04","530.53" - "neural_net (Iterations)","7.73","8.66" - "numeric_sort (Iterations)","557.39","628.41" - "string_sort (Iterations)","146.35","163.93" + :header: "Benchmarks","am62xx_lp_sk-fs: perf","am62xx_sk-fs: perf","am62xxsip_sk-fs: perf" + + "assignment (Iterations)","12.94","14.49","12.90" + "fourier (Iterations)","20380.00","22832.00","20384.00" + "fp_emulation (Iterations)","192.48","215.64","192.53" + "huffman (Iterations)","1057.10","1184.20","1057.10" + "idea (Iterations)","3075.10","3444.70","3075.50" + "lu_decomposition (Iterations)","473.03","530.04","473.87" + "neural_net (Iterations)","7.73","8.66","7.73" + "numeric_sort (Iterations)","561.61","629.58","557.46" + "string_sort (Iterations)","146.34","163.94","146.35" + + Stream -^^^^^^^^^^^^^^^^^^^^^^^^^^^ +^^^^^^ STREAM is a microbenchmark for measuring data memory system performance without any data reuse. It is designed to miss on caches and exercise data prefetcher and speculative accesses. @@ -270,51 +283,60 @@ Execute the benchmark with the following: stream_c .. csv-table:: Stream Benchmarks - :header: "Benchmarks","am62xx_lp_sk-fs: perf","am62xx_sk-fs: perf" + :header: "Benchmarks","am62xx_lp_sk-fs: perf","am62xx_sk-fs: perf","am62xxsip_sk-fs: perf" + + "add (MB/s)","1415.00","1447.90","1692.60" + "copy (MB/s)","1455.40","1638.00","1875.50" + "scale (MB/s)","1595.70","1792.70","1938.40" + "triad (MB/s)","1384.60","1477.10","1657.50" + + - "add (MB/s)","1366.70","1450.00" - "copy (MB/s)","1423.40","1637.80" - "scale (MB/s)","1567.90","1792.50" - "triad (MB/s)","1345.00","1478.40" CoreMarkPro -^^^^^^^^^^^^^^^^^^^^^^^^^^^ +^^^^^^^^^^^ CoreMark®-Pro is a comprehensive, advanced processor benchmark that works with and enhances the market-proven industry-standard EEMBC CoreMark® benchmark. While CoreMark stresses the CPU pipeline, CoreMark-Pro tests the entire processor, adding comprehensive support for multicore technology, a combination of integer and floating-point workloads, and data sets for utilizing larger memory subsystems. + .. csv-table:: CoreMarkPro Benchmarks - :header: "Benchmarks","am62xx_lp_sk-fs: perf","am62xx_sk-fs: perf" - - "cjpeg-rose7-preset (workloads/)","37.17","41.49" - "core (workloads/)","0.27","0.30" - "coremark-pro ()","781.81","863.81" - "linear_alg-mid-100x100-sp (workloads/)","13.10","14.68" - "loops-all-mid-10k-sp (workloads/)","0.60","0.66" - "nnet_test (workloads/)","0.97","1.09" - "parser-125k (workloads/)","7.35","8.26" - "radix2-big-64k (workloads/)","41.24","41.57" - "sha-test (workloads/)","71.94","80.65" - "zip-test (workloads/)","18.52","20.83" + :header: "Benchmarks","am62xx_lp_sk-fs: perf","am62xx_sk-fs: perf","am62xxsip_sk-fs: perf" + + "cjpeg-rose7-preset (workloads/)","37.17","41.84","37.59" + "core (workloads/)","0.27","0.30","0.27" + "coremark-pro ()","789.18","887.88","823.42" + "linear_alg-mid-100x100-sp (workloads/)","13.10","14.68","13.11" + "loops-all-mid-10k-sp (workloads/)","0.59","0.66","0.63" + "nnet_test (workloads/)","0.97","1.09","0.97" + "parser-125k (workloads/)","7.41","8.20","7.69" + "radix2-big-64k (workloads/)","45.59","53.18","57.64" + "sha-test (workloads/)","71.43","80.65","72.46" + "zip-test (workloads/)","18.52","20.83","18.87" + + + .. csv-table:: CoreMarkProTwoCore Benchmarks - :header: "Benchmarks","am62xx_lp_sk-fs: perf","am62xx_sk-fs: perf" - - "cjpeg-rose7-preset (workloads/)","74.07","82.64" - "core (workloads/)","0.54","0.60" - "coremark-pro ()","1352.68","1503.32" - "linear_alg-mid-100x100-sp (workloads/)","26.19","29.31" - "loops-all-mid-10k-sp (workloads/)","1.03","1.15" - "nnet_test (workloads/)","1.93","2.17" - "parser-125k (workloads/)","11.05","11.56" - "radix2-big-64k (workloads/)","37.88","43.39" - "sha-test (workloads/)","142.86","158.73" - "zip-test (workloads/)","34.48","38.46" + :header: "Benchmarks","am62xx_lp_sk-fs: perf","am62xx_sk-fs: perf","am62xxsip_sk-fs: perf" + + "cjpeg-rose7-preset (workloads/)","74.07","83.33","74.07" + "core (workloads/)","0.54","0.60","0.54" + "coremark-pro ()","1329.74","1487.25","1358.05" + "linear_alg-mid-100x100-sp (workloads/)","26.21","29.36","26.18" + "loops-all-mid-10k-sp (workloads/)","1.03","1.15","1.05" + "nnet_test (workloads/)","1.94","2.17","1.94" + "parser-125k (workloads/)","9.43","11.70","10.64" + "radix2-big-64k (workloads/)","38.37","38.53","40.37" + "sha-test (workloads/)","140.85","158.73","140.85" + "zip-test (workloads/)","34.48","38.46","34.48" + + MultiBench -^^^^^^^^^^^^^^^^^^^^^^^^^^^ +^^^^^^^^^^ MultiBench™ is a suite of benchmarks that allows processor and system designers to analyze, test, and improve multicore processors. It uses three forms of concurrency: Data decomposition: multiple threads cooperating on achieving a unified goal and @@ -331,135 +353,163 @@ abstraction layer and provides a flexible interface to allow a wide variety of thread-enabled workloads to be tested. .. csv-table:: Multibench Benchmarks - :header: "Benchmarks","am62xx_lp_sk-fs: perf","am62xx_sk-fs: perf" - - "4m-check (workloads/)","277.78","300.52" - "4m-check-reassembly (workloads/)","58.41","63.41" - "4m-check-reassembly-tcp (workloads/)","37.88","41.19" - "4m-check-reassembly-tcp-cmykw2-rotatew2 (workloads/)","21.83","23.79" - "4m-check-reassembly-tcp-x264w2 (workloads/)","1.60","1.77" - "4m-cmykw2 (workloads/)","180.67","199.01" - "4m-cmykw2-rotatew2 (workloads/)","35.40","38.96" - "4m-reassembly (workloads/)","45.45","49.38" - "4m-rotatew2 (workloads/)","41.02","43.92" - "4m-tcp-mixed (workloads/)","93.57","102.56" - "4m-x264w2 (workloads/)","1.61","1.86" - "empty-wld (workloads/)","1.00","1.00" - "idct-4m (workloads/)","16.47","18.59" - "idct-4mw1 (workloads/)","16.49","18.58" - "ippktcheck-4m (workloads/)","278.15","304.77" - "ippktcheck-4mw1 (workloads/)","274.54","304.84" - "ipres-4m (workloads/)","59.43","64.05" - "ipres-4mw1 (workloads/)","58.92","64.29" - "md5-4m (workloads/)","22.84","25.21" - "md5-4mw1 (workloads/)","22.78","25.43" - "rgbcmyk-4m (workloads/)","56.92","63.51" - "rgbcmyk-4mw1 (workloads/)","56.67","63.67" - "rotate-4ms1 (workloads/)","16.71","18.14" - "rotate-4ms1w1 (workloads/)","17.68","18.16" - "rotate-4ms64 (workloads/)","16.86","18.29" - "rotate-4ms64w1 (workloads/)","16.89","18.31" - "x264-4mq (workloads/)","0.50","0.56" - "x264-4mqw1 (workloads/)","0.50","0.56" + :header: "Benchmarks","am62xx_lp_sk-fs: perf","am62xx_sk-fs: perf","am62xxsip_sk-fs: perf" + + "4m-check (workloads/)","281.25","301.64","283.22" + "4m-check-reassembly (workloads/)","62.19","62.93","59.24" + "4m-check-reassembly-tcp (workloads/)","38.88","40.98","38.76" + "4m-check-reassembly-tcp-cmykw2-rotatew2 (workloads/)","22.26","23.69","22.50" + "4m-check-reassembly-tcp-x264w2 (workloads/)","1.61","1.75","1.59" + "4m-cmykw2 (workloads/)","185.01","199.01","181.32" + "4m-cmykw2-rotatew2 (workloads/)","36.17","39.14","35.38" + "4m-reassembly (workloads/)","46.88","49.53","47.57" + "4m-rotatew2 (workloads/)","41.34","44.29","41.08" + "4m-tcp-mixed (workloads/)","94.67","102.56","94.12" + "4m-x264w2 (workloads/)","1.67","1.83","1.64" + "empty-wld (workloads/)","1.00","1.00","1.00" + "idct-4m (workloads/)","16.83","18.59","16.84" + "idct-4mw1 (workloads/)","16.82","18.59","16.84" + "ippktcheck-4m (workloads/)","282.01","302.52","283.42" + "ippktcheck-4mw1 (workloads/)","282.90","299.76","283.45" + "ipres-4m (workloads/)","59.34","64.32","60.90" + "ipres-4mw1 (workloads/)","60.98","64.91","61.37" + "md5-4m (workloads/)","22.20","24.69","22.06" + "md5-4mw1 (workloads/)","22.63","24.76","22.28" + "rgbcmyk-4m (workloads/)","56.95","63.67","56.77" + "rgbcmyk-4mw1 (workloads/)","57.16","63.55","56.71" + "rotate-4ms1 (workloads/)","16.98","18.11","16.91" + "rotate-4ms1w1 (workloads/)","17.00","18.12","16.89" + "rotate-4ms64 (workloads/)","17.19","18.37","17.05" + "rotate-4ms64w1 (workloads/)","17.18","18.27","16.95" + "x264-4mq (workloads/)","0.50","0.56","0.50" + "x264-4mqw1 (workloads/)","0.50","0.56","0.50" + + + + Boot-time Measurement --------------------- + Boot media: MMCSD ^^^^^^^^^^^^^^^^^ .. csv-table:: Linux boot time MMCSD - :header: "Boot Configuration","am62xx_sk-fs: Boot time in seconds: avg(min,max)" + :header: "Boot Configuration","am62xx_sk-fs: Boot time in seconds: avg(min,max)","am62xxsip_sk-fs: Boot time in seconds: avg(min,max)" - "Linux boot time from SD with default rootfs (20 boot cycles)","16.20 (min 15.97, max 16.41)" + "Linux boot time from SD with default rootfs (20 boot cycles)","16.24 (min 15.96, max 16.42)","15.09 (min 14.92, max 15.32)" -.. csv-table:: Linux boot time MMCSD - :header: "Boot Configuration","am62xx_lp_sk-fs: Boot time in seconds: avg(min,max)","am62xx_sk-fs: Boot time in seconds: avg(min,max)" - - "Linux boot time from SD with default rootfs (20 boot cycles)","20.38 (min 20.27, max 20.48)","16.20 (min 15.97, max 16.41)" Boot time numbers [avg, min, max] are measured from "Starting kernel" to Linux prompt across 20 boot cycles. + + | ALSA SoC Audio Driver -------------------------- +--------------------- #. Access type - RW\_INTERLEAVED #. Channels - 2 #. Format - S16\_LE #. Period size - 64 + .. csv-table:: Audio Capture - :header: "Sampling Rate (Hz)","am62xx_lp_sk-fs: Throughput (bits/sec)","am62xx_lp_sk-fs: CPU Load (%)","am62xx_sk-fs: Throughput (bits/sec)","am62xx_sk-fs: CPU Load (%)" - - "8000","255998.00","0.23","256001.00","0.10" - "11025","352797.00","0.21","352801.00","0.19" - "16000","511996.00","0.15","512002.00","0.12" - "22050","705595.00","0.37","705602.00","0.33" - "24000","705593.00","0.34","705596.00","0.31" - "32000","1023988.00","0.23","1023987.00","0.57" - "44100","1411184.00","0.55","1411186.00","0.49" - "48000","1535983.00","0.29","1535988.00","0.24" - "88200","2822364.00","1.07","2822378.00","0.97" - "96000","3071961.00","2.29","3071976.00","0.42" + :header: "Sampling Rate (Hz)","am62xx_lp_sk-fs: Throughput (bits/sec)","am62xx_lp_sk-fs: CPU Load (%)","am62xx_sk-fs: Throughput (bits/sec)","am62xx_sk-fs: CPU Load (%)","am62xxsip_sk-fs: Throughput (bits/sec)","am62xxsip_sk-fs: CPU Load (%)" + + "11025","352798.00","0.12","352799.00","0.12","352799.00","0.28" + "16000","511998.00","0.48","511999.00","0.10","512000.00","0.31" + "22050","705593.00","0.16","705596.00","0.16","705594.00","0.31" + "24000","705595.00","0.19","705598.00","0.19","705597.00","0.39" + "32000","1023993.00","0.11","1023996.00","0.11","1023997.00","0.31" + "44100","1411190.00","0.27","1411195.00","0.25","1411196.00","0.48" + "48000","1535990.00","1.02","1535994.00","0.20","1535996.00","0.38" + "88200","2822374.00","0.51","2822385.00","0.51","2822389.00","0.59" + "96000","3071972.00","0.22","3071984.00","0.19","3071987.00","0.36" + + + + +.. csv-table:: Audio Playback + :header: "Sampling Rate (Hz)","am62xx_lp_sk-fs: Throughput (bits/sec)","am62xx_lp_sk-fs: CPU Load (%)","am62xx_sk-fs: Throughput (bits/sec)","am62xx_sk-fs: CPU Load (%)","am62xxsip_sk-fs: Throughput (bits/sec)","am62xxsip_sk-fs: CPU Load (%)" + + "11025","352945.00","0.13","352946.00","0.11","352948.00","0.10" + "16000","512212.00","0.26","512213.00","0.11","512216.00","0.09" + "22050","705888.00","0.14","705892.00","0.12","705887.00","0.12" + "24000","705890.00","0.17","705893.00","0.14","705891.00","0.14" + "32000","1024421.00","0.13","1024426.00","0.10","1024425.00","0.10" + "44100","1411781.00","0.21","1411787.00","0.19","1411789.00","0.18" + "48000","1536633.00","0.22","1536639.00","0.19","","" + "88200","2823558.00","0.36","2823567.00","0.35","","" + "96000","3073259.00","0.28","3073271.00","0.22","","" + + | Graphics SGX/RGX Driver -------------------------- +----------------------- + + GFXBench -^^^^^^^^^^^^^^^^^^^^^^^^^^^ +^^^^^^^^ Run GFXBench and capture performance reported (Score and Display rate in fps). All display outputs (HDMI, Displayport and/or LCD) are connected when running these tests .. csv-table:: GFXBench Performance :header: "Benchmark","am62xx_lp_sk-fs: Score","am62xx_lp_sk-fs: Fps","am62xx_sk-fs: Score","am62xx_sk-fs: Fps" - " GFXBench 3.x gl_manhattan_off","81.16","1.31" - " GFXBench 3.x gl_trex_off","123.21","2.20" - " GFXBench 5.x gl_5_high_off","11.20","0.17","11.66","0.18" + " GFXBench 3.x gl_manhattan_off","80.18","1.29" + " GFXBench 3.x gl_trex_off","125.77","2.25" + " GFXBench 5.x gl_5_high_off","11.41","0.18","11.48","0.18" + + + Glmark2 -^^^^^^^^^^^^^^^^^^^^^^^^^^^ +^^^^^^^ Run Glmark2 and capture performance reported (Score). All display outputs (HDMI, Displayport and/or LCD) are connected when running these tests .. csv-table:: Glmark2 Performance - :header: "Benchmark","am62xx_lp_sk-fs: Score","am62xx_sk-fs: Score" + :header: "Benchmark","am62xx_lp_sk-fs: Score","am62xx_sk-fs: Score","am62xxsip_sk-fs: Score" + + "Glmark2-DRM","51.00","61.00","54.00" + "Glmark2-Wayland","199.00","217.00","210.00" + "Glmark2-Wayland 1920x1080","63.00","66.00" + - "Glmark2-DRM","51.00","61.00" - "Glmark2-Wayland","203.00","216.00" - "Glmark2-Wayland 1920x1080","63.00","67.00" | Ethernet ------------------ +-------- Ethernet performance benchmarks were measured using Netperf 2.7.1 https://hewlettpackard.github.io/netperf/doc/netperf.html Test procedures were modeled after those defined in RFC-2544: -https://tools.ietf.org/html/rfc2544, where the DUT is the TI device +https://tools.ietf.org/html/rfc2544, where the DUT is the TI device and the "tester" used was a Linux PC. To produce consistent results, -it is recommended to carry out performance tests in a private network and to avoid -running NFS on the same interface used in the test. In these results, +it is recommended to carry out performance tests in a private network and to avoid +running NFS on the same interface used in the test. In these results, CPU utilization was captured as the total percentage used across all cores on the device, -while running the performance test over one external interface. +while running the performance test over one external interface. UDP Throughput (0% loss) was measured by the procedure defined in RFC-2544 section 26.1: Throughput. In this scenario, netperf options burst_size (-b) and wait_time (-w) are used to limit bandwidth -during different trials of the test, with the goal of finding the highest rate at which +during different trials of the test, with the goal of finding the highest rate at which no loss is seen. For example, to limit bandwidth to 500Mbits/sec with 1472B datagram: :: burst_size = / 8 (bits -> bytes) / / 100 (seconds -> 10 ms) - burst_size = 500000000 / 8 / 1472 / 100 = 425 + burst_size = 500000000 / 8 / 1472 / 100 = 425 wait_time = 10 milliseconds (minimum supported by Linux PC used for testing) UDP Throughput (possible loss) was measured by capturing throughput and packet loss statistics when -running the netperf test with no bandwidth limit (remove -b/-w options). +running the netperf test with no bandwidth limit (remove -b/-w options). In order to start a netperf client on one device, the other device must have netserver running. To start netserver: @@ -468,10 +518,10 @@ To start netserver: netserver [-p ] [-4 (IPv4 addressing)] [-6 (IPv6 addressing)] -Running the following shell script from the DUT will trigger netperf clients to measure +Running the following shell script from the DUT will trigger netperf clients to measure bidirectional TCP performance for 60 seconds and report CPU utilization. Parameter -k is used in -client commands to summarize selected statistics on their own line and -j is used to gain -additional timing measurements during the test. +client commands to summarize selected statistics on their own line and -j is used to gain +additional timing measurements during the test. :: @@ -480,316 +530,598 @@ additional timing measurements during the test. do netperf -H -j -c -l 60 -t TCP_STREAM -- -k DIRECTION,THROUGHPUT,MEAN_LATENCY,LOCAL_CPU_UTIL,REMOTE_CPU_UTIL,LOCAL_BYTES_SENT,REMOTE_BYTES_RECVD,LOCAL_SEND_SIZE & - + netperf -H -j -c -l 60 -t TCP_MAERTS -- -k DIRECTION,THROUGHPUT,MEAN_LATENCY,LOCAL_CPU_UTIL,REMOTE_CPU_UTIL,LOCAL_BYTES_SENT,REMOTE_BYTES_RECVD,LOCAL_SEND_SIZE & done -Running the following commands will trigger netperf clients to measure UDP burst performance for -60 seconds at various burst/datagram sizes and report CPU utilization. +Running the following commands will trigger netperf clients to measure UDP burst performance for +60 seconds at various burst/datagram sizes and report CPU utilization. -- For UDP egress tests, run netperf client from DUT and start netserver on tester. +- For UDP egress tests, run netperf client from DUT and start netserver on tester. :: - netperf -H -j -c -l 60 -t UDP_STREAM -b -w -- -m - -k DIRECTION,THROUGHPUT,MEAN_LATENCY,LOCAL_CPU_UTIL,REMOTE_CPU_UTIL,LOCAL_BYTES_SENT,REMOTE_BYTES_RECVD,LOCAL_SEND_SIZE + netperf -H -j -c -l 60 -t UDP_STREAM -b -w -- -m + -k DIRECTION,THROUGHPUT,MEAN_LATENCY,LOCAL_CPU_UTIL,REMOTE_CPU_UTIL,LOCAL_BYTES_SENT,REMOTE_BYTES_RECVD,LOCAL_SEND_SIZE -- For UDP ingress tests, run netperf client from tester and start netserver on DUT. +- For UDP ingress tests, run netperf client from tester and start netserver on DUT. :: netperf -H -j -C -l 60 -t UDP_STREAM -b -w -- -m - -k DIRECTION,THROUGHPUT,MEAN_LATENCY,LOCAL_CPU_UTIL,REMOTE_CPU_UTIL,LOCAL_BYTES_SENT,REMOTE_BYTES_RECVD,LOCAL_SEND_SIZE + -k DIRECTION,THROUGHPUT,MEAN_LATENCY,LOCAL_CPU_UTIL,REMOTE_CPU_UTIL,LOCAL_BYTES_SENT,REMOTE_BYTES_RECVD,LOCAL_SEND_SIZE -CPSW/CPSW2g/CPSW3g Ethernet Driver + +CPSW/CPSW2g/CPSW3g Ethernet Driver ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ -- CPSW3g: AM62x +- CPSW3g: AM64x, AM62x, AM62ax, AM62px + -.. rubric:: TCP Bidirectional Throughput +.. rubric:: TCP Bidirectional Throughput :name: CPSW2g-tcp-bidirectional-throughput .. csv-table:: CPSW2g TCP Bidirectional Throughput - :header: "Command Used","am62xx_lp_sk-fs: THROUGHPUT (Mbits/sec)","am62xx_lp_sk-fs: CPU Load % (LOCAL_CPU_UTIL)","am62xx_sk-fs: THROUGHPUT (Mbits/sec)","am62xx_sk-fs: CPU Load % (LOCAL_CPU_UTIL)" + :header: "Command Used","am62xx_lp_sk-fs: THROUGHPUT (Mbits/sec)","am62xx_lp_sk-fs: CPU Load % (LOCAL_CPU_UTIL)","am62xx_sk-fs: THROUGHPUT (Mbits/sec)","am62xx_sk-fs: CPU Load % (LOCAL_CPU_UTIL)","am62xxsip_sk-fs: THROUGHPUT (Mbits/sec)","am62xxsip_sk-fs: CPU Load % (LOCAL_CPU_UTIL)" + + "netperf -H 192.168.0.1 -j -c -C -l 60 -t TCP_STREAM; netperf -H 192.168.0.1 -j -c -C -l 60 -t TCP_MAERTS","1685.26","64.99","1767.54","66.70","1647.55","62.58" + + - "netperf -H 192.168.0.1 -j -c -C -l 60 -t TCP_STREAM; netperf -H 192.168.0.1 -j -c -C -l 60 -t TCP_MAERTS","1808.53","71.31","1830.85","70.45" .. rubric:: TCP Bidirectional Throughput Interrupt Pacing :name: CPSW2g-tcp-bidirectional-throughput-interrupt-pacing .. csv-table:: CPSW2g TCP Bidirectional Throughput Interrupt Pacing - :header: "Command Used","am62xx_lp_sk-fs: THROUGHPUT (Mbits/sec)","am62xx_lp_sk-fs: CPU Load % (LOCAL_CPU_UTIL)","am62xx_sk-fs: THROUGHPUT (Mbits/sec)","am62xx_sk-fs: CPU Load % (LOCAL_CPU_UTIL)" + :header: "Command Used","am62xx_lp_sk-fs: THROUGHPUT (Mbits/sec)","am62xx_lp_sk-fs: CPU Load % (LOCAL_CPU_UTIL)","am62xx_sk-fs: THROUGHPUT (Mbits/sec)","am62xx_sk-fs: CPU Load % (LOCAL_CPU_UTIL)","am62xxsip_sk-fs: THROUGHPUT (Mbits/sec)","am62xxsip_sk-fs: CPU Load % (LOCAL_CPU_UTIL)" + + "netperf -H 192.168.0.1 -j -c -C -l 60 -t TCP_STREAM; netperf -H 192.168.0.1 -j -c -C -l 60 -t TCP_MAERTS","326.74","11.98","1586.14","47.14","1807.73","48.75" + + - "netperf -H 192.168.0.1 -j -c -C -l 60 -t TCP_STREAM; netperf -H 192.168.0.1 -j -c -C -l 60 -t TCP_MAERTS","1423.79","39.02","1824.36","56.36" .. rubric:: UDP Throughput :name: CPSW2g-udp-throughput-0-loss .. csv-table:: CPSW2g UDP Egress Throughput 0 loss - :header: "Frame Size(bytes)","am62xx_lp_sk-fs: UDP Datagram Size(bytes) (LOCAL_SEND_SIZE)","am62xx_lp_sk-fs: THROUGHPUT (Mbits/sec)","am62xx_lp_sk-fs: Packets Per Second (kPPS)","am62xx_lp_sk-fs: CPU Load % (LOCAL_CPU_UTIL)","am62xx_sk-fs: UDP Datagram Size(bytes) (LOCAL_SEND_SIZE)","am62xx_sk-fs: THROUGHPUT (Mbits/sec)","am62xx_sk-fs: Packets Per Second (kPPS)","am62xx_sk-fs: CPU Load % (LOCAL_CPU_UTIL)" + :header: "Frame Size(bytes)","am62xx_lp_sk-fs: UDP Datagram Size(bytes) (LOCAL_SEND_SIZE)","am62xx_lp_sk-fs: THROUGHPUT (Mbits/sec)","am62xx_lp_sk-fs: Packets Per Second (kPPS)","am62xx_lp_sk-fs: CPU Load % (LOCAL_CPU_UTIL)","am62xx_sk-fs: UDP Datagram Size(bytes) (LOCAL_SEND_SIZE)","am62xx_sk-fs: THROUGHPUT (Mbits/sec)","am62xx_sk-fs: Packets Per Second (kPPS)","am62xx_sk-fs: CPU Load % (LOCAL_CPU_UTIL)","am62xxsip_sk-fs: UDP Datagram Size(bytes) (LOCAL_SEND_SIZE)","am62xxsip_sk-fs: THROUGHPUT (Mbits/sec)","am62xxsip_sk-fs: Packets Per Second (kPPS)","am62xxsip_sk-fs: CPU Load % (LOCAL_CPU_UTIL)" + + "64","18.00","40.95","80.00","37.35","18.00","18.74","37.00","10.12","","48.25","94.00","39.37" + "128","82.00","89.57","87.00","38.47","82.00","99.23","97.00","38.69","","97.04","95.00","26.11" + "256","210.00","175.06","85.00","38.18","210.00","198.56","97.00","38.59","","179.37","88.00","38.40" + "1024","978.00","363.71","44.00","21.25","978.00","785.68","96.00","38.33","","708.67","87.00","38.11" + "1518","1472.00","183.19","15.00","10.59","1472.00","770.81","63.00","36.09","","706.43","58.00","35.94" + + - "64","18.00","11.19","78.00","36.85","18.00","14.71","102.00","39.11" - "128","82.00","58.84","90.00","25.07","82.00","19.72","30.00","19.16" - "256","210.00","147.65","88.00","30.25","210.00","162.53","97.00","38.66" - "1024","978.00","183.54","23.00","10.13","978.00","122.05","16.00","7.09" - "1518","1472.00","956.37","81.00","43.48","1472.00","299.10","25.00","9.72" .. csv-table:: CPSW2g UDP Ingress Throughput 0 loss - :header: "Frame Size(bytes)","am62xx_lp_sk-fs: UDP Datagram Size(bytes) (LOCAL_SEND_SIZE)","am62xx_lp_sk-fs: THROUGHPUT (Mbits/sec)","am62xx_lp_sk-fs: Packets Per Second (kPPS)","am62xx_lp_sk-fs: CPU Load % (LOCAL_CPU_UTIL)","am62xx_sk-fs: UDP Datagram Size(bytes) (LOCAL_SEND_SIZE)","am62xx_sk-fs: THROUGHPUT (Mbits/sec)","am62xx_sk-fs: Packets Per Second (kPPS)","am62xx_sk-fs: CPU Load % (LOCAL_CPU_UTIL)" + :header: "Frame Size(bytes)","am62xx_lp_sk-fs: UDP Datagram Size(bytes) (LOCAL_SEND_SIZE)","am62xx_lp_sk-fs: THROUGHPUT (Mbits/sec)","am62xx_lp_sk-fs: Packets Per Second (kPPS)","am62xx_lp_sk-fs: CPU Load % (LOCAL_CPU_UTIL)","am62xx_sk-fs: UDP Datagram Size(bytes) (LOCAL_SEND_SIZE)","am62xx_sk-fs: THROUGHPUT (Mbits/sec)","am62xx_sk-fs: Packets Per Second (kPPS)","am62xx_sk-fs: CPU Load % (LOCAL_CPU_UTIL)","am62xxsip_sk-fs: UDP Datagram Size(bytes) (LOCAL_SEND_SIZE)","am62xxsip_sk-fs: THROUGHPUT (Mbits/sec)","am62xxsip_sk-fs: Packets Per Second (kPPS)","am62xxsip_sk-fs: CPU Load % (LOCAL_CPU_UTIL)" + + "64","","3.38","7.00","1.44","","3.07","6.00","3.06","","2.92","6.00","0.97" + "128","82.00","6.76","7.00","1.28","82.00","7.58","7.00","2.11","","7.17","7.00","1.63" + "256","210.00","15.77","8.00","4.60","210.00","15.36","8.00","2.22","","15.16","7.00","2.19" + "1024","978.00","60.62","7.00","2.80","978.00","61.44","8.00","1.82","","61.44","8.00","4.89" + "1518","1472.00","38.86","3.00","2.09","1472.00","2.43","0.00","0.59" + + - "64","18.00","1.04","7.00","0.94" - "128","82.00","5.25","8.00","1.46","82.00","3.94","6.00","2.87" - "256","210.00","13.94","8.00","1.68","210.00","13.61","8.00","1.80" - "1024","978.00","936.44","120.00","42.20","978.00","65.72","8.00","2.08" - "1518","1472.00","98.92","8.00","5.88","1472.00","98.92","8.00","2.97" .. csv-table:: CPSW2g UDP Ingress Throughput possible loss - :header: "Frame Size(bytes)","am62xx_lp_sk-fs: UDP Datagram Size(bytes) (LOCAL_SEND_SIZE)","am62xx_lp_sk-fs: THROUGHPUT (Mbits/sec)","am62xx_lp_sk-fs: Packets Per Second (kPPS)","am62xx_lp_sk-fs: CPU Load % (LOCAL_CPU_UTIL)","am62xx_lp_sk-fs: Packet Loss %","am62xx_sk-fs: UDP Datagram Size(bytes) (LOCAL_SEND_SIZE)","am62xx_sk-fs: THROUGHPUT (Mbits/sec)","am62xx_sk-fs: Packets Per Second (kPPS)","am62xx_sk-fs: CPU Load % (LOCAL_CPU_UTIL)","am62xx_sk-fs: Packet Loss %" + :header: "Frame Size(bytes)","am62xx_lp_sk-fs: UDP Datagram Size(bytes) (LOCAL_SEND_SIZE)","am62xx_lp_sk-fs: THROUGHPUT (Mbits/sec)","am62xx_lp_sk-fs: Packets Per Second (kPPS)","am62xx_lp_sk-fs: CPU Load % (LOCAL_CPU_UTIL)","am62xx_lp_sk-fs: Packet Loss %","am62xx_sk-fs: UDP Datagram Size(bytes) (LOCAL_SEND_SIZE)","am62xx_sk-fs: THROUGHPUT (Mbits/sec)","am62xx_sk-fs: Packets Per Second (kPPS)","am62xx_sk-fs: CPU Load % (LOCAL_CPU_UTIL)","am62xx_sk-fs: Packet Loss %","am62xxsip_sk-fs: UDP Datagram Size(bytes) (LOCAL_SEND_SIZE)","am62xxsip_sk-fs: THROUGHPUT (Mbits/sec)","am62xxsip_sk-fs: Packets Per Second (kPPS)","am62xxsip_sk-fs: CPU Load % (LOCAL_CPU_UTIL)","am62xxsip_sk-fs: Packet Loss %" + + "64","","71.78","140.00","65.81","74.13","","97.99","191.00","40.57","61.06","","92.95","182.00","42.37","75.92" + "128","82.00","173.15","169.00","45.12","65.86","82.00","193.99","189.00","41.95","67.47","","183.40","179.00","43.96","71.84" + "256","210.00","271.82","133.00","41.55","0.42","210.00","357.95","175.00","41.46","52.51","","324.57","158.00","43.50","58.28" + "1024","978.00","939.00","115.00","44.49","0.03","978.00","904.64","110.00","42.08","0.38","","738.20","90.00","41.83","0.55" + "1518","1472.00","13.16","1.00","20.85","97.98","1472.00","3.90","0.00","24.85","99.58" + - "64","18.00","33.17","230.00","45.90","73.39" - "128","82.00","113.43","173.00","45.16","66.03","82.00","125.24","191.00","44.79","37.76" - "256","210.00","277.06","165.00","44.13","60.85","210.00","317.12","189.00","43.41","39.20" - "1024","978.00","936.44","120.00","42.20","0.00","978.00","932.18","119.00","41.73","0.09" - "1518","1472.00","916.84","78.00","42.74","0.53","1472.00","922.97","78.00","41.03","0.34" | Linux OSPI Flash Driver -------------------------- +----------------------- + AM62XX-SK -^^^^^^^^^^^^^^^^^^^^^^^^^^^ +^^^^^^^^^ + RAW -""""""""""""""""""""""""""" +""" .. csv-table:: OSPI Raw Flash Driver :header: "File size (Mbytes)","am62xx_sk-fs: Raw Read Throughput (Mbytes/sec)" - "50","192.31" + "50","172.41" + .. csv-table:: OSPI Raw Flash Driver :header: "File size (Mbytes)","am62xx_lp_sk-fs: Raw Read Throughput (Mbytes/sec)" - "50","29.24" - -| - -UBoot QSPI/OSPI Driver -------------------------- - -AM62XX-SK -^^^^^^^^^^^^^^^^^^^^^^^^^^^ + "50","29.07" -.. csv-table:: UBOOT QSPI or OSPI - :header: "File size (bytes in hex)","am62xx_sk-fs: Write Throughput (Kbytes/sec)","am62xx_sk-fs: Read Throughput (Kbytes/sec)" - - "400000","391.66","195047.62" - "800000","391.27","248242.42" - "1000000","392.73","273066.67" - "2000000","390.59","284939.13" EMMC Driver ----------- -.. important:: +.. warning:: - The performance numbers can be severely affected if the media is + **IMPORTANT**: The performance numbers can be severely affected if the media is mounted in sync mode. Hot plug scripts in the filesystem mount removable media in sync mode to ensure data integrity. For performance sensitive applications, umount the auto-mounted filesystem and re-mount in async mode. + + EMMC EXT4 FIO 1G ^^^^^^^^^^^^^^^^ + .. csv-table:: EMMC EXT4 FIO 1G :header: "Buffer size (bytes)","am62xx_sk-fs: Write EXT4 Throughput (Mbytes/sec)","am62xx_sk-fs: Write EXT4 CPU Load (%)","am62xx_sk-fs: Read EXT4 Throughput (Mbytes/sec)","am62xx_sk-fs: Read EXT4 CPU Load (%)" - "1m","59.80","0.08","175.00","0.08" - "4m","60.00","0.10","175.00","0.06" - "4k","53.00","0.06","56.40","0.05" - "256k","60.30","0.05","174.00","0.05" + "1m","44.60","1.43","174.00","2.30" + "4m","44.90","1.06","174.00","1.70" + "4k","5.32","2.19","36.40","10.09" + "256k","35.40","1.49","173.00","2.84" + + .. csv-table:: EMMC EXT4 FIO 1G :header: "Buffer size (bytes)","am62xx_lp_sk-fs: Write EXT4 Throughput (Mbytes/sec)","am62xx_lp_sk-fs: Write EXT4 CPU Load (%)","am62xx_lp_sk-fs: Read EXT4 Throughput (Mbytes/sec)","am62xx_lp_sk-fs: Read EXT4 CPU Load (%)" - "1m","61.60","2.01","175.00","2.53" - "4m","61.60","1.43","175.00","1.97" - "4k","51.40","19.40","56.80","16.84" - "256k","61.60","2.34","174.00","3.36" + "1m","61.50","1.99","175.00","2.30" + "4m","61.70","1.33","175.00","1.97" + "4k","51.40","19.45","56.80","16.71" + "256k","61.60","2.33","174.00","3.23" + + EMMC EXT4 ^^^^^^^^^ + .. csv-table:: EMMC EXT4 :header: "Buffer size (bytes)","am62xx_sk-fs: Write EXT4 Throughput (Mbytes/sec)","am62xx_sk-fs: Write EXT4 CPU Load (%)","am62xx_sk-fs: Read EXT4 Throughput (Mbytes/sec)","am62xx_sk-fs: Read EXT4 CPU Load (%)" - "102400","54.37 (min 51.87, max 55.19)","4.84 (min 4.31, max 5.84)","179.35","10.78" - "262144","53.97 (min 51.33, max 54.81)","4.82 (min 4.31, max 5.77)","182.78","12.39" - "524288","53.43 (min 50.44, max 54.44)","4.74 (min 4.38, max 5.90)","182.99","14.47" - "1048576","53.56 (min 51.51, max 54.38)","4.71 (min 4.02, max 6.14)","182.93","12.33" - "5242880","53.84 (min 50.47, max 54.90)","4.55 (min 4.05, max 5.55)","182.73","12.72" + "102400","51.83 (min 49.52, max 52.61)","5.03 (min 4.49, max 6.52)","177.99","13.56" + "262144","50.97 (min 48.59, max 51.96)","5.21 (min 4.57, max 6.73)","181.94","14.16" + "524288","50.87 (min 48.19, max 51.68)","5.79 (min 4.94, max 7.10)","182.94","15.35" + "1048576","51.33 (min 48.92, max 52.08)","5.32 (min 4.73, max 6.67)","182.83","15.72" + "5242880","51.15 (min 49.02, max 51.80)","5.18 (min 4.59, max 6.56)","181.99","15.65" + + .. csv-table:: EMMC EXT4 :header: "Buffer size (bytes)","am62xx_lp_sk-fs: Write EXT4 Throughput (Mbytes/sec)","am62xx_lp_sk-fs: Write EXT4 CPU Load (%)","am62xx_lp_sk-fs: Read EXT4 Throughput (Mbytes/sec)","am62xx_lp_sk-fs: Read EXT4 CPU Load (%)" - "102400","50.36 (min 47.80, max 51.29)","5.49 (min 4.78, max 6.86)","177.03","13.68" - "262144","49.57 (min 47.01, max 50.44)","5.60 (min 5.19, max 6.88)","180.97","16.23" - "524288","49.52 (min 46.68, max 50.44)","5.76 (min 5.29, max 7.23)","182.44","17.11" - "1048576","49.57 (min 47.05, max 50.43)","5.87 (min 5.39, max 7.29)","182.30","17.03" - "5242880","49.85 (min 47.38, max 50.51)","5.64 (min 5.07, max 6.91)","181.88","17.39" + "102400","51.16 (min 48.23, max 52.47)","5.88 (min 5.39, max 7.27)","176.55","13.73" + "262144","49.96 (min 47.07, max 50.94)","5.98 (min 5.45, max 7.53)","180.51","15.04" + "524288","49.88 (min 47.07, max 50.91)","6.06 (min 5.47, max 7.62)","182.72","17.03" + "1048576","50.54 (min 47.88, max 51.64)","5.91 (min 5.45, max 7.31)","182.53","16.81" + "5242880","50.27 (min 47.13, max 51.35)","5.73 (min 5.24, max 7.20)","181.68","17.11" + + + + +EMMC EXT2 +^^^^^^^^^ + + +.. csv-table:: EMMC EXT2 + :header: "Buffer size (bytes)","am62xx_sk-fs: Write EXT2 Throughput (Mbytes/sec)","am62xx_sk-fs: Write EXT2 CPU Load (%)","am62xx_sk-fs: Read EXT2 Throughput (Mbytes/sec)","am62xx_sk-fs: Read EXT2 CPU Load (%)" + + "102400","39.34 (min 35.40, max 40.46)","4.49 (min 3.94, max 6.48)","174.99","12.45" + "262144","40.53 (min 36.16, max 41.74)","4.57 (min 3.80, max 6.57)","177.47","15.45" + "524288","40.11 (min 35.88, max 41.23)","4.54 (min 3.93, max 6.59)","178.47","14.59" + "1048576","40.30 (min 36.06, max 41.48)","4.49 (min 3.86, max 6.46)","177.97","14.66" + "5242880","40.32 (min 35.99, max 41.47)","4.52 (min 3.85, max 6.69)","177.38","14.96" + + + +.. csv-table:: EMMC EXT2 + :header: "Buffer size (bytes)","am62xx_lp_sk-fs: Write EXT2 Throughput (Mbytes/sec)","am62xx_lp_sk-fs: Write EXT2 CPU Load (%)","am62xx_lp_sk-fs: Read EXT2 Throughput (Mbytes/sec)","am62xx_lp_sk-fs: Read EXT2 CPU Load (%)" + + "102400","50.92 (min 43.73, max 52.79)","6.83 (min 5.89, max 9.41)","173.27","13.56" + "262144","50.82 (min 43.46, max 52.86)","6.61 (min 5.74, max 9.33)","177.56","16.88" + "524288","50.63 (min 43.23, max 52.92)","6.74 (min 5.62, max 9.18)","176.84","16.74" + "1048576","50.67 (min 43.58, max 52.56)","6.62 (min 5.64, max 9.44)","177.34","17.17" + "5242880","50.70 (min 42.92, max 53.20)","6.46 (min 5.52, max 9.29)","177.49","17.45" + + + EMMC VFAT ^^^^^^^^^ -.. csv-table:: EMMC EXT4 - :header: "Buffer size (bytes)","am62xx_sk-fs: Write EXT4 Throughput (Mbytes/sec)","am62xx_sk-fs: Write EXT4 CPU Load (%)","am62xx_sk-fs: Read EXT4 Throughput (Mbytes/sec)","am62xx_sk-fs: Read EXT4 CPU Load (%)" - "102400","54.37 (min 51.87, max 55.19)","4.84 (min 4.31, max 5.84)","179.35","10.78" - "262144","53.97 (min 51.33, max 54.81)","4.82 (min 4.31, max 5.77)","182.78","12.39" - "524288","53.43 (min 50.44, max 54.44)","4.74 (min 4.38, max 5.90)","182.99","14.47" - "1048576","53.56 (min 51.51, max 54.38)","4.71 (min 4.02, max 6.14)","182.93","12.33" - "5242880","53.84 (min 50.47, max 54.90)","4.55 (min 4.05, max 5.55)","182.73","12.72" +.. csv-table:: EMMC VFAT + :header: "Buffer size (bytes)","am62xx_sk-fs: Write VFAT Throughput (Mbytes/sec)","am62xx_sk-fs: Write VFAT CPU Load (%)","am62xx_sk-fs: Read VFAT Throughput (Mbytes/sec)","am62xx_sk-fs: Read VFAT CPU Load (%)" + + "102400","47.06 (min 41.03, max 48.78)","6.33 (min 5.66, max 8.51)","169.15","13.77" + "262144","48.92 (min 42.39, max 50.79)","6.86 (min 6.18, max 9.11)","169.46","15.16" + "524288","48.17 (min 41.79, max 49.95)","6.59 (min 5.95, max 8.88)","169.19","15.32" + "1048576","48.61 (min 42.40, max 50.41)","6.61 (min 5.84, max 9.10)","168.56","14.29" + "5242880","48.94 (min 42.48, max 50.83)","6.65 (min 5.99, max 8.93)","168.66","14.98" + + + + +.. csv-table:: EMMC VFAT + :header: "Buffer size (bytes)","am62xx_lp_sk-fs: Write VFAT Throughput (Mbytes/sec)","am62xx_lp_sk-fs: Write VFAT CPU Load (%)","am62xx_lp_sk-fs: Read VFAT Throughput (Mbytes/sec)","am62xx_lp_sk-fs: Read VFAT CPU Load (%)" + + "102400","46.05 (min 39.74, max 47.77)","6.88 (min 5.82, max 9.39)","168.34","14.29" + "262144","47.61 (min 40.82, max 49.71)","7.33 (min 6.35, max 10.21)","169.78","17.89" + "524288","47.73 (min 40.99, max 49.58)","7.30 (min 6.47, max 9.78)","168.38","18.25" + "1048576","47.70 (min 41.07, max 49.70)","7.31 (min 6.53, max 9.81)","168.33","17.67" + "5242880","47.80 (min 40.93, max 49.80)","7.16 (min 6.26, max 9.67)","167.94","16.87" + -.. csv-table:: EMMC EXT4 - :header: "Buffer size (bytes)","am62xx_lp_sk-fs: Write EXT4 Throughput (Mbytes/sec)","am62xx_lp_sk-fs: Write EXT4 CPU Load (%)","am62xx_lp_sk-fs: Read EXT4 Throughput (Mbytes/sec)","am62xx_lp_sk-fs: Read EXT4 CPU Load (%)" - "102400","50.36 (min 47.80, max 51.29)","5.49 (min 4.78, max 6.86)","177.03","13.68" - "262144","49.57 (min 47.01, max 50.44)","5.60 (min 5.19, max 6.88)","180.97","16.23" - "524288","49.52 (min 46.68, max 50.44)","5.76 (min 5.29, max 7.23)","182.44","17.11" - "1048576","49.57 (min 47.05, max 50.43)","5.87 (min 5.39, max 7.29)","182.30","17.03" - "5242880","49.85 (min 47.38, max 50.51)","5.64 (min 5.07, max 6.91)","181.88","17.39" UBoot EMMC Driver ----------------- + .. csv-table:: UBOOT EMMC RAW :header: "File size (bytes in hex)","am62xx_sk-fs: Write Throughput (Kbytes/sec)","am62xx_sk-fs: Read Throughput (Kbytes/sec)" - "2000000","59254.97","170666.67" - "4000000","59201.45","174297.87" + "2000000","99296.97","171560.21" + "4000000","97523.81","175229.95" + + + .. csv-table:: UBOOT EMMC RAW :header: "File size (bytes in hex)","am62xx_lp_sk-fs: Write Throughput (Kbytes/sec)","am62xx_lp_sk-fs: Read Throughput (Kbytes/sec)" - "2000000","60457.56","171560.21" - "4000000","61536.15","174297.87" + "2000000","57588.75","169782.38" + "4000000","57893.99","173375.66" + + +.. csv-table:: UBOOT EMMC RAW + :header: "File size (bytes in hex)","am62xxsip_sk-fs: Write Throughput (Kbytes/sec)","am62xxsip_sk-fs: Read Throughput (Kbytes/sec)" + + "2000000","47627.91","91275.77" + "4000000","52261.56","92434.41" + + MMCSD ----- -.. important:: +.. warning:: - The performance numbers can be severely affected if the media is + **IMPORTANT**: The performance numbers can be severely affected if the media is mounted in sync mode. Hot plug scripts in the filesystem mount removable media in sync mode to ensure data integrity. For performance sensitive applications, umount the auto-mounted filesystem and re-mount in async mode. + MMC EXT4 FIO 1G ^^^^^^^^^^^^^^^ + .. csv-table:: MMC EXT4 FIO 1G :header: "Buffer size (bytes)","am62xx_sk-fs: Write EXT4 Throughput (Mbytes/sec)","am62xx_sk-fs: Write EXT4 CPU Load (%)","am62xx_sk-fs: Read EXT4 Throughput (Mbytes/sec)","am62xx_sk-fs: Read EXT4 CPU Load (%)" - "1m","43.60","1.21","87.30","1.22" - "4m","42.30","0.79","87.10","0.93" - "4k","2.83","1.64","12.90","4.19" - "256k","39.70","1.44","83.60","1.65" + "1m","42.70","1.44","87.30","1.53" + "4m","41.60","1.07","87.20","1.24" + "4k","2.77","1.78","13.00","4.45" + "256k","37.60","1.72","83.70","1.79" + + + .. csv-table:: MMC EXT4 FIO 1G :header: "Buffer size (bytes)","am62xx_lp_sk-fs: Write EXT4 Throughput (Mbytes/sec)","am62xx_lp_sk-fs: Write EXT4 CPU Load (%)","am62xx_lp_sk-fs: Read EXT4 Throughput (Mbytes/sec)","am62xx_lp_sk-fs: Read EXT4 CPU Load (%)" - "1m","42.00","1.55","87.20","1.89" - "4m","43.00","1.14","87.30","1.18" - "4k","2.78","2.00","13.00","4.94" - "256k","38.40","1.82","83.50","2.02" + "1m","42.60","1.57","87.20","1.56" + "4m","42.20","1.07","87.30","1.13" + "4k","2.77","1.99","12.90","5.04" + "256k","38.60","1.70","84.00","1.88" + + + + +.. csv-table:: MMC EXT4 FIO 1G + :header: "Buffer size (bytes)","am62xxsip_sk-fs: Write EXT4 Throughput (Mbytes/sec)","am62xxsip_sk-fs: Write EXT4 CPU Load (%)","am62xxsip_sk-fs: Read EXT4 Throughput (Mbytes/sec)","am62xxsip_sk-fs: Read EXT4 CPU Load (%)" + + "1m","42.40","1.97","87.30","1.86" + "4m","40.70","1.40","87.00","1.79" + "4k","2.77","2.14","13.20","4.98" + "256k","37.80","1.90","83.40","1.90" + MMC EXT4 ^^^^^^^^ + .. csv-table:: MMC EXT4 :header: "Buffer size (bytes)","am62xx_sk-fs: Write Raw Throughput (Mbytes/sec)","am62xx_sk-fs: Write Raw CPU Load (%)","am62xx_sk-fs: Read Raw Throughput (Mbytes/sec)","am62xx_sk-fs: Read Raw CPU Load (%)" - "102400","10.63 (min 10.54, max 10.67)","0.98 (min 0.91, max 1.18)","10.71","0.95" - "262144","10.31 (min 10.23, max 10.42)","1.05 (min 0.91, max 1.37)","11.22","0.96" - "524288","10.35 (min 10.22, max 10.49)","0.99 (min 0.86, max 1.25)","11.39","0.90" - "1048576","10.36 (min 10.29, max 10.51)","1.04 (min 0.84, max 1.40)","11.07","0.77" - "5242880","10.77 (min 10.21, max 11.15)","1.00 (min 0.85, max 1.31)","11.97","0.83" + "102400","10.57 (min 10.49, max 10.60)","1.19 (min 1.06, max 1.45)","11.06","1.08" + "262144","10.30 (min 9.97, max 10.61)","1.24 (min 1.10, max 1.54)","10.78","1.03" + "524288","10.31 (min 10.18, max 10.46)","1.18 (min 1.02, max 1.55)","11.44","1.01" + "1048576","10.35 (min 10.23, max 10.63)","1.22 (min 1.03, max 1.52)","11.44","1.12" + "5242880","10.55 (min 10.41, max 10.77)","1.20 (min 1.09, max 1.54)","10.95","1.05" + + + + +.. csv-table:: MMC EXT4 + :header: "Buffer size (bytes)","am62xx_lp_sk-fs: Write Raw Throughput (Mbytes/sec)","am62xx_lp_sk-fs: Write Raw CPU Load (%)","am62xx_lp_sk-fs: Read Raw Throughput (Mbytes/sec)","am62xx_lp_sk-fs: Read Raw CPU Load (%)" + + "102400","10.58 (min 10.19, max 10.92)","1.36 (min 1.22, max 1.63)","11.78","1.10" + "262144","10.12 (min 10.10, max 10.15)","1.42 (min 1.23, max 1.76)","10.95","1.18" + "524288","10.17 (min 10.10, max 10.23)","1.37 (min 1.17, max 1.66)","11.06","1.19" + "1048576","10.20 (min 10.14, max 10.30)","1.38 (min 1.21, max 1.70)","11.05","1.21" + "5242880","10.19 (min 10.16, max 10.28)","1.32 (min 1.14, max 1.81)","11.52","1.13" + + + .. csv-table:: MMC EXT4 + :header: "Buffer size (bytes)","am62xxsip_sk-fs: Write Raw Throughput (Mbytes/sec)","am62xxsip_sk-fs: Write Raw CPU Load (%)","am62xxsip_sk-fs: Read Raw Throughput (Mbytes/sec)","am62xxsip_sk-fs: Read Raw CPU Load (%)" + + "102400","10.51 (min 10.47, max 10.57)","1.40 (min 1.06, max 2.18)","10.87","1.19" + "262144","10.59 (min 10.55, max 10.63)","1.13 (min 0.99, max 1.49)","11.19","1.15" + "524288","10.71 (min 10.54, max 10.82)","1.02 (min 0.92, max 1.16)","11.01","1.08" + "1048576","10.61 (min 10.25, max 10.81)","1.16 (min 1.11, max 1.33)","10.83","1.16" + "5242880","11.04 (min 10.55, max 11.30)","1.11 (min 0.95, max 1.29)","12.02","1.12" + + +MMC EXT3 +^^^^^^^^ + + +.. csv-table:: MMC EXT3 + :header: "Buffer size (bytes)","am62xx_sk-fs: Write Raw Throughput (Mbytes/sec)","am62xx_sk-fs: Write Raw CPU Load (%)","am62xx_sk-fs: Read Raw Throughput (Mbytes/sec)","am62xx_sk-fs: Read Raw CPU Load (%)" + + "102400","36.77 (min 32.06, max 38.32)","4.58 (min 3.89, max 6.66)","79.49","6.11" + "262144","34.68 (min 31.54, max 35.94)","4.19 (min 3.63, max 6.25)","85.12","7.36" + "524288","34.44 (min 29.06, max 36.66)","4.13 (min 3.68, max 5.76)","89.51","7.51" + "1048576","34.73 (min 30.73, max 36.01)","4.02 (min 3.28, max 5.95)","89.59","7.73" + "5242880","38.45 (min 34.26, max 40.01)","4.49 (min 3.77, max 6.55)","84.49","8.22" + + + + +.. csv-table:: MMC EXT3 :header: "Buffer size (bytes)","am62xx_lp_sk-fs: Write Raw Throughput (Mbytes/sec)","am62xx_lp_sk-fs: Write Raw CPU Load (%)","am62xx_lp_sk-fs: Read Raw Throughput (Mbytes/sec)","am62xx_lp_sk-fs: Read Raw CPU Load (%)" - "102400","10.55 (min 10.43, max 10.61)","1.43 (min 1.28, max 1.84)","10.89","1.22" - "262144","10.26 (min 10.15, max 10.57)","1.44 (min 1.24, max 1.79)","10.80","1.24" - "524288","10.24 (min 10.00, max 10.50)","1.41 (min 1.17, max 1.85)","11.03","1.26" - "1048576","10.28 (min 10.09, max 10.45)","1.48 (min 1.28, max 1.60)","11.28","1.35" - "5242880","10.52 (min 10.32, max 10.75)","1.40 (min 1.22, max 1.94)","11.97","1.28" + "102400","35.99 (min 31.77, max 37.53)","5.11 (min 4.33, max 7.44)","79.52","6.86" + "262144","33.79 (min 30.12, max 34.90)","4.95 (min 3.92, max 7.10)","84.64","9.05" + "524288","34.60 (min 30.25, max 35.84)","4.67 (min 3.92, max 7.00)","89.40","8.76" + "1048576","34.20 (min 29.86, max 35.48)","4.64 (min 3.71, max 6.63)","84.61","8.69" + "5242880","34.30 (min 30.30, max 35.96)","4.52 (min 3.81, max 6.73)","88.98","8.53" + + + + +.. csv-table:: MMC EXT3 + :header: "Buffer size (bytes)","am62xxsip_sk-fs: Write Raw Throughput (Mbytes/sec)","am62xxsip_sk-fs: Write Raw CPU Load (%)","am62xxsip_sk-fs: Read Raw Throughput (Mbytes/sec)","am62xxsip_sk-fs: Read Raw CPU Load (%)" + + "102400","35.61 (min 27.97, max 40.71)","5.54 (min 2.82, max 9.13)","79.32","6.99" + "262144","41.13 (min 40.19, max 42.21)","5.30 (min 2.86, max 8.60)","85.39","7.99" + "524288","42.68 (min 41.84, max 43.27)","4.09 (min 2.83, max 7.93)","84.63","7.74" + "1048576","42.30 (min 40.04, max 43.70)","3.76 (min 2.77, max 7.16)","89.59","8.72" + "5242880","42.01 (min 41.13, max 43.09)","3.97 (min 2.62, max 8.43)","89.58","8.33" + + + +MMC EXT2 +^^^^^^^^ + + +.. csv-table:: MMC EXT2 + :header: "Buffer size (bytes)","am62xx_sk-fs: Write Raw Throughput (Mbytes/sec)","am62xx_sk-fs: Write Raw CPU Load (%)","am62xx_sk-fs: Read Raw Throughput (Mbytes/sec)","am62xx_sk-fs: Read Raw CPU Load (%)" + + "102400","38.68 (min 34.18, max 40.13)","4.40 (min 3.65, max 6.69)","80.23","6.53" + "262144","35.29 (min 31.27, max 36.87)","4.22 (min 3.38, max 5.82)","85.76","7.80" + "524288","35.20 (min 32.48, max 36.36)","4.16 (min 3.34, max 6.66)","89.26","7.68" + "1048576","34.82 (min 31.53, max 37.17)","4.12 (min 3.34, max 6.15)","84.22","7.47" + "5242880","39.27 (min 35.85, max 40.49)","4.39 (min 3.68, max 6.50)","89.89","7.73" + + + + +.. csv-table:: MMC EXT2 + :header: "Buffer size (bytes)","am62xx_lp_sk-fs: Write Raw Throughput (Mbytes/sec)","am62xx_lp_sk-fs: Write Raw CPU Load (%)","am62xx_lp_sk-fs: Read Raw Throughput (Mbytes/sec)","am62xx_lp_sk-fs: Read Raw CPU Load (%)" + + "102400","38.00 (min 33.58, max 39.35)","4.74 (min 4.13, max 7.07)","75.79","6.22" + "262144","35.16 (min 31.38, max 36.39)","4.55 (min 3.80, max 6.53)","85.71","8.59" + "524288","35.08 (min 31.38, max 36.20)","4.61 (min 3.86, max 6.72)","89.64","8.37" + "1048576","38.47 (min 34.47, max 39.93)","4.85 (min 4.14, max 7.38)","89.35","8.33" + "5242880","35.09 (min 31.53, max 36.31)","4.41 (min 3.64, max 6.68)","89.43","9.15" + + + + +.. csv-table:: MMC EXT2 + :header: "Buffer size (bytes)","am62xxsip_sk-fs: Write Raw Throughput (Mbytes/sec)","am62xxsip_sk-fs: Write Raw CPU Load (%)","am62xxsip_sk-fs: Read Raw Throughput (Mbytes/sec)","am62xxsip_sk-fs: Read Raw CPU Load (%)" + + "102400","36.89 (min 27.05, max 39.82)","4.68 (min 2.56, max 7.97)","76.13","6.58" + "262144","43.94 (min 43.06, max 44.49)","3.91 (min 2.71, max 7.56)","86.00","7.44" + "524288","43.69 (min 41.15, max 45.18)","3.87 (min 3.04, max 6.82)","89.93","8.19" + "1048576","44.00 (min 42.71, max 45.20)","3.85 (min 2.86, max 7.47)","89.56","7.96" + "5242880","42.13 (min 40.07, max 44.07)","3.47 (min 2.70, max 5.73)","89.51","8.14" + + + + The performance numbers were captured using the following: - SanDisk Max Endurance SD card (SDSQQVR-032G-GN6IA) - Partition was mounted with async option - + + UBoot MMCSD ----------- + UBOOT MMCSD FAT ^^^^^^^^^^^^^^^ + .. csv-table:: UBOOT MMCSD FAT :header: "File size (bytes in hex)","am62xx_sk-fs: Write Throughput (Kbytes/sec)","am62xx_sk-fs: Read Throughput (Kbytes/sec)" - "400000","37577.98","81920.00" - "800000","42010.26","87148.94" - "1000000","49498.49","89530.05" + "400000","39009.52","83591.84" + "800000","40554.46","88086.02" + "1000000","44643.05","89530.05" + + + .. csv-table:: UBOOT MMCSD FAT :header: "File size (bytes in hex)","am62xx_lp_sk-fs: Write Throughput (Kbytes/sec)","am62xx_lp_sk-fs: Read Throughput (Kbytes/sec)" - "400000","37577.98","83591.84" - "800000","32899.60","87148.94" - "1000000","49201.20","89530.05" + "400000","35008.55","81920.00" + "800000","32379.45","87148.94" + "1000000","49799.39","90021.98" + + + + +.. csv-table:: UBOOT MMCSD FAT + :header: "File size (bytes in hex)","am62xxsip_sk-fs: Write Throughput (Kbytes/sec)","am62xxsip_sk-fs: Read Throughput (Kbytes/sec)" + + "400000","31507.69","81920.00" + "800000","42890.05","87148.94" + "1000000","49648.48","89530.05" + + The performance numbers were captured using the following: - SanDisk Max Endurance SD card (SDSQQVR-032G-GN6IA) + | USB Driver -------------------------- +---------- + USB Device Controller -^^^^^^^^^^^^^^^^^^^^^^^^^^^ +^^^^^^^^^^^^^^^^^^^^^ + .. csv-table:: USBDEVICE HIGHSPEED SLAVE_READ_THROUGHPUT - :header: "Number of Blocks","am62xx_lp_sk-fs: Throughput (MB/sec)","am62xx_sk-fs: Throughput (MB/sec)" + :header: "Number of Blocks","am62xx_lp_sk-fs: Throughput (MB/sec)","am62xx_sk-fs: Throughput (MB/sec)","am62xxsip_sk-fs: Throughput (MB/sec)" + + "150","33.00","42.50","43.50" + - "150","31.40","42.20" .. csv-table:: USBDEVICE HIGHSPEED SLAVE_WRITE_THROUGHPUT - :header: "Number of Blocks","am62xx_lp_sk-fs: Throughput (MB/sec)","am62xx_sk-fs: Throughput (MB/sec)" + :header: "Number of Blocks","am62xx_lp_sk-fs: Throughput (MB/sec)","am62xx_sk-fs: Throughput (MB/sec)","am62xxsip_sk-fs: Throughput (MB/sec)" + + "150","30.70","39.80","37.10" - "150","27.40","40.60" | CRYPTO Driver -------------------------- +------------- + + +OpenSSL Performance +^^^^^^^^^^^^^^^^^^^ + +.. csv-table:: OpenSSL Performance + :header: "Algorithm","Buffer Size (in bytes)","am62xx_lp_sk-fs: throughput (KBytes/Sec)","am62xx_sk-fs: throughput (KBytes/Sec)","am62xxsip_sk-fs: throughput (KBytes/Sec)" + + "aes-128-cbc","1024","19332.10","21814.61","21901.65" + "aes-128-cbc","16","344.83","380.01","370.25" + "aes-128-cbc","16384","112579.93","116370.09","124261.72" + "aes-128-cbc","256","6121.81","6160.55","6302.89" + "aes-128-cbc","64","1532.95","1542.44","1626.22" + "aes-128-cbc","8192","84322.99","88956.93","93115.73" + "aes-128-ecb","1024","20105.56","22847.49","22268.59" + "aes-128-ecb","16","362.49","383.55","378.92" + "aes-128-ecb","16384","114682.54","120569.86","127860.74" + "aes-128-ecb","256","6092.37","6171.05","6463.66" + "aes-128-ecb","64","1538.13","1548.27","1663.08" + "aes-128-ecb","8192","87067.31","92536.83","95701.67" + "aes-192-cbc","1024","19188.39","21499.22","21587.29" + "aes-192-cbc","16","345.95","383.05","370.81" + "aes-192-cbc","16384","105316.35","109843.80","116539.39" + "aes-192-cbc","256","6100.82","6159.79","6238.46" + "aes-192-cbc","64","1526.98","1543.04","1620.65" + "aes-192-cbc","8192","80538.28","84535.98","89066.15" + "aes-192-ecb","1024","19876.52","22543.70","22193.49" + "aes-192-ecb","16","361.37","383.07","382.09" + "aes-192-ecb","16384","111351.13","113628.50","121503.74" + "aes-192-ecb","256","6097.66","6170.03","6481.15" + "aes-192-ecb","64","1538.15","1546.99","1680.51" + "aes-192-ecb","8192","84372.14","88003.93","92105.39" + "aes-256-cbc","1024","18930.69","21275.65","21359.96" + "aes-256-cbc","16","346.96","382.25","370.81" + "aes-256-cbc","16384","98604.37","102520.15","109614.42" + "aes-256-cbc","256","6085.97","6156.37","6247.08" + "aes-256-cbc","64","1530.07","1540.33","1618.54" + "aes-256-cbc","8192","77313.37","81087.15","84762.62" + "aes-256-ecb","1024","19814.06","22423.21","22128.98" + "aes-256-ecb","16","362.64","383.66","376.89" + "aes-256-ecb","16384","103180.97","106779.99","114005.33" + "aes-256-ecb","256","6135.13","6169.94","6384.81" + "aes-256-ecb","64","1537.32","1547.33","1658.11" + "aes-256-ecb","8192","79563.43","84475.90","87960.23" + "sha256","1024","24798.21","29029.38","30274.56" + "sha256","16","426.65","481.71","501.07" + "sha256","16384","242909.18","261947.39","251062.95" + "sha256","256","6633.39","7654.49","7859.88" + "sha256","64","1676.12","1916.42","1985.81" + "sha256","8192","156958.72","174140.07","168424.79" + "sha512","1024","20430.17","23322.97","21332.99" + "sha512","16","398.69","480.28","483.14" + "sha512","16384","59561.30","65650.69","60030.98" + "sha512","256","6163.97","6760.70","6936.66" + "sha512","64","1589.91","1920.30","1929.90" + "sha512","8192","52232.19","58804.91","53398.19" + + + +.. csv-table:: OpenSSL CPU Load + :header: "Algorithm","am62xx_lp_sk-fs: CPU Load","am62xx_sk-fs: CPU Load","am62xxsip_sk-fs: CPU Load" + + "aes-128-cbc","35.00","35.00","35.00" + "aes-128-ecb","37.00","38.00","36.00" + "aes-192-cbc","35.00","36.00","35.00" + "aes-192-ecb","36.00","37.00","36.00" + "aes-256-cbc","35.00","35.00","34.00" + "aes-256-ecb","36.00","36.00","36.00" + "sha256","94.00","94.00","95.00" + "sha512","94.00","95.00","95.00" + + + +Listed for each algorithm are the code snippets used to run each + benchmark test. + +:: + + time -v openssl speed -elapsed -evp aes-128-cbc + + IPSec Software Performance -^^^^^^^^^^^^^^^^^^^^^^^^^^^ +^^^^^^^^^^^^^^^^^^^^^^^^^^ .. csv-table:: IPSec Software Performance - :header: "Algorithm","am62xx_sk-fs: Throughput (Mbps)","am62xx_sk-fs: Packets/Sec","am62xx_sk-fs: CPU Load" - - "aes128","1.20","0.00","51.33" - "aes192","2.30","0.00","75.29" + :header: "Algorithm","am62xx_sk-fs: Throughput (Mbps)","am62xx_sk-fs: Packets/Sec","am62xx_sk-fs: CPU Load","am62xxsip_sk-fs: Throughput (Mbps)","am62xxsip_sk-fs: Packets/Sec","am62xxsip_sk-fs: CPU Load" -| + "aes128","1.20","0.00","51.32","4.20","0.00","70.26" + "aes192","1.30","0.00","51.25","2.90","0.00","79.48" + "aes256","1.20","0.00","51.52","216.40","19.00","28.32" Low Power Performance ------------------------- diff --git a/source/devices/AM62X/linux/RT_Linux_Performance_Guide.rst b/source/devices/AM62X/linux/RT_Linux_Performance_Guide.rst index 513560424..35de6053f 100644 --- a/source/devices/AM62X/linux/RT_Linux_Performance_Guide.rst +++ b/source/devices/AM62X/linux/RT_Linux_Performance_Guide.rst @@ -1,6 +1,6 @@ -====================================== - RT-linux 11.00 Performance Guide -====================================== +================================= + RT-linux 11.01 Performance Guide +================================= .. rubric:: **Read This First** :name: read-this-first-kernel-perf-guide @@ -38,7 +38,7 @@ https://e2e.ti.com/ or https://support.ti.com/ | System Benchmarks -------------------------- +----------------- Stress-ng and Cyclic Test diff --git a/source/devices/AM62X/linux/Release_Specific_Release_Notes.rst b/source/devices/AM62X/linux/Release_Specific_Release_Notes.rst index 796052c9e..d3a802a9b 100644 --- a/source/devices/AM62X/linux/Release_Specific_Release_Notes.rst +++ b/source/devices/AM62X/linux/Release_Specific_Release_Notes.rst @@ -45,8 +45,8 @@ What's new **Processor SDK Linux AM62X Release has following new features:** -- Second 2025 LTS Reference Release Including RT combined branch model - - Falcon mode through R5 SPL + - Second 2025 LTS Reference Release Including RT combined branch model + - Falcon mode through R5 SPL :ref:`U-Boot Falcon Mode ` - Important Bug Fixes on top of Processor SDK 11.00.09.04 Release - Review Issue Tracker Section for the new fixes. diff --git a/source/devices/AM64X/linux/Linux_Performance_Guide.rst b/source/devices/AM64X/linux/Linux_Performance_Guide.rst index ea34f9d83..1b76ee48e 100644 --- a/source/devices/AM64X/linux/Linux_Performance_Guide.rst +++ b/source/devices/AM64X/linux/Linux_Performance_Guide.rst @@ -1,6 +1,6 @@ -====================================== - Linux 11.00 Performance Guide -====================================== +============================== + Linux 11.01 Performance Guide +============================== .. rubric:: **Read This First** diff --git a/source/devices/AM64X/linux/RT_Linux_Performance_Guide.rst b/source/devices/AM64X/linux/RT_Linux_Performance_Guide.rst index 533dba10b..ccd6cd7a7 100644 --- a/source/devices/AM64X/linux/RT_Linux_Performance_Guide.rst +++ b/source/devices/AM64X/linux/RT_Linux_Performance_Guide.rst @@ -1,6 +1,6 @@ -====================================== -RT-linux 11.00 Performance Guide -====================================== +================================ +RT-linux 11.01 Performance Guide +================================ .. rubric:: **Read This First** :name: read-this-first-rt-kernel-perf-guide @@ -34,7 +34,7 @@ https://e2e.ti.com/ or https://support.ti.com/ | System Benchmarks -------------------------- +----------------- Stress-ng and Cyclic Test ^^^^^^^^^^^^^^^^^^^^^^^^^ @@ -72,13 +72,13 @@ The latencies observed with this DK are summarized below: | LMBench -^^^^^^^^^^^^^^^^^^^^^^^^^^^ -LMBench is a collection of microbenchmarks of which the memory bandwidth -and latency related ones are typically used to estimate processor -memory system performance. More information about lmbench at -https://lmbench.sourceforge.net/whatis_lmbench.html and -https://lmbench.sourceforge.net/man/lmbench.8.html - +^^^^^^^ +LMBench is a collection of microbenchmarks of which the memory bandwidth +and latency related ones are typically used to estimate processor +memory system performance. More information about lmbench at +http://lmbench.sourceforge.net/whatis_lmbench.html and +http://lmbench.sourceforge.net/man/lmbench.8.html + **Latency**: lat_mem_rd-stride128-szN, where N is equal to or smaller than the cache size at given level measures the cache miss penalty. N that is at least double the size of last level cache is the latency to external memory. @@ -99,124 +99,126 @@ Execute the LMBench with the following: .. csv-table:: LMBench Benchmarks :header: "Benchmarks","am64xx-hsevm: perf" - "af_unix_sock_stream_latency (microsec)","39.20" - "af_unix_socket_stream_bandwidth (MBs)","535.26" - "bw_file_rd-io-1mb (MB/s)","873.06" - "bw_file_rd-o2c-1mb (MB/s)","500.58" - "bw_mem-bcopy-16mb (MB/s)","976.92" - "bw_mem-bcopy-1mb (MB/s)","995.36" - "bw_mem-bcopy-2mb (MB/s)","987.65" - "bw_mem-bcopy-4mb (MB/s)","912.30" - "bw_mem-bcopy-8mb (MB/s)","987.41" - "bw_mem-bzero-16mb (MB/s)","2117.24" - "bw_mem-bzero-1mb (MB/s)","1556.44 (min 995.36, max 2117.52)" - "bw_mem-bzero-2mb (MB/s)","1553.34 (min 987.65, max 2119.02)" - "bw_mem-bzero-4mb (MB/s)","1513.23 (min 912.30, max 2114.16)" - "bw_mem-bzero-8mb (MB/s)","1551.49 (min 987.41, max 2115.56)" - "bw_mem-cp-16mb (MB/s)","533.72" - "bw_mem-cp-1mb (MB/s)","1553.21 (min 542.79, max 2563.63)" - "bw_mem-cp-2mb (MB/s)","1448.26 (min 557.34, max 2339.18)" - "bw_mem-cp-4mb (MB/s)","1393.63 (min 562.98, max 2224.28)" - "bw_mem-cp-8mb (MB/s)","1343.05 (min 539.30, max 2146.79)" - "bw_mem-fcp-16mb (MB/s)","1084.82" - "bw_mem-fcp-1mb (MB/s)","1600.87 (min 1084.21, max 2117.52)" - "bw_mem-fcp-2mb (MB/s)","1579.89 (min 1040.76, max 2119.02)" - "bw_mem-fcp-4mb (MB/s)","1586.67 (min 1059.18, max 2114.16)" - "bw_mem-fcp-8mb (MB/s)","1565.07 (min 1014.58, max 2115.56)" - "bw_mem-frd-16mb (MB/s)","1307.30" - "bw_mem-frd-1mb (MB/s)","1174.33 (min 1084.21, max 1264.45)" - "bw_mem-frd-2mb (MB/s)","1191.30 (min 1040.76, max 1341.83)" - "bw_mem-frd-4mb (MB/s)","1177.79 (min 1059.18, max 1296.39)" - "bw_mem-frd-8mb (MB/s)","1125.34 (min 1014.58, max 1236.09)" - "bw_mem-fwr-16mb (MB/s)","2126.25" - "bw_mem-fwr-1mb (MB/s)","1914.04 (min 1264.45, max 2563.63)" - "bw_mem-fwr-2mb (MB/s)","1840.51 (min 1341.83, max 2339.18)" - "bw_mem-fwr-4mb (MB/s)","1760.34 (min 1296.39, max 2224.28)" - "bw_mem-fwr-8mb (MB/s)","1691.44 (min 1236.09, max 2146.79)" - "bw_mem-rd-16mb (MB/s)","1339.70" - "bw_mem-rd-1mb (MB/s)","1067.51 (min 833.89, max 1301.12)" - "bw_mem-rd-2mb (MB/s)","1096.69 (min 826.56, max 1366.82)" - "bw_mem-rd-4mb (MB/s)","1084.80 (min 834.03, max 1335.56)" - "bw_mem-rd-8mb (MB/s)","1110.87 (min 876.52, max 1345.22)" - "bw_mem-rdwr-16mb (MB/s)","870.23" - "bw_mem-rdwr-1mb (MB/s)","677.24 (min 542.79, max 811.69)" - "bw_mem-rdwr-2mb (MB/s)","706.14 (min 557.34, max 854.94)" - "bw_mem-rdwr-4mb (MB/s)","697.94 (min 562.98, max 832.90)" - "bw_mem-rdwr-8mb (MB/s)","697.96 (min 539.30, max 856.62)" - "bw_mem-wr-16mb (MB/s)","886.23" - "bw_mem-wr-1mb (MB/s)","822.79 (min 811.69, max 833.89)" - "bw_mem-wr-2mb (MB/s)","840.75 (min 826.56, max 854.94)" - "bw_mem-wr-4mb (MB/s)","833.47 (min 832.90, max 834.03)" - "bw_mem-wr-8mb (MB/s)","866.57 (min 856.62, max 876.52)" - "bw_mmap_rd-mo-1mb (MB/s)","1327.77" - "bw_mmap_rd-o2c-1mb (MB/s)","466.78" - "bw_pipe (MB/s)","543.51" - "bw_unix (MB/s)","535.26" - "lat_connect (us)","75.45" - "lat_ctx-2-128k (us)","14.36" - "lat_ctx-2-256k (us)","32.22" - "lat_ctx-4-128k (us)","11.99" - "lat_ctx-4-256k (us)","19.26" - "lat_fs-0k (num_files)","180.00" - "lat_fs-10k (num_files)","84.00" - "lat_fs-1k (num_files)","120.00" - "lat_fs-4k (num_files)","123.00" - "lat_mem_rd-stride128-sz1000k (ns)","48.32" - "lat_mem_rd-stride128-sz125k (ns)","7.85" - "lat_mem_rd-stride128-sz250k (ns)","9.46" - "lat_mem_rd-stride128-sz31k (ns)","3.06" + "af_unix_sock_stream_latency (microsec)","43.92 (min 40.29, max 49.64)" + "af_unix_socket_stream_bandwidth (mb\s)","534.11 (min 521.33, max 538.82)" + "bw_file_rd-io-1mb (mb/s)","848.55 (min 833.06, max 870.93)" + "bw_file_rd-o2c-1mb (mb/s)","484.02 (min 461.04, max 493.83)" + "bw_mem-bcopy-16mb (mb/s)","985.60 (min 970.29, max 999.00)" + "bw_mem-bcopy-1mb (mb/s)","948.66 (min 922.65, max 989.94)" + "bw_mem-bcopy-2mb (mb/s)","966.59 (min 936.48, max 1006.88)" + "bw_mem-bcopy-4mb (mb/s)","942.89 (min 904.98, max 972.41)" + "bw_mem-bcopy-8mb (mb/s)","967.26 (min 924.00, max 997.13)" + "bw_mem-bzero-16mb (mb/s)","2116.75 (min 2116.40, max 2116.96)" + "bw_mem-bzero-1mb (mb/s)","1531.23 (min 922.65, max 2117.90)" + "bw_mem-bzero-2mb (mb/s)","1540.94 (min 936.48, max 2118.27)" + "bw_mem-bzero-4mb (mb/s)","1526.44 (min 904.98, max 2113.05)" + "bw_mem-bzero-8mb (mb/s)","1541.80 (min 924.00, max 2118.36)" + "bw_mem-cp-16mb (mb/s)","568.31 (min 522.04, max 598.00)" + "bw_mem-cp-1mb (mb/s)","1609.38 (min 584.20, max 2690.24)" + "bw_mem-cp-2mb (mb/s)","1447.17 (min 534.47, max 2359.88)" + "bw_mem-cp-4mb (mb/s)","1374.97 (min 536.91, max 2217.70)" + "bw_mem-cp-8mb (mb/s)","1342.76 (min 529.07, max 2152.56)" + "bw_mem-fcp-16mb (mb/s)","1012.13 (min 983.47, max 1038.08)" + "bw_mem-fcp-1mb (mb/s)","1597.21 (min 1037.88, max 2117.90)" + "bw_mem-fcp-2mb (mb/s)","1576.33 (min 1004.02, max 2118.27)" + "bw_mem-fcp-4mb (mb/s)","1577.93 (min 985.71, max 2113.05)" + "bw_mem-fcp-8mb (mb/s)","1573.48 (min 969.93, max 2118.36)" + "bw_mem-frd-16mb (mb/s)","1302.42 (min 1283.39, max 1312.98)" + "bw_mem-frd-1mb (mb/s)","1176.71 (min 1037.88, max 1287.47)" + "bw_mem-frd-2mb (mb/s)","1173.57 (min 1004.02, max 1329.12)" + "bw_mem-frd-4mb (mb/s)","1164.97 (min 985.71, max 1321.22)" + "bw_mem-frd-8mb (mb/s)","1173.06 (min 969.93, max 1334.22)" + "bw_mem-fwr-16mb (mb/s)","2121.88 (min 2118.36, max 2125.96)" + "bw_mem-fwr-1mb (mb/s)","1951.85 (min 1252.46, max 2690.24)" + "bw_mem-fwr-2mb (mb/s)","1824.75 (min 1296.39, max 2359.88)" + "bw_mem-fwr-4mb (mb/s)","1742.03 (min 1258.06, max 2217.70)" + "bw_mem-fwr-8mb (mb/s)","1729.04 (min 1285.76, max 2152.56)" + "bw_mem-rd-16mb (mb/s)","1338.76 (min 1331.78, max 1342.62)" + "bw_mem-rd-1mb (mb/s)","1078.25 (min 788.52, max 1365.85)" + "bw_mem-rd-2mb (mb/s)","1079.56 (min 790.93, max 1355.01)" + "bw_mem-rd-4mb (mb/s)","1086.93 (min 810.62, max 1351.58)" + "bw_mem-rd-8mb (mb/s)","1111.91 (min 863.19, max 1350.44)" + "bw_mem-rdwr-16mb (mb/s)","860.25 (min 847.73, max 871.32)" + "bw_mem-rdwr-1mb (mb/s)","705.72 (min 584.20, max 847.17)" + "bw_mem-rdwr-2mb (mb/s)","687.14 (min 534.47, max 851.18)" + "bw_mem-rdwr-4mb (mb/s)","681.11 (min 536.91, max 839.72)" + "bw_mem-rdwr-8mb (mb/s)","695.52 (min 529.07, max 856.16)" + "bw_mem-wr-16mb (mb/s)","889.48 (min 888.15, max 890.47)" + "bw_mem-wr-1mb (mb/s)","818.14 (min 785.67, max 847.17)" + "bw_mem-wr-2mb (mb/s)","817.43 (min 773.69, max 851.18)" + "bw_mem-wr-4mb (mb/s)","821.37 (min 784.01, max 843.79)" + "bw_mem-wr-8mb (mb/s)","862.85 (min 844.24, max 885.54)" + "bw_mmap_rd-mo-1mb (mb/s)","1297.19 (min 1274.58, max 1315.05)" + "bw_mmap_rd-o2c-1mb (mb/s)","467.73 (min 455.72, max 477.86)" + "bw_pipe (mb/s)","550.83 (min 538.93, max 563.59)" + "bw_unix (mb/s)","534.11 (min 521.33, max 538.82)" + "lat_connect (us)","76.50 (min 75.69, max 77.20)" + "lat_ctx-2-128k (us)","10.52 (min 9.79, max 11.79)" + "lat_ctx-2-256k (us)","25.39 (min 15.41, max 30.59)" + "lat_ctx-4-128k (us)","11.98 (min 11.81, max 12.29)" + "lat_ctx-4-256k (us)","12.54 (min 6.99, max 15.42)" + "lat_fs-0k (num_files)","195.75 (min 188.00, max 202.00)" + "lat_fs-10k (num_files)","86.50 (min 83.00, max 90.00)" + "lat_fs-1k (num_files)","134.25 (min 129.00, max 139.00)" + "lat_fs-4k (num_files)","119.75 (min 112.00, max 128.00)" + "lat_mem_rd-stride128-sz1000k (ns)","48.27 (min 48.18, max 48.38)" + "lat_mem_rd-stride128-sz125k (ns)","7.83 (min 7.81, max 7.86)" + "lat_mem_rd-stride128-sz250k (ns)","12.79 (min 9.36, max 18.06)" + "lat_mem_rd-stride128-sz31k (ns)","5.68 (min 5.13, max 6.57)" "lat_mem_rd-stride128-sz50 (ns)","3.02" - "lat_mem_rd-stride128-sz500k (ns)","44.34" - "lat_mem_rd-stride128-sz62k (ns)","7.37" - "lat_mmap-1m (us)","65.00" - "lat_ops-double-add (ns)","4.02" - "lat_ops-double-div (ns)","22.11" - "lat_ops-double-mul (ns)","4.02" + "lat_mem_rd-stride128-sz500k (ns)","43.72 (min 42.55, max 44.62)" + "lat_mem_rd-stride128-sz62k (ns)","7.41 (min 7.37, max 7.47)" + "lat_mmap-1m (us)","71.50 (min 65.00, max 79.00)" + "lat_ops-double-add (ns)","4.02 (min 4.02, max 4.03)" + "lat_ops-double-div (ns)","22.12 (min 22.12, max 22.13)" + "lat_ops-double-mul (ns)","4.02 (min 4.02, max 4.03)" "lat_ops-float-add (ns)","4.02" - "lat_ops-float-div (ns)","13.07" + "lat_ops-float-div (ns)","13.07 (min 13.07, max 13.08)" "lat_ops-float-mul (ns)","4.02" "lat_ops-int-add (ns)","1.01" "lat_ops-int-bit (ns)","0.67" - "lat_ops-int-div (ns)","6.03" - "lat_ops-int-mod (ns)","6.37" - "lat_ops-int-mul (ns)","4.32" + "lat_ops-int-div (ns)","6.04" + "lat_ops-int-mod (ns)","6.38 (min 6.37, max 6.39)" + "lat_ops-int-mul (ns)","4.35 (min 4.32, max 4.40)" "lat_ops-int64-add (ns)","1.01" "lat_ops-int64-bit (ns)","0.67" - "lat_ops-int64-div (ns)","9.55" - "lat_ops-int64-mod (ns)","7.37" - "lat_ops-int64-mul (ns)","5.00" - "lat_pagefault (us)","1.82" - "lat_pipe (us)","26.36" - "lat_proc-exec (us)","1157.20" - "lat_proc-fork (us)","1010.83" + "lat_ops-int64-div (ns)","9.56 (min 9.55, max 9.56)" + "lat_ops-int64-mod (ns)","7.38 (min 7.37, max 7.38)" + "lat_ops-int64-mul (ns)","5.01 (min 4.99, max 5.06)" + "lat_pagefault (us)","1.83 (min 1.81, max 1.84)" + "lat_pipe (us)","26.04 (min 25.91, max 26.16)" + "lat_proc-exec (us)","1208.90 (min 1194.20, max 1221.80)" + "lat_proc-fork (us)","1035.45 (min 1016.67, max 1052.33)" "lat_proc-proccall (us)","0.01" - "lat_select (us)","46.54" - "lat_sem (us)","3.37" - "lat_sig-catch (us)","5.98" - "lat_sig-install (us)","0.91" - "lat_sig-prot (us)","0.36" - "lat_syscall-fstat (us)","2.33" - "lat_syscall-null (us)","0.58" - "lat_syscall-open (us)","406.36" - "lat_syscall-read (us)","0.84" - "lat_syscall-stat (us)","6.48" - "lat_syscall-write (us)","0.77" - "lat_tcp (us)","1.11" - "lat_unix (us)","39.20" - "latency_for_0.50_mb_block_size (nanosec)","44.34" - "latency_for_1.00_mb_block_size (nanosec)","24.16 (min 0.00, max 48.32)" - "pipe_bandwidth (MBs)","543.51" - "pipe_latency (microsec)","26.36" + "lat_select (us)","47.62 (min 46.19, max 49.96)" + "lat_sem (us)","3.47 (min 3.28, max 3.55)" + "lat_sig-catch (us)","5.64 (min 5.50, max 5.97)" + "lat_sig-install (us)","0.92 (min 0.90, max 0.94)" + "lat_sig-prot (us)","0.77 (min 0.49, max 0.88)" + "lat_syscall-fstat (us)","2.48 (min 2.33, max 2.69)" + "lat_syscall-null (us)","0.57 (min 0.57, max 0.58)" + "lat_syscall-open (us)","415.84 (min 400.86, max 439.75)" + "lat_syscall-read (us)","0.86 (min 0.82, max 0.93)" + "lat_syscall-stat (us)","6.86 (min 6.57, max 7.24)" + "lat_syscall-write (us)","0.80 (min 0.79, max 0.83)" + "lat_tcp (us)","1.10 (min 1.09, max 1.11)" + "lat_unix (us)","43.92 (min 40.29, max 49.64)" + "latency_for_0.50_mb_block_size (nanosec)","43.72 (min 42.55, max 44.62)" + "latency_for_1.00_mb_block_size (nanosec)","24.13 (min 0.00, max 48.38)" + "pipe_bandwidth (mb\s)","550.83 (min 538.93, max 563.59)" + "pipe_latency (microsec)","26.04 (min 25.91, max 26.16)" "procedure_call (microsec)","0.01" - "select_on_200_tcp_fds (microsec)","46.54" - "semaphore_latency (microsec)","3.37" - "signal_handler_latency (microsec)","0.91" - "signal_handler_overhead (microsec)","5.98" - "tcp_ip_connection_cost_to_localhost (microsec)","75.45" - "tcp_latency_using_localhost (microsec)","1.11" + "select_on_200_tcp_fds (microsec)","47.62 (min 46.19, max 49.96)" + "semaphore_latency (microsec)","3.47 (min 3.28, max 3.55)" + "signal_handler_latency (microsec)","0.92 (min 0.90, max 0.94)" + "signal_handler_overhead (microsec)","5.64 (min 5.50, max 5.97)" + "tcp_ip_connection_cost_to_localhost (microsec)","76.50 (min 75.69, max 77.20)" + "tcp_latency_using_localhost (microsec)","1.10 (min 1.09, max 1.11)" + + Dhrystone -^^^^^^^^^^^^^^^^^^^^^^^^^^^ +^^^^^^^^^ Dhrystone is a core only benchmark that runs from warm L1 caches in all modern processors. It scales linearly with clock speed. For standard ARM cores the DMIPS/MHz score will be identical with the same compiler and flags. @@ -224,86 +226,73 @@ cores the DMIPS/MHz score will be identical with the same compiler and flags. .. csv-table:: Dhrystone Benchmarks :header: "Benchmarks","am64xx-hsevm: perf" - "cpu_clock (MHz)","1000.00" - "dhrystone_per_mhz (DMIPS/MHz)","2.90" - "dhrystone_per_second (DhrystoneP)","5128205.00" + "cpu_clock (mhz)","1000.00" + "dhrystone_per_mhz (dmips/mhz)","2.88 (min 2.80, max 2.90)" + "dhrystone_per_second (dhrystonep)","5096153.75 (min 5000000.00, max 5128205.00)" + Whetstone -^^^^^^^^^^^^^^^^^^^^^^^^^^^ +^^^^^^^^^ .. csv-table:: Whetstone Benchmarks :header: "Benchmarks","am64xx-hsevm: perf" - "whetstone (MIPS)","5000.00" + "whetstone (mips)","5000.00" + + Linpack -^^^^^^^^^^^^^^^^^^^^^^^^^^^ +^^^^^^^ Linpack measures peak double precision (64 bit) floating point performance in solving a dense linear system. .. csv-table:: Linpack Benchmarks :header: "Benchmarks","am64xx-hsevm: perf" - "linpack (Kflops)","407227.00" - -Stream -^^^^^^^^^^^^^^^^^^^^^^^^^^^ -STREAM is a microbenchmark for measuring data memory system performance without -any data reuse. It is designed to miss on caches and exercise data prefetcher -and speculative accesses. -It uses double precision floating point (64bit) but in -most modern processors the memory access will be the bottleneck. -The four individual scores are copy, scale as in multiply by constant, -add two numbers, and triad for multiply accumulate. -For bandwidth, a byte read counts as one and a byte written counts as one, -resulting in a score that is double the bandwidth LMBench will show. - -.. csv-table:: Stream Benchmarks - :header: "Benchmarks","am64xx-hsevm: perf" + "linpack (kflops)","409028.75 (min 408609.00, max 409559.00)" - "add (MB/s)","1619.70" - "copy (MB/s)","2019.30" - "scale (MB/s)","2250.40" - "triad (MB/s)","1615.70" CoreMarkPro -^^^^^^^^^^^^^^^^^^^^^^^^^^^ +^^^^^^^^^^^ CoreMark®-Pro is a comprehensive, advanced processor benchmark that works with and enhances the market-proven industry-standard EEMBC CoreMark® benchmark. While CoreMark stresses the CPU pipeline, CoreMark-Pro tests the entire processor, adding comprehensive support for multicore technology, a combination of integer and floating-point workloads, and data sets for utilizing larger memory subsystems. + .. csv-table:: CoreMarkPro Benchmarks :header: "Benchmarks","am64xx-hsevm: perf" - "cjpeg-rose7-preset (workloads/)","29.41" + "cjpeg-rose7-preset (workloads/)","29.56 (min 29.50, max 29.67)" "core (workloads/)","0.21" - "coremark-pro ()","587.13" - "linear_alg-mid-100x100-sp (workloads/)","10.42" + "coremark-pro ()","587.32 (min 583.40, max 594.37)" + "linear_alg-mid-100x100-sp (workloads/)","10.41 (min 10.40, max 10.42)" "loops-all-mid-10k-sp (workloads/)","0.48" "nnet_test (workloads/)","0.77" - "parser-125k (workloads/)","5.68" - "radix2-big-64k (workloads/)","19.25" + "parser-125k (workloads/)","5.50 (min 5.35, max 5.71)" + "radix2-big-64k (workloads/)","20.12 (min 19.27, max 21.25)" "sha-test (workloads/)","57.80" - "zip-test (workloads/)","15.63" + "zip-test (workloads/)","15.44 (min 15.38, max 15.63)" + .. csv-table:: CoreMarkProTwoCore Benchmarks :header: "Benchmarks","am64xx-hsevm: perf" - "cjpeg-rose7-preset (workloads/)","58.82" - "core (workloads/)","0.42" - "coremark-pro ()","1034.74" - "linear_alg-mid-100x100-sp (workloads/)","20.65" - "loops-all-mid-10k-sp (workloads/)","0.86" - "nnet_test (workloads/)","1.53" - "parser-125k (workloads/)","5.43" - "radix2-big-64k (workloads/)","32.63" - "sha-test (workloads/)","114.94" - "zip-test (workloads/)","28.17" + "cjpeg-rose7-preset (workloads/)","58.15 (min 57.47, max 59.17)" + "core (workloads/)","0.43" + "coremark-pro ()","1047.54 (min 1043.17, max 1054.31)" + "linear_alg-mid-100x100-sp (workloads/)","20.80 (min 20.77, max 20.82)" + "loops-all-mid-10k-sp (workloads/)","0.87 (min 0.87, max 0.88)" + "nnet_test (workloads/)","1.54" + "parser-125k (workloads/)","5.84 (min 5.71, max 5.93)" + "radix2-big-64k (workloads/)","32.26 (min 30.87, max 33.46)" + "sha-test (workloads/)","115.28 (min 114.94, max 116.28)" + "zip-test (workloads/)","28.37 (min 28.17, max 28.57)" + MultiBench -^^^^^^^^^^^^^^^^^^^^^^^^^^^ +^^^^^^^^^^ MultiBench™ is a suite of benchmarks that allows processor and system designers to analyze, test, and improve multicore processors. It uses three forms of concurrency: Data decomposition: multiple threads cooperating on achieving a unified goal and @@ -322,75 +311,83 @@ thread-enabled workloads to be tested. .. csv-table:: Multibench Benchmarks :header: "Benchmarks","am64xx-hsevm: perf" - "4m-check (workloads/)","281.69" - "4m-check-reassembly (workloads/)","61.50" - "4m-check-reassembly-tcp (workloads/)","36.28" - "4m-check-reassembly-tcp-cmykw2-rotatew2 (workloads/)","14.43" - "4m-check-reassembly-tcp-x264w2 (workloads/)","0.75" - "4m-cmykw2 (workloads/)","85.14" - "4m-cmykw2-rotatew2 (workloads/)","18.51" - "4m-reassembly (workloads/)","56.18" - "4m-rotatew2 (workloads/)","19.53" - "4m-tcp-mixed (workloads/)","81.63" - "4m-x264w2 (workloads/)","0.76" + "4m-check (workloads/)","281.56 (min 281.15, max 281.85)" + "4m-check-reassembly (workloads/)","61.33 (min 61.09, max 61.61)" + "4m-check-reassembly-tcp (workloads/)","36.25 (min 35.77, max 36.60)" + "4m-check-reassembly-tcp-cmykw2-rotatew2 (workloads/)","14.55 (min 14.38, max 14.71)" + "4m-check-reassembly-tcp-x264w2 (workloads/)","0.75 (min 0.74, max 0.76)" + "4m-cmykw2 (workloads/)","85.64 (min 84.64, max 86.24)" + "4m-cmykw2-rotatew2 (workloads/)","18.45 (min 17.29, max 19.80)" + "4m-reassembly (workloads/)","56.74 (min 56.63, max 56.88)" + "4m-rotatew2 (workloads/)","21.89 (min 20.05, max 23.13)" + "4m-tcp-mixed (workloads/)","85.56 (min 85.11, max 86.02)" + "4m-x264w2 (workloads/)","0.76 (min 0.76, max 0.77)" "empty-wld (workloads/)","1.00" - "idct-4m (workloads/)","13.62" - "idct-4mw1 (workloads/)","13.62" - "ippktcheck-4m (workloads/)","281.82" - "ippktcheck-4mw1 (workloads/)","280.93" - "ipres-4m (workloads/)","72.74" - "ipres-4mw1 (workloads/)","72.18" - "md5-4m (workloads/)","19.67" - "md5-4mw1 (workloads/)","19.81" - "rgbcmyk-4m (workloads/)","44.61" - "rgbcmyk-4mw1 (workloads/)","44.71" - "rotate-4ms1 (workloads/)","16.56" - "rotate-4ms1w1 (workloads/)","16.52" - "rotate-4ms64 (workloads/)","16.73" - "rotate-4ms64w1 (workloads/)","16.77" + "idct-4m (workloads/)","13.68 (min 13.68, max 13.69)" + "idct-4mw1 (workloads/)","13.68 (min 13.67, max 13.69)" + "ippktcheck-4m (workloads/)","281.33 (min 280.65, max 282.01)" + "ippktcheck-4mw1 (workloads/)","281.37 (min 280.87, max 281.85)" + "ipres-4m (workloads/)","72.63 (min 72.46, max 72.92)" + "ipres-4mw1 (workloads/)","71.88 (min 71.29, max 72.39)" + "md5-4m (workloads/)","19.75 (min 19.58, max 19.93)" + "md5-4mw1 (workloads/)","19.74 (min 19.45, max 20.00)" + "rgbcmyk-4m (workloads/)","44.70 (min 44.68, max 44.72)" + "rgbcmyk-4mw1 (workloads/)","44.69 (min 44.68, max 44.70)" + "rotate-4ms1 (workloads/)","16.57 (min 16.51, max 16.59)" + "rotate-4ms1w1 (workloads/)","16.60 (min 16.59, max 16.61)" + "rotate-4ms64 (workloads/)","16.79 (min 16.79, max 16.80)" + "rotate-4ms64w1 (workloads/)","16.79 (min 16.76, max 16.80)" "x264-4mq (workloads/)","0.41" "x264-4mqw1 (workloads/)","0.41" + + + + Boot-time Measurement --------------------- + Boot media: MMCSD ^^^^^^^^^^^^^^^^^ .. csv-table:: Linux boot time MMCSD :header: "Boot Configuration","am64xx-hsevm: Boot time in seconds: avg(min,max)" - "Linux boot time from SD with default rootfs (20 boot cycles)","21.36 (min 20.52, max 21.88)" + "Linux boot time from SD with default rootfs (20 boot cycles)","22.75 (min 20.01, max 29.10)" + + Boot time numbers [avg, min, max] are measured from "Starting kernel" to Linux prompt across 20 boot cycles. + | Ethernet ------------------ +-------- Ethernet performance benchmarks were measured using Netperf 2.7.1 https://hewlettpackard.github.io/netperf/doc/netperf.html Test procedures were modeled after those defined in RFC-2544: -https://tools.ietf.org/html/rfc2544, where the DUT is the TI device +https://tools.ietf.org/html/rfc2544, where the DUT is the TI device and the "tester" used was a Linux PC. To produce consistent results, -it is recommended to carry out performance tests in a private network and to avoid -running NFS on the same interface used in the test. In these results, +it is recommended to carry out performance tests in a private network and to avoid +running NFS on the same interface used in the test. In these results, CPU utilization was captured as the total percentage used across all cores on the device, -while running the performance test over one external interface. +while running the performance test over one external interface. UDP Throughput (0% loss) was measured by the procedure defined in RFC-2544 section 26.1: Throughput. In this scenario, netperf options burst_size (-b) and wait_time (-w) are used to limit bandwidth -during different trials of the test, with the goal of finding the highest rate at which +during different trials of the test, with the goal of finding the highest rate at which no loss is seen. For example, to limit bandwidth to 500Mbits/sec with 1472B datagram: :: burst_size = / 8 (bits -> bytes) / / 100 (seconds -> 10 ms) - burst_size = 500000000 / 8 / 1472 / 100 = 425 + burst_size = 500000000 / 8 / 1472 / 100 = 425 wait_time = 10 milliseconds (minimum supported by Linux PC used for testing) UDP Throughput (possible loss) was measured by capturing throughput and packet loss statistics when -running the netperf test with no bandwidth limit (remove -b/-w options). +running the netperf test with no bandwidth limit (remove -b/-w options). In order to start a netperf client on one device, the other device must have netserver running. To start netserver: @@ -398,11 +395,11 @@ To start netserver: :: netserver [-p ] [-4 (IPv4 addressing)] [-6 (IPv6 addressing)] - -Running the following shell script from the DUT will trigger netperf clients to measure + +Running the following shell script from the DUT will trigger netperf clients to measure bidirectional TCP performance for 60 seconds and report CPU utilization. Parameter -k is used in -client commands to summarize selected statistics on their own line and -j is used to gain -additional timing measurements during the test. +client commands to summarize selected statistics on their own line and -j is used to gain +additional timing measurements during the test. :: @@ -411,51 +408,58 @@ additional timing measurements during the test. do netperf -H -j -c -l 60 -t TCP_STREAM -- -k DIRECTION,THROUGHPUT,MEAN_LATENCY,LOCAL_CPU_UTIL,REMOTE_CPU_UTIL,LOCAL_BYTES_SENT,REMOTE_BYTES_RECVD,LOCAL_SEND_SIZE & - + netperf -H -j -c -l 60 -t TCP_MAERTS -- -k DIRECTION,THROUGHPUT,MEAN_LATENCY,LOCAL_CPU_UTIL,REMOTE_CPU_UTIL,LOCAL_BYTES_SENT,REMOTE_BYTES_RECVD,LOCAL_SEND_SIZE & done -Running the following commands will trigger netperf clients to measure UDP burst performance for -60 seconds at various burst/datagram sizes and report CPU utilization. +Running the following commands will trigger netperf clients to measure UDP burst performance for +60 seconds at various burst/datagram sizes and report CPU utilization. - For UDP egress tests, run netperf client from DUT and start netserver on tester. :: - netperf -H -j -c -l 60 -t UDP_STREAM -b -w -- -m - -k DIRECTION,THROUGHPUT,MEAN_LATENCY,LOCAL_CPU_UTIL,REMOTE_CPU_UTIL,LOCAL_BYTES_SENT,REMOTE_BYTES_RECVD,LOCAL_SEND_SIZE + netperf -H -j -c -l 60 -t UDP_STREAM -b -w -- -m + -k DIRECTION,THROUGHPUT,MEAN_LATENCY,LOCAL_CPU_UTIL,REMOTE_CPU_UTIL,LOCAL_BYTES_SENT,REMOTE_BYTES_RECVD,LOCAL_SEND_SIZE -- For UDP ingress tests, run netperf client from tester and start netserver on DUT. +- For UDP ingress tests, run netperf client from tester and start netserver on DUT. :: netperf -H -j -C -l 60 -t UDP_STREAM -b -w -- -m - -k DIRECTION,THROUGHPUT,MEAN_LATENCY,LOCAL_CPU_UTIL,REMOTE_CPU_UTIL,LOCAL_BYTES_SENT,REMOTE_BYTES_RECVD,LOCAL_SEND_SIZE + -k DIRECTION,THROUGHPUT,MEAN_LATENCY,LOCAL_CPU_UTIL,REMOTE_CPU_UTIL,LOCAL_BYTES_SENT,REMOTE_BYTES_RECVD,LOCAL_SEND_SIZE -CPSW/CPSW2g/CPSW3g Ethernet Driver + +CPSW/CPSW2g/CPSW3g Ethernet Driver ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ - CPSW3g: AM64x -.. rubric:: TCP Bidirectional Throughput + +.. rubric:: TCP Bidirectional Throughput :name: CPSW2g-tcp-bidirectional-throughput .. csv-table:: CPSW2g TCP Bidirectional Throughput :header: "Command Used","am64xx-hsevm: THROUGHPUT (Mbits/sec)","am64xx-hsevm: CPU Load % (LOCAL_CPU_UTIL)" - "netperf -H 192.168.0.1 -j -c -C -l 60 -t TCP_STREAM; netperf -H 192.168.0.1 -j -c -C -l 60 -t TCP_MAERTS","1095.34","98.12" + "netperf -H 192.168.0.1 -j -c -C -l 60 -t TCP_STREAM; netperf -H 192.168.0.1 -j -c -C -l 60 -t TCP_MAERTS","1058.45 (min 964.24, max 1158.01)","88.12 (min 79.19, max 99.75)" + -ICSSG Ethernet Driver -^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ -.. rubric:: TCP Bidirectional Throughput +ICSSG Ethernet Driver +^^^^^^^^^^^^^^^^^^^^^ + +.. rubric:: TCP Bidirectional Throughput :name: tcp-bidirectional-throughput .. csv-table:: ICSSG TCP Bidirectional Throughput :header: "Command Used","am64xx-hsevm: THROUGHPUT (Mbits/sec)","am64xx-hsevm: CPU Load % (LOCAL_CPU_UTIL)" - "netperf -H 192.168.2.1 -j -c -C -l 60 -t TCP_STREAM; netperf -H 192.168.2.1 -j -c -C -l 60 -t TCP_MAERTS","183.94","37.22" + "netperf -H 192.168.2.1 -j -c -C -l 60 -t TCP_STREAM; netperf -H 192.168.2.1 -j -c -C -l 60 -t TCP_MAERTS","889.39 (min 356.03, max 1078.23)","85.19 (min 72.62, max 99.42)" + + + .. rubric:: TCP Bidirectional Throughput Interrupt Pacing :name: ICSSG-tcp-bidirectional-throughput-interrupt-pacing @@ -463,76 +467,115 @@ ICSSG Ethernet Driver .. csv-table:: ICSSG TCP Bidirectional Throughput Interrupt Pacing :header: "Command Used","am64xx-hsevm: THROUGHPUT (Mbits/sec)","am64xx-hsevm: CPU Load % (LOCAL_CPU_UTIL)" - "netperf -H 192.168.2.1 -j -c -C -l 60 -t TCP_STREAM; netperf -H 192.168.2.1 -j -c -C -l 60 -t TCP_MAERTS","359.51","53.51" + "netperf -H 192.168.2.1 -j -c -C -l 60 -t TCP_STREAM; netperf -H 192.168.2.1 -j -c -C -l 60 -t TCP_MAERTS","640.11 (min 185.59, max 1029.32)","55.82 (min 28.49, max 79.66)" + + + + +.. rubric:: UDP Egress Throughput + :name: udp-egress-throughput-0-loss + +.. csv-table:: ICSSG UDP Egress Throughput 0 loss + :header: "Frame Size(bytes)","am64xx-hsevm: UDP Datagram Size(bytes) (LOCAL_SEND_SIZE)","am64xx-hsevm: THROUGHPUT (Mbits/sec)","am64xx-hsevm: Packets Per Second (kPPS)","am64xx-hsevm: CPU Load % (LOCAL_CPU_UTIL)" + + "64","18.00","13.33 (min 6.27, max 22.35)","55.67 (min 44.00, max 79.00)","68.09 (min 58.78, max 86.60)" + "128","82.00","55.06 (min 29.23, max 79.35)","62.20 (min 43.00, max 77.00)","77.28 (min 60.07, max 89.36)" + "256","210.00","123.93 (min 70.38, max 154.02)","63.00 (min 42.00, max 75.00)","78.99 (min 58.73, max 89.81)" + "1024","978.00","292.51 (min 21.12, max 581.33)","36.75 (min 3.00, max 71.00)","46.82 (min 2.71, max 91.06)" + "1518","1472.00","685.55 (min 585.79, max 792.91)","57.25 (min 48.00, max 67.00)","87.26 (min 84.56, max 89.28)" + -.. rubric:: UDP Ingress Throughput - :name: ICSSG-udp-ingress-throughput + + +.. rubric:: UDP Ingress Throughput + :name: udp-ingress-throughput-0-loss .. csv-table:: ICSSG UDP Ingress Throughput 0 loss :header: "Frame Size(bytes)","am64xx-hsevm: UDP Datagram Size(bytes) (LOCAL_SEND_SIZE)","am64xx-hsevm: THROUGHPUT (Mbits/sec)","am64xx-hsevm: Packets Per Second (kPPS)","am64xx-hsevm: CPU Load %" - "128","82.00","1.97","3.00","0.64" - "256","210.00","7.39","4.00","0.71" - "1024","978.00","40.68","5.00","6.72" - "1518","1472.00","94.21","8.00","9.70" + "64","","1.44 (min 1.33, max 1.54)","3.00","0.10" + "128","82.00","1.97","3.00","3.07 (min 2.68, max 3.46)" + "256","210.00","9.27 (min 8.06, max 10.24)","5.00","3.87 (min 0.20, max 7.68)" + "1024","978.00","39.84 (min 37.56, max 40.96)","5.00","5.38 (min 2.89, max 7.92)" + "1472","1472.00","167.00 (min 56.52, max 374.79)","14.25 (min 5.00, max 32.00)","17.31 (min 0.25, max 49.37)" + + | OSPI Flash Driver -------------------------- +----------------- + +AM64XX-EVM +^^^^^^^^^^ + RAW -^^^ +""" .. csv-table:: OSPI Raw Flash Driver :header: "File size (Mbytes)","am64xx-hsevm: Raw Read Throughput (Mbytes/sec)" - "50","108.69" + "50","142.86" + EMMC Driver ----------- -.. important:: +.. warning:: - The performance numbers can be severely affected if the media is + **IMPORTANT**: The performance numbers can be severely affected if the media is mounted in sync mode. Hot plug scripts in the filesystem mount removable media in sync mode to ensure data integrity. For performance sensitive applications, umount the auto-mounted filesystem and re-mount in async mode. + EMMC EXT4 FIO 1G -^^^^^^^^^^^^^^^^ +"""""""""""""""" .. csv-table:: EMMC EXT4 FIO 1G :header: "Buffer size (bytes)","am64xx-hsevm: Write EXT4 Throughput (Mbytes/sec)","am64xx-hsevm: Write EXT4 CPU Load (%)","am64xx-hsevm: Read EXT4 Throughput (Mbytes/sec)","am64xx-hsevm: Read EXT4 CPU Load (%)" - "1m","60.90","4.78","175.00","8.43" - "4m","60.80","3.41","174.00","6.83" - "4k","48.40","54.81","55.80","48.81" - "256k","60.60","6.56","174.00","9.49" + "1m","60.85 (min 60.00, max 61.40)","3.92 (min 3.78, max 4.06)","175.00","7.36 (min 6.89, max 7.69)" + "4m","60.95 (min 60.10, max 61.60)","2.97 (min 2.80, max 3.07)","174.50 (min 174.00, max 175.00)","6.15 (min 5.69, max 6.48)" + "4k","49.05 (min 48.60, max 49.50)","55.20 (min 54.38, max 56.44)","56.08 (min 55.90, max 56.30)","48.50 (min 46.00, max 50.60)" + "256k","61.00 (min 60.50, max 61.40)","5.92 (min 5.70, max 6.06)","174.00","8.89 (min 8.66, max 9.25)" EMMC EXT4 -^^^^^^^^^ +""""""""" .. csv-table:: EMMC EXT4 :header: "Buffer size (bytes)","am64xx-hsevm: Write EXT4 Throughput (Mbytes/sec)","am64xx-hsevm: Write EXT4 CPU Load (%)","am64xx-hsevm: Read EXT4 Throughput (Mbytes/sec)","am64xx-hsevm: Read EXT4 CPU Load (%)" - "102400","53.73 (min 51.22, max 54.75)","11.69 (min 10.55, max 15.06)","176.86","31.48" - "262144","53.21 (min 49.80, max 54.17)","11.37 (min 10.31, max 15.20)","181.69","30.56" - "524288","53.20 (min 49.68, max 54.36)","12.17 (min 10.91, max 15.48)","182.95","28.57" - "1048576","53.49 (min 49.83, max 54.53)","11.91 (min 10.76, max 15.31)","182.97","27.68" - "5242880","53.50 (min 50.17, max 54.57)","10.88 (min 9.69, max 14.59)","182.94","27.68" + "102400","54.23 (min 49.95, max 56.05)","11.05 (min 9.63, max 15.65)","174.86 (min 168.13, max 177.19)","29.58 (min 27.83, max 30.56)" + "262144","53.37 (min 48.45, max 55.70)","10.88 (min 9.40, max 15.57)","180.84 (min 177.78, max 181.98)","33.37 (min 30.28, max 40.18)" + "524288","53.33 (min 49.35, max 55.13)","11.16 (min 9.74, max 14.88)","182.92 (min 182.86, max 183.01)","27.58 (min 27.03, max 27.93)" + "1048576","53.90 (min 49.57, max 56.36)","10.75 (min 9.38, max 14.46)","182.86 (min 182.82, max 182.90)","27.68 (min 26.79, max 28.57)" + "5242880","53.72 (min 49.33, max 55.38)","10.40 (min 9.11, max 14.18)","182.79 (min 182.63, max 182.88)","27.64 (min 27.03, max 28.57)" + +EMMC EXT2 +""""""""" + +.. csv-table:: EMMC EXT2 + :header: "Buffer size (bytes)","am64xx-hsevm: Write EXT2 Throughput (Mbytes/sec)","am64xx-hsevm: Write EXT2 CPU Load (%)","am64xx-hsevm: Read EXT2 Throughput (Mbytes/sec)","am64xx-hsevm: Read EXT2 CPU Load (%)" + + "102400","50.24 (min 41.49, max 52.87)","11.45 (min 9.39, max 18.33)","173.05 (min 172.80, max 173.40)","31.11 (min 30.28, max 32.14)" + "262144","51.48 (min 42.38, max 54.31)","11.49 (min 9.18, max 18.46)","176.38 (min 173.55, max 177.82)","32.51 (min 30.28, max 36.28)" + "524288","51.72 (min 42.47, max 54.95)","11.45 (min 9.44, max 18.60)","177.33 (min 174.83, max 178.79)","28.11 (min 27.19, max 28.70)" + "1048576","51.82 (min 42.51, max 54.93)","11.24 (min 9.23, max 18.05)","178.63 (min 178.27, max 178.86)","27.70 (min 27.43, max 27.83)" + "5242880","51.62 (min 41.99, max 54.77)","11.30 (min 9.32, max 18.16)","178.55 (min 178.35, max 178.84)","28.07 (min 27.68, max 28.70)" EMMC VFAT -^^^^^^^^^ +""""""""" .. csv-table:: EMMC VFAT :header: "Buffer size (bytes)","am64xx-hsevm: Write VFAT Throughput (Mbytes/sec)","am64xx-hsevm: Write VFAT CPU Load (%)","am64xx-hsevm: Read VFAT Throughput (Mbytes/sec)","am64xx-hsevm: Read VFAT CPU Load (%)" - "102400","49.53 (min 41.41, max 51.76)","14.70 (min 12.69, max 20.76)","165.41","31.30" - "262144","50.47 (min 41.92, max 52.82)","14.72 (min 12.98, max 20.97)","170.09","31.03" - "524288","50.37 (min 41.84, max 52.80)","14.89 (min 13.10, max 20.72)","169.50","29.51" - "1048576","50.62 (min 42.25, max 52.97)","14.22 (min 12.66, max 19.76)","169.28","27.27" - "5242880","50.74 (min 42.19, max 53.16)","13.71 (min 11.90, max 19.84)","169.63","27.73" + "102400","48.75 (min 39.83, max 52.16)","13.28 (min 11.30, max 19.92)","166.81 (min 166.61, max 167.05)","31.22 (min 30.70, max 32.17)" + "262144","50.50 (min 41.55, max 53.41)","13.72 (min 11.39, max 20.25)","168.89 (min 164.53, max 170.76)","34.76 (min 30.70, max 46.28)" + "524288","50.46 (min 41.61, max 53.75)","13.86 (min 11.72, max 19.63)","169.99 (min 169.68, max 170.18)","27.74 (min 27.50, max 28.10)" + "1048576","50.81 (min 41.99, max 54.38)","13.60 (min 11.53, max 20.17)","169.57 (min 169.41, max 169.83)","27.50" + "5242880","50.90 (min 41.73, max 54.15)","13.40 (min 11.56, max 19.68)","169.96 (min 169.64, max 170.06)","28.10 (min 27.50, max 28.69)" UBoot EMMC Driver ----------------- @@ -540,122 +583,161 @@ UBoot EMMC Driver .. csv-table:: UBOOT EMMC RAW :header: "File size (bytes in hex)","am64xx-hsevm: Write Throughput (Kbytes/sec)","am64xx-hsevm: Read Throughput (Kbytes/sec)" - "2000000","59041.44","168907.22" - "4000000","59308.60","172463.16" + "2000000","60406.03 (min 60014.65, max 61248.60)","169126.01 (min 168907.22, max 169782.38)" + "4000000","61165.23 (min 60907.06, max 61826.42)","173032.57 (min 172918.21, max 173375.66)" MMCSD ----- -.. important:: +.. warning:: - The performance numbers can be severely affected if the media is + **IMPORTANT**: The performance numbers can be severely affected if the media is mounted in sync mode. Hot plug scripts in the filesystem mount removable media in sync mode to ensure data integrity. For performance sensitive applications, umount the auto-mounted filesystem and re-mount in async mode. + MMC EXT4 FIO 1G -^^^^^^^^^^^^^^^ +""""""""""""""" + .. csv-table:: MMC EXT4 FIO 1G :header: "Buffer size (bytes)","am64xx-hsevm: Write EXT4 Throughput (Mbytes/sec)","am64xx-hsevm: Write EXT4 CPU Load (%)","am64xx-hsevm: Read EXT4 Throughput (Mbytes/sec)","am64xx-hsevm: Read EXT4 CPU Load (%)" - "1m","41.50","3.23","87.00","5.01" - "4m","41.80","2.57","87.20","3.57" - "4k","2.81","8.28","12.80","14.46" - "256k","35.90","4.53","83.10","6.21" + "4m","42.15 (min 41.50, max 43.00)","2.62 (min 2.56, max 2.71)","86.95 (min 86.80, max 87.10)","3.73 (min 3.60, max 3.92)" + "1m","41.95 (min 41.10, max 42.30)","3.52 (min 3.34, max 3.70)","86.95 (min 86.90, max 87.10)","5.01 (min 4.85, max 5.31)" + "4k","2.76 (min 2.71, max 2.82)","7.73 (min 7.65, max 7.76)","12.70 (min 12.50, max 12.80)","15.07 (min 14.49, max 15.69)" + "256k","36.50 (min 35.50, max 37.30)","4.48 (min 4.17, max 4.66)","83.10 (min 83.00, max 83.20)","6.78 (min 5.99, max 7.20)" + MMC EXT4 -^^^^^^^^ +"""""""" + + .. csv-table:: MMC EXT4 :header: "Buffer size (bytes)","am64xx-hsevm: Write Raw Throughput (Mbytes/sec)","am64xx-hsevm: Write Raw CPU Load (%)","am64xx-hsevm: Read Raw Throughput (Mbytes/sec)","am64xx-hsevm: Read Raw CPU Load (%)" - "102400","28.52 (min 27.85, max 29.34)","6.12 (min 5.46, max 8.49)","39.35","7.63" - "262144","28.36 (min 27.89, max 29.10)","6.54 (min 5.35, max 8.79)","40.40","6.82" - "524288","29.14 (min 28.88, max 29.71)","5.85 (min 4.99, max 8.69)","45.51","6.99" - "1048576","29.44 (min 29.04, max 29.61)","5.75 (min 4.73, max 8.64)","45.37","6.77" - "5242880","29.29 (min 28.70, max 30.30)","5.69 (min 4.86, max 8.71)","45.33","6.97" + "102400","35.51 (min 33.07, max 37.45)","7.64 (min 6.31, max 11.15)","73.92 (min 70.24, max 78.38)","13.81 (min 12.28, max 15.00)" + "262144","35.56 (min 34.43, max 37.31)","7.63 (min 6.38, max 11.15)","79.33 (min 76.51, max 82.53)","13.86 (min 13.06, max 14.92)" + "524288","36.53 (min 35.02, max 37.89)","7.53 (min 6.06, max 11.27)","90.56 (min 90.39, max 90.83)","13.27 (min 12.78, max 13.60)" + "1048576","36.79 (min 35.77, max 38.36)","7.50 (min 6.21, max 11.07)","89.75 (min 88.59, max 90.86)","12.83 (min 12.45, max 13.16)" + "5242880","36.94 (min 36.11, max 38.11)","7.24 (min 6.02, max 11.13)","90.18 (min 88.74, max 90.76)","13.42 (min 13.16, max 13.60)" + + +MMC EXT3 +"""""""" + + +.. csv-table:: MMC EXT3 + :header: "Buffer size (bytes)","am64xx-hsevm: Write Raw Throughput (Mbytes/sec)","am64xx-hsevm: Write Raw CPU Load (%)","am64xx-hsevm: Read Raw Throughput (Mbytes/sec)","am64xx-hsevm: Read Raw CPU Load (%)" + + "102400","35.60 (min 29.10, max 38.72)","9.59 (min 7.56, max 16.54)","75.63 (min 73.59, max 77.51)","14.68 (min 13.67, max 15.38)" + "262144","34.94 (min 29.45, max 38.84)","8.94 (min 7.00, max 15.25)","85.00 (min 84.90, max 85.05)","14.19 (min 12.77, max 15.06)" + "524288","35.04 (min 29.43, max 40.16)","8.87 (min 6.99, max 15.13)","89.16 (min 88.79, max 89.60)","13.62 (min 13.36, max 14.29)" + "1048576","34.48 (min 29.87, max 36.97)","8.60 (min 6.82, max 14.08)","88.16 (min 84.43, max 89.47)","13.86 (min 13.42, max 14.22)" + "5242880","36.10 (min 29.24, max 40.71)","8.85 (min 6.89, max 15.43)","85.76 (min 84.45, max 89.52)","14.01 (min 13.58, max 14.69)" + + +MMC EXT2 +"""""""" + +.. csv-table:: MMC EXT2 + :header: "Buffer size (bytes)","am64xx-hsevm: Write Raw Throughput (Mbytes/sec)","am64xx-hsevm: Write Raw CPU Load (%)","am64xx-hsevm: Read Raw Throughput (Mbytes/sec)","am64xx-hsevm: Read Raw CPU Load (%)" + + "102400","37.83 (min 32.12, max 39.96)","8.88 (min 7.14, max 14.78)","77.57 (min 77.14, max 77.81)","14.41 (min 13.90, max 15.06)" + "262144","35.82 (min 31.13, max 38.76)","8.27 (min 6.17, max 13.51)","84.06 (min 80.46, max 85.35)","13.15 (min 12.12, max 14.51)" + "524288","37.30 (min 31.09, max 41.07)","8.11 (min 6.16, max 14.38)","87.09 (min 84.25, max 89.67)","13.89 (min 13.52, max 14.29)" + "1048576","35.75 (min 29.84, max 39.26)","7.80 (min 6.20, max 13.56)","89.35 (min 88.93, max 89.64)","13.62 (min 13.36, max 13.85)" + "5242880","36.21 (min 30.33, max 40.10)","7.78 (min 6.06, max 14.01)","88.34 (min 84.79, max 89.80)","14.20 (min 13.85, max 14.69)" + The performance numbers were captured using the following: - SanDisk Max Endurance SD card (SDSQQVR-032G-GN6IA) - Partition was mounted with async option + | CRYPTO Driver -------------------------- +------------- OpenSSL Performance -^^^^^^^^^^^^^^^^^^^^^^^^^^^ +^^^^^^^^^^^^^^^^^^^ .. csv-table:: OpenSSL Performance :header: "Algorithm","Buffer Size (in bytes)","am64xx-hsevm: throughput (KBytes/Sec)" - "aes-128-cbc","1024","20972.89" - "aes-128-cbc","16","308.04" - "aes-128-cbc","16384","133207.38" - "aes-128-cbc","256","5377.96" - "aes-128-cbc","64","1348.82" - "aes-128-cbc","8192","98050.05" - "aes-128-ecb","1024","20139.01" - "aes-128-ecb","16","311.34" - "aes-128-ecb","16384","139935.74" - "aes-128-ecb","256","5468.25" - "aes-128-ecb","64","1381.89" - "aes-128-ecb","8192","99188.74" - "aes-192-cbc","1024","20745.56" - "aes-192-cbc","16","313.45" - "aes-192-cbc","16384","126697.47" - "aes-192-cbc","256","5491.88" - "aes-192-cbc","64","1363.61" - "aes-192-cbc","8192","92886.36" - "aes-192-ecb","1024","21486.93" - "aes-192-ecb","16","311.40" - "aes-192-ecb","16384","130913.62" - "aes-192-ecb","256","5517.31" - "aes-192-ecb","64","1386.69" - "aes-192-ecb","8192","96488.11" - "aes-256-cbc","1024","18767.87" - "aes-256-cbc","16","293.50" - "aes-256-cbc","16384","120034.65" - "aes-256-cbc","256","5228.37" - "aes-256-cbc","64","1313.64" - "aes-256-cbc","8192","89967.27" - "aes-256-ecb","1024","21326.85" - "aes-256-ecb","16","311.80" - "aes-256-ecb","16384","122404.86" - "aes-256-ecb","256","5472.26" - "aes-256-ecb","64","1403.71" - "aes-256-ecb","8192","92056.23" - "sha256","1024","25799.00" - "sha256","16","429.11" - "sha256","16384","208180.57" - "sha256","256","6721.79" - "sha256","64","1698.90" - "sha256","8192","139889.32" - "sha512","1024","17905.66" - "sha512","16","415.45" - "sha512","16384","47966.89" - "sha512","256","5931.52" - "sha512","64","1662.76" - "sha512","8192","43048.96" + "aes-128-cbc","1024","20959.06 (min 19965.27, max 21320.36)" + "aes-128-cbc","16","314.21 (min 306.05, max 318.36)" + "aes-128-cbc","16384","133997.91 (min 131858.43, max 135550.29)" + "aes-128-cbc","256","5437.10 (min 5301.33, max 5531.31)" + "aes-128-cbc","64","1338.92 (min 1300.12, max 1358.95)" + "aes-128-cbc","8192","97014.44 (min 95608.83, max 97957.21)" + "aes-128-ecb","1024","20929.19 (min 19151.87, max 21648.38)" + "aes-128-ecb","16","316.54 (min 312.22, max 321.90)" + "aes-128-ecb","16384","138835.29 (min 138067.97, max 139225.77)" + "aes-128-ecb","256","5569.86 (min 5543.42, max 5597.95)" + "aes-128-ecb","64","1381.94 (min 1351.47, max 1404.52)" + "aes-128-ecb","8192","98376.36 (min 97315.50, max 99846.83)" + "aes-192-cbc","1024","20155.73 (min 18997.25, max 21175.98)" + "aes-192-cbc","16","301.27 (min 292.46, max 316.99)" + "aes-192-cbc","16384","126993.75 (min 126528.17, max 127391.06)" + "aes-192-cbc","256","5340.86 (min 5230.76, max 5599.91)" + "aes-192-cbc","64","1325.11 (min 1298.11, max 1391.21)" + "aes-192-cbc","8192","93233.83 (min 92883.63, max 93566.29)" + "aes-192-ecb","1024","21606.31 (min 21497.86, max 21739.18)" + "aes-192-ecb","16","308.54 (min 297.98, max 319.48)" + "aes-192-ecb","16384","130969.60 (min 130804.39, max 131055.62)" + "aes-192-ecb","256","5439.27 (min 5317.21, max 5579.35)" + "aes-192-ecb","64","1347.27 (min 1306.52, max 1391.27)" + "aes-192-ecb","8192","95673.69 (min 95341.23, max 96144.04)" + "aes-256-cbc","1024","20149.93 (min 18618.71, max 21155.16)" + "aes-256-cbc","16","305.02 (min 291.87, max 317.27)" + "aes-256-cbc","16384","119682.39 (min 119177.22, max 120083.80)" + "aes-256-cbc","256","5467.90 (min 5210.97, max 5604.61)" + "aes-256-cbc","64","1361.79 (min 1306.24, max 1398.59)" + "aes-256-cbc","8192","89436.84 (min 89022.46, max 89754.28)" + "aes-256-ecb","1024","21332.99 (min 20724.05, max 21572.27)" + "aes-256-ecb","16","309.47 (min 297.86, max 319.90)" + "aes-256-ecb","16384","123528.53 (min 122836.31, max 124245.33)" + "aes-256-ecb","256","5514.73 (min 5286.66, max 5620.48)" + "aes-256-ecb","64","1361.30 (min 1325.67, max 1396.20)" + "aes-256-ecb","8192","91379.71 (min 90923.01, max 92198.23)" + "sha256","1024","26057.22 (min 25825.96, max 26321.58)" + "sha256","16","433.83 (min 431.97, max 435.66)" + "sha256","16384","209231.87 (min 208125.95, max 210031.96)" + "sha256","256","6812.05 (min 6772.82, max 6881.45)" + "sha256","64","1721.55 (min 1711.42, max 1737.02)" + "sha256","8192","141288.79 (min 140610.22, max 141890.90)" + "sha512","1024","18045.35 (min 18011.14, max 18106.71)" + "sha512","16","420.20 (min 419.00, max 421.30)" + "sha512","16384","48089.77 (min 48016.04, max 48147.11)" + "sha512","256","5989.50 (min 5978.62, max 6010.03)" + "sha512","64","1682.55 (min 1675.90, max 1690.75)" + "sha512","8192","43149.99 (min 43092.65, max 43242.84)" + .. csv-table:: OpenSSL CPU Load :header: "Algorithm","am64xx-hsevm: CPU Load" - "aes-128-cbc","43.00" - "aes-128-ecb","45.00" - "aes-192-cbc","44.00" - "aes-192-ecb","44.00" - "aes-256-cbc","42.00" - "aes-256-ecb","44.00" + "aes-128-cbc","43.25 (min 43.00, max 44.00)" + "aes-128-ecb","45.00 (min 44.00, max 46.00)" + "aes-192-cbc","42.75 (min 42.00, max 44.00)" + "aes-192-ecb","44.25 (min 44.00, max 45.00)" + "aes-256-cbc","42.75 (min 42.00, max 44.00)" + "aes-256-ecb","43.75 (min 43.00, max 44.00)" "sha256","93.00" "sha512","93.00" + Listed for each algorithm are the code snippets used to run each benchmark test. :: time -v openssl speed -elapsed -evp aes-128-cbc + + +| diff --git a/source/linux/Foundational_Components/U-Boot/UG-Falcon-Mode.rst b/source/linux/Foundational_Components/U-Boot/UG-Falcon-Mode.rst index 2036345f5..4c9782e1d 100644 --- a/source/linux/Foundational_Components/U-Boot/UG-Falcon-Mode.rst +++ b/source/linux/Foundational_Components/U-Boot/UG-Falcon-Mode.rst @@ -1,3 +1,5 @@ +.. _U-Boot-Falcon-Mode: + ################## U-Boot Falcon Mode ################## diff --git a/source/linux/Overview/GCC_ToolChain.rst b/source/linux/Overview/GCC_ToolChain.rst index b90887398..9d3e1acf6 100644 --- a/source/linux/Overview/GCC_ToolChain.rst +++ b/source/linux/Overview/GCC_ToolChain.rst @@ -322,10 +322,10 @@ commands of the Foundational Components in this document. host# COMPILER_PATH=/opt/arm-toolchain host# mkdir -p $COMPILER_PATH - host# wget -c https://developer.arm.com/-/media/Files/downloads/gnu/11.3.rel1/binrel/arm-gnu-toolchain-11.3.rel1-x86_64-arm-none-linux-gnueabihf.tar.xz -O - | tar -xv -J -C $COMPILER_PATH - host# wget -c https://developer.arm.com/-/media/Files/downloads/gnu/11.3.rel1/binrel/arm-gnu-toolchain-11.3.rel1-x86_64-aarch64-none-linux-gnu.tar.xz -O - | tar -xv -J -C $COMPILER_PATH - host# export CROSS_COMPILE_64=$COMPILER_PATH/arm-gnu-toolchain-11.3.rel1-x86_64-aarch64-none-linux-gnu/bin/aarch64-none-linux-gnu- - host# export CROSS_COMPILE_32=$COMPILER_PATH/arm-gnu-toolchain-11.3.rel1-x86_64-arm-none-linux-gnueabihf/bin/arm-none-linux-gnueabihf- + host# wget -c https://developer.arm.com/-/media/Files/downloads/gnu/13.3.rel1/binrel/arm-gnu-toolchain-13.3.rel1-x86_64-arm-none-linux-gnueabihf.tar.xz -O - | tar -xv -J -C $COMPILER_PATH + host# wget -c https://developer.arm.com/-/media/Files/downloads/gnu/13.3.rel1/binrel/arm-gnu-toolchain-13.3.rel1-x86_64-aarch64-none-linux-gnu.tar.xz -O - | tar -xv -J -C $COMPILER_PATH + host# export CROSS_COMPILE_64=$COMPILER_PATH/arm-gnu-toolchain-13.3.rel1-x86_64-aarch64-none-linux-gnu/bin/aarch64-none-linux-gnu- + host# export CROSS_COMPILE_32=$COMPILER_PATH/arm-gnu-toolchain-13.3.rel1-x86_64-arm-none-linux-gnueabihf/bin/arm-none-linux-gnueabihf- host# export CC_64="${CROSS_COMPILE_64}gcc" host# export CC_32="${CROSS_COMPILE_32}gcc"