[libc] Polish GPU benchmarking #153900

leandrolcampos · 2025-08-15T23:17:39Z

This patch provides cleanups and improvements for the GPU benchmarking infrastructure. The key changes are:

Fix benchmark convergence bug: Round up the scaled iteration count (ceil) to ensure it grows properly. The previous truncation logic causes the iteration count to get stuck.
Resolve remaining compiler warning.
Remove unused BenchmarkLogger files: This is dead code that added maintenance and cognitive overhead without providing functionality.
Improve build hygiene: Clean up headers and CMake dependencies to strictly follow the 'include what you use' (IWYU) principle.

llvmbot · 2025-08-15T23:18:10Z

@llvm/pr-subscribers-backend-amdgpu

@llvm/pr-subscribers-libc

Author: Leandro Lacerda (leandrolcampos)

Changes

This patch provides cleanups and improvements for the GPU benchmarking infrastructure. The key changes are:

Fix benchmark convergence bug: Round up the scaled iteration count (ceil) to ensure it grows properly. The previous truncation logic causes the iteration count to get stuck.
Resolve remaining compiler warning.
Remove unused BenchmarkLogger files: This is dead code that added maintenance and cognitive overhead without providing functionality.
Improve build hygiene: Clean up headers and CMake dependencies to strictly follow the 'include what you use' (IWYU) principle.

Full diff: https://github.com/llvm/llvm-project/pull/153900.diff

9 Files Affected:

(removed) libc/benchmarks/gpu/BenchmarkLogger.cpp (-97)
(removed) libc/benchmarks/gpu/BenchmarkLogger.h (-29)
(modified) libc/benchmarks/gpu/CMakeLists.txt (+5-11)
(modified) libc/benchmarks/gpu/LibcGpuBenchmark.cpp (+11-7)
(modified) libc/benchmarks/gpu/LibcGpuBenchmark.h (+1-3)
(modified) libc/benchmarks/gpu/timing/amdgpu/CMakeLists.txt (+2-1)
(modified) libc/benchmarks/gpu/timing/amdgpu/timing.h (-1)
(modified) libc/benchmarks/gpu/timing/nvptx/CMakeLists.txt (+2-1)
(modified) libc/benchmarks/gpu/timing/nvptx/timing.h (+1-3)

diff --git a/libc/benchmarks/gpu/BenchmarkLogger.cpp b/libc/benchmarks/gpu/BenchmarkLogger.cpp
deleted file mode 100644
index d5996a74f6dd7..0000000000000
--- a/libc/benchmarks/gpu/BenchmarkLogger.cpp
+++ /dev/null
@@ -1,97 +0,0 @@
-#include "benchmarks/gpu/BenchmarkLogger.h"
-#include "hdr/stdint_proxy.h"
-#include "src/__support/CPP/string.h"
-#include "src/__support/CPP/string_view.h"
-#include "src/__support/OSUtil/io.h"               // write_to_stderr
-#include "src/__support/big_int.h"                 // is_big_int
-#include "src/__support/macros/config.h"
-#include "src/__support/macros/properties/types.h" // LIBC_TYPES_HAS_INT128
-#include "src/__support/uint128.h"
-
-namespace LIBC_NAMESPACE_DECL {
-namespace benchmarks {
-
-// cpp::string_view specialization
-template <>
-BenchmarkLogger &
-    BenchmarkLogger::operator<< <cpp::string_view>(cpp::string_view str) {
-  LIBC_NAMESPACE::write_to_stderr(str);
-  return *this;
-}
-
-// cpp::string specialization
-template <>
-BenchmarkLogger &BenchmarkLogger::operator<< <cpp::string>(cpp::string str) {
-  return *this << static_cast<cpp::string_view>(str);
-}
-
-// const char* specialization
-template <>
-BenchmarkLogger &BenchmarkLogger::operator<< <const char *>(const char *str) {
-  return *this << cpp::string_view(str);
-}
-
-// char* specialization
-template <> BenchmarkLogger &BenchmarkLogger::operator<< <char *>(char *str) {
-  return *this << cpp::string_view(str);
-}
-
-// char specialization
-template <> BenchmarkLogger &BenchmarkLogger::operator<<(char ch) {
-  return *this << cpp::string_view(&ch, 1);
-}
-
-// bool specialization
-template <> BenchmarkLogger &BenchmarkLogger::operator<<(bool cond) {
-  return *this << (cond ? "true" : "false");
-}
-
-// void * specialization
-template <> BenchmarkLogger &BenchmarkLogger::operator<<(void *addr) {
-  return *this << "0x" << cpp::to_string(reinterpret_cast<uintptr_t>(addr));
-}
-
-template <typename T> BenchmarkLogger &BenchmarkLogger::operator<<(T t) {
-  if constexpr (is_big_int_v<T> ||
-                (cpp::is_integral_v<T> && cpp::is_unsigned_v<T> &&
-                 (sizeof(T) > sizeof(uint64_t)))) {
-    static_assert(sizeof(T) % 8 == 0, "Unsupported size of UInt");
-    const IntegerToString<T, radix::Hex::WithPrefix> buffer(t);
-    return *this << buffer.view();
-  } else {
-    return *this << cpp::to_string(t);
-  }
-}
-
-// is_integral specializations
-// char is already specialized to handle character
-template BenchmarkLogger &BenchmarkLogger::operator<< <short>(short);
-template BenchmarkLogger &BenchmarkLogger::operator<< <int>(int);
-template BenchmarkLogger &BenchmarkLogger::operator<< <long>(long);
-template BenchmarkLogger &BenchmarkLogger::operator<< <long long>(long long);
-template BenchmarkLogger &
-    BenchmarkLogger::operator<< <unsigned char>(unsigned char);
-template BenchmarkLogger &
-    BenchmarkLogger::operator<< <unsigned short>(unsigned short);
-template BenchmarkLogger &
-    BenchmarkLogger::operator<< <unsigned int>(unsigned int);
-template BenchmarkLogger &
-    BenchmarkLogger::operator<< <unsigned long>(unsigned long);
-template BenchmarkLogger &
-    BenchmarkLogger::operator<< <unsigned long long>(unsigned long long);
-
-#ifdef LIBC_TYPES_HAS_INT128
-template BenchmarkLogger &
-    BenchmarkLogger::operator<< <__uint128_t>(__uint128_t);
-#endif // LIBC_TYPES_HAS_INT128
-template BenchmarkLogger &BenchmarkLogger::operator<< <UInt<128>>(UInt<128>);
-template BenchmarkLogger &BenchmarkLogger::operator<< <UInt<192>>(UInt<192>);
-template BenchmarkLogger &BenchmarkLogger::operator<< <UInt<256>>(UInt<256>);
-template BenchmarkLogger &BenchmarkLogger::operator<< <UInt<320>>(UInt<320>);
-
-// TODO: Add floating point formatting once it's supported by StringStream.
-
-BenchmarkLogger log;
-
-} // namespace benchmarks
-} // namespace LIBC_NAMESPACE_DECL
diff --git a/libc/benchmarks/gpu/BenchmarkLogger.h b/libc/benchmarks/gpu/BenchmarkLogger.h
deleted file mode 100644
index 2b22aba085f86..0000000000000
--- a/libc/benchmarks/gpu/BenchmarkLogger.h
+++ /dev/null
@@ -1,29 +0,0 @@
-//===-- Utilities to log to standard output during tests --------*- C++ -*-===//
-//
-// Part of the LLVM Project, under the Apache License v2.0 with LLVM Exceptions.
-// See https://llvm.org/LICENSE.txt for license information.
-// SPDX-License-Identifier: Apache-2.0 WITH LLVM-exception
-//
-//===----------------------------------------------------------------------===//
-
-#ifndef LLVM_LIBC_BENCHMARKS_GPU_BENCHMARKLOGGER_H
-#define LLVM_LIBC_BENCHMARKS_GPU_BENCHMARKLOGGER_H
-
-#include "src/__support/macros/config.h"
-
-namespace LIBC_NAMESPACE_DECL {
-namespace benchmarks {
-
-// A class to log to standard output in the context of hermetic tests.
-struct BenchmarkLogger {
-  constexpr BenchmarkLogger() = default;
-  template <typename T> BenchmarkLogger &operator<<(T);
-};
-
-// A global TestLogger instance to be used in tests.
-extern BenchmarkLogger log;
-
-} // namespace benchmarks
-} // namespace LIBC_NAMESPACE_DECL
-
-#endif /* LLVM_LIBC_BENCHMARKS_GPU_BENCHMARKLOGGER_H */
diff --git a/libc/benchmarks/gpu/CMakeLists.txt b/libc/benchmarks/gpu/CMakeLists.txt
index beedac78d4826..6ca134b12a479 100644
--- a/libc/benchmarks/gpu/CMakeLists.txt
+++ b/libc/benchmarks/gpu/CMakeLists.txt
@@ -38,31 +38,25 @@ add_unittest_framework_library(
   SRCS
     LibcGpuBenchmark.cpp
     LibcGpuBenchmarkMain.cpp
-    BenchmarkLogger.cpp
   HDRS
     LibcGpuBenchmark.h
-    BenchmarkLogger.h
   DEPENDS
+    libc.benchmarks.gpu.timing.timing
     libc.hdr.stdint_proxy
-    libc.src.__support.big_int
-    libc.src.__support.c_string
     libc.src.__support.CPP.string
     libc.src.__support.CPP.string_view
     libc.src.__support.CPP.type_traits
-    libc.src.__support.CPP.limits
     libc.src.__support.CPP.algorithm
     libc.src.__support.CPP.atomic
     libc.src.__support.CPP.array
-    libc.src.__support.fixed_point.fx_rep
-    libc.src.__support.macros.properties.types
-    libc.src.__support.OSUtil.osutil
-    libc.src.__support.uint128
     libc.src.__support.FPUtil.fp_bits
+    libc.src.__support.FPUtil.nearest_integer_operations
     libc.src.__support.FPUtil.sqrt
     libc.src.__support.fixedvector
-    libc.src.time.clock
-    libc.benchmarks.gpu.timing.timing
+    libc.src.__support.GPU.utils
+    libc.src.__support.time.gpu.time_utils
     libc.src.stdio.printf
+    libc.src.time.clock
 )
 
 add_subdirectory(src)
diff --git a/libc/benchmarks/gpu/LibcGpuBenchmark.cpp b/libc/benchmarks/gpu/LibcGpuBenchmark.cpp
index ef816c51a87d7..a4a0ff4ec46e5 100644
--- a/libc/benchmarks/gpu/LibcGpuBenchmark.cpp
+++ b/libc/benchmarks/gpu/LibcGpuBenchmark.cpp
@@ -2,16 +2,17 @@
 
 #include "hdr/stdint_proxy.h"
 #include "src/__support/CPP/algorithm.h"
-#include "src/__support/CPP/array.h"
 #include "src/__support/CPP/atomic.h"
 #include "src/__support/CPP/string.h"
 #include "src/__support/FPUtil/FPBits.h"
+#include "src/__support/FPUtil/NearestIntegerOperations.h"
 #include "src/__support/FPUtil/sqrt.h"
 #include "src/__support/GPU/utils.h"
 #include "src/__support/fixedvector.h"
 #include "src/__support/macros/config.h"
 #include "src/__support/time/gpu/time_utils.h"
 #include "src/stdio/printf.h"
+#include "src/time/clock.h"
 
 namespace LIBC_NAMESPACE_DECL {
 namespace benchmarks {
@@ -134,11 +135,13 @@ void print_results(Benchmark *b) {
   cpp::atomic_thread_fence(cpp::MemoryOrder::RELEASE);
 
   LIBC_NAMESPACE::printf(
-      "%-24s |%15.0f |%9.0f |%8llu |%8llu |%11llu |%9u |\n",
+      "%-24s |%15.0f |%9.0f |%8llu |%8llu |%15llu |%9u |\n",
       b->get_test_name().data(), final_result.cycles,
-      final_result.standard_deviation, (unsigned long long)final_result.min,
-      (unsigned long long)final_result.max,
-      (unsigned long long)final_result.total_iterations, (unsigned)num_threads);
+      final_result.standard_deviation,
+      static_cast<unsigned long long>(final_result.min),
+      static_cast<unsigned long long>(final_result.max),
+      static_cast<unsigned long long>(final_result.total_iterations),
+      static_cast<unsigned>(num_threads));
 }
 
 void print_header() {
@@ -147,7 +150,7 @@ void print_header() {
                          benchmarks[0]->get_suite_name().data());
   LIBC_NAMESPACE::printf("%s", RESET);
   cpp::string titles = "Benchmark                |  Cycles (Mean) |   Stddev | "
-                       "    Min |     Max | Iterations |  Threads |\n";
+                       "    Min |     Max |     Iterations |  Threads |\n";
   LIBC_NAMESPACE::printf(titles.data());
 
   cpp::string separator(titles.size(), '-');
@@ -226,7 +229,8 @@ BenchmarkResult benchmark(const BenchmarkOptions &options,
         change_ratio < options.epsilon)
       break;
 
-    iterations = static_cast<uint32_t>(iterations * options.scaling_factor);
+    iterations = static_cast<uint32_t>(
+        fputil::ceil(iterations * options.scaling_factor));
   }
 
   const auto &estimator = rep.get_estimator();
diff --git a/libc/benchmarks/gpu/LibcGpuBenchmark.h b/libc/benchmarks/gpu/LibcGpuBenchmark.h
index 60f69edf86556..e36e93c7efc18 100644
--- a/libc/benchmarks/gpu/LibcGpuBenchmark.h
+++ b/libc/benchmarks/gpu/LibcGpuBenchmark.h
@@ -1,18 +1,16 @@
 #ifndef LLVM_LIBC_BENCHMARKS_LIBC_GPU_BENCHMARK_H
 #define LLVM_LIBC_BENCHMARKS_LIBC_GPU_BENCHMARK_H
 
-#include "benchmarks/gpu/BenchmarkLogger.h"
 #include "benchmarks/gpu/timing/timing.h"
+
 #include "hdr/stdint_proxy.h"
 #include "src/__support/CPP/algorithm.h"
 #include "src/__support/CPP/array.h"
-#include "src/__support/CPP/limits.h"
 #include "src/__support/CPP/string_view.h"
 #include "src/__support/CPP/type_traits.h"
 #include "src/__support/FPUtil/FPBits.h"
 #include "src/__support/FPUtil/sqrt.h"
 #include "src/__support/macros/config.h"
-#include "src/time/clock.h"
 
 namespace LIBC_NAMESPACE_DECL {
 
diff --git a/libc/benchmarks/gpu/timing/amdgpu/CMakeLists.txt b/libc/benchmarks/gpu/timing/amdgpu/CMakeLists.txt
index d6a89d04dab97..f85152e69c346 100644
--- a/libc/benchmarks/gpu/timing/amdgpu/CMakeLists.txt
+++ b/libc/benchmarks/gpu/timing/amdgpu/CMakeLists.txt
@@ -4,10 +4,11 @@ add_header_library(
     timing.h
   DEPENDS
     libc.hdr.stdint_proxy
-    libc.src.__support.common
     libc.src.__support.macros.config
     libc.src.__support.macros.attributes
     libc.src.__support.CPP.algorithm
     libc.src.__support.CPP.array
+    libc.src.__support.CPP.atomic
     libc.src.__support.CPP.type_traits
+    libc.src.__support.GPU.utils
 )
diff --git a/libc/benchmarks/gpu/timing/amdgpu/timing.h b/libc/benchmarks/gpu/timing/amdgpu/timing.h
index de721a2d6ce6b..b4a174f729817 100644
--- a/libc/benchmarks/gpu/timing/amdgpu/timing.h
+++ b/libc/benchmarks/gpu/timing/amdgpu/timing.h
@@ -15,7 +15,6 @@
 #include "src/__support/CPP/atomic.h"
 #include "src/__support/CPP/type_traits.h"
 #include "src/__support/GPU/utils.h"
-#include "src/__support/common.h"
 #include "src/__support/macros/attributes.h"
 #include "src/__support/macros/config.h"
 
diff --git a/libc/benchmarks/gpu/timing/nvptx/CMakeLists.txt b/libc/benchmarks/gpu/timing/nvptx/CMakeLists.txt
index 801080e7a6e98..4615f53e3d247 100644
--- a/libc/benchmarks/gpu/timing/nvptx/CMakeLists.txt
+++ b/libc/benchmarks/gpu/timing/nvptx/CMakeLists.txt
@@ -4,10 +4,11 @@ add_header_library(
     timing.h
   DEPENDS
     libc.hdr.stdint_proxy
-    libc.src.__support.common
     libc.src.__support.macros.config
     libc.src.__support.macros.attributes
     libc.src.__support.CPP.algorithm
     libc.src.__support.CPP.array
+    libc.src.__support.CPP.atomic
     libc.src.__support.CPP.type_traits
+    libc.src.__support.GPU.utils
 )
diff --git a/libc/benchmarks/gpu/timing/nvptx/timing.h b/libc/benchmarks/gpu/timing/nvptx/timing.h
index 133032ca08423..0c93a67129b8d 100644
--- a/libc/benchmarks/gpu/timing/nvptx/timing.h
+++ b/libc/benchmarks/gpu/timing/nvptx/timing.h
@@ -13,9 +13,7 @@
 #include "src/__support/CPP/algorithm.h"
 #include "src/__support/CPP/array.h"
 #include "src/__support/CPP/atomic.h"
-#include "src/__support/CPP/type_traits.h"
 #include "src/__support/GPU/utils.h"
-#include "src/__support/common.h"
 #include "src/__support/macros/attributes.h"
 #include "src/__support/macros/config.h"
 
@@ -66,7 +64,7 @@ template <typename F, typename T>
   uint64_t stop = gpu::processor_clock();
   cpp::atomic_thread_fence(cpp::MemoryOrder::ACQ_REL);
   asm("" ::"r"(stop));
-  volatile T output = result;
+  volatile auto output = result;
 
   // Return the time elapsed.
   return stop - start;

leandrolcampos · 2025-08-15T23:20:04Z

Here's what I get on my NVIDIA GeForce RTX 4070 Laptop GPU.

Note that the Iterations column now shows the global total: this makes the column consistent with the global Cycles (Mean) and Stddev.

[1/4] Running hermetic test libc.benchmarks.gpu.src.ctype.isalnum_benchmark
Running Suite: LlvmLibcIsAlNumGpuBenchmark
Benchmark                |  Cycles (Mean) |   Stddev |     Min |     Max |     Iterations |  Threads |
------------------------------------------------------------------------------------------------------
IsAlnum                  |             53 |        0 |      53 |      53 |          11904 |       64 |
IsAlnumSingleThread      |             53 |        0 |      53 |      53 |            186 |        1 |
IsAlnumSingleWave        |             53 |        0 |      53 |      53 |           5952 |       32 |
IsAlnumCapital           |             53 |        0 |      53 |      53 |          11904 |       64 |
IsAlnumNotAlnum          |             43 |        0 |      43 |      43 |          11904 |       64 |
[2/4] Running hermetic test libc.benchmarks.gpu.src.ctype.isalpha_benchmark
Running Suite: LlvmLibcIsAlphaGpuBenchmark
Benchmark                |  Cycles (Mean) |   Stddev |     Min |     Max |     Iterations |  Threads |
------------------------------------------------------------------------------------------------------
IsAlpha                  |             53 |        0 |      53 |      53 |            186 |        1 |
[3/4] Running hermetic test libc.benchmarks.gpu.src.math.sin_benchmark
Running Suite: LlvmLibcSinGpuBenchmark
Benchmark                |  Cycles (Mean) |   Stddev |     Min |     Max |     Iterations |  Threads |
------------------------------------------------------------------------------------------------------
Sin_1                    |           3120 |      153 |    2934 |    3610 |        2735008 |       32 |
Sin_128                  |            364 |        6 |     346 |     382 |          12032 |       32 |
Sin_1024                 |            354 |        2 |     348 |     359 |           5952 |       32 |
Sin_4096                 |            361 |        1 |     358 |     363 |           4160 |       32 |
SinTwoPi_1               |           2217 |       11 |    2204 |    2494 |          24032 |       32 |
SinTwoPi_128             |            262 |        2 |     258 |     268 |          24032 |       32 |
SinTwoPi_1024            |            271 |        1 |     271 |     276 |          24032 |       32 |
SinTwoPi_4096            |            282 |        0 |     281 |     282 |           1984 |       32 |
SinTwoPow30_1            |           3133 |       11 |    3116 |    3177 |           1984 |       32 |
SinTwoPow30_128          |            348 |        2 |     345 |     352 |           1344 |       32 |
SinTwoPow30_1024         |            359 |        1 |     358 |     361 |           1984 |       32 |
SinTwoPow30_4096         |            368 |        0 |     367 |     368 |            576 |       32 |
SinVeryLarge_1           |           2855 |       15 |    2821 |    3082 |           8480 |       32 |
SinVeryLarge_128         |            316 |        2 |     310 |     324 |          17024 |       32 |
SinVeryLarge_1024        |            317 |        1 |     316 |     320 |           1344 |       32 |
SinVeryLarge_4096        |            325 |        0 |     325 |     326 |            896 |       32 |
NvSin_1                  |           2521 |       69 |    2264 |    2876 |           5952 |       32 |
NvSin_128                |           1860 |        2 |    1857 |    1865 |            576 |       32 |
NvSin_1024               |           2069 |        1 |    2068 |    2071 |            352 |       32 |
NvSin_4096               |           2087 |        0 |    2087 |    2088 |            352 |       32 |
NvSinTwoPi_1             |           1106 |        1 |    1104 |    1108 |           1344 |       32 |
NvSinTwoPi_128           |            925 |        0 |     925 |     925 |            352 |       32 |
NvSinTwoPi_1024          |           1136 |        0 |    1136 |    1136 |            352 |       32 |
NvSinTwoPi_4096          |           1155 |        0 |    1155 |    1155 |            352 |       32 |
NvSinTwoPow30_1          |           1106 |        1 |    1104 |    1107 |           1984 |       32 |
NvSinTwoPow30_128        |            925 |        0 |     925 |     927 |            576 |       32 |
NvSinTwoPow30_1024       |           1136 |        0 |    1136 |    1136 |            352 |       32 |
NvSinTwoPow30_4096       |           1155 |        0 |    1155 |    1155 |            352 |       32 |
NvSinVeryLarge_1         |           2496 |       23 |    2251 |    2840 |          12032 |       32 |
NvSinVeryLarge_128       |           1827 |        1 |    1826 |    1829 |            576 |       32 |
NvSinVeryLarge_1024      |           2034 |        0 |    2034 |    2035 |            352 |       32 |
NvSinVeryLarge_4096      |           2052 |        0 |    2052 |    2052 |            352 |       32 |
Sinf_1                   |           2200 |      170 |    1521 |    2397 |         507776 |       32 |
Sinf_128                 |            240 |        5 |     218 |     255 |          47616 |       32 |
Sinf_1024                |            240 |        2 |     235 |     248 |           5952 |       32 |
Sinf_4096                |            259 |        1 |     257 |     261 |            896 |       32 |
SinfTwoPi_1              |           1441 |       11 |    1425 |    1760 |          33856 |       32 |
SinfTwoPi_128            |            147 |        1 |     146 |     150 |          12032 |       32 |
SinfTwoPi_1024           |            147 |        0 |     145 |     148 |           4160 |       32 |
SinfTwoPi_4096           |            165 |        0 |     165 |     167 |           4160 |       32 |
SinfTwoPow30_1           |           1087 |       10 |    1078 |    1160 |           1984 |       32 |
SinfTwoPow30_128         |            102 |        1 |     101 |     104 |            576 |       32 |
SinfTwoPow30_1024        |            102 |        0 |     100 |     103 |           2880 |       32 |
SinfTwoPow30_4096        |            121 |        0 |     121 |     122 |           4160 |       32 |
SinfVeryLarge_1          |           1924 |       14 |    1867 |    2277 |          24032 |       32 |
SinfVeryLarge_128        |            205 |        1 |     205 |     207 |            896 |       32 |
SinfVeryLarge_1024       |            206 |        0 |     204 |     206 |           5952 |       32 |
SinfVeryLarge_4096       |            225 |        0 |     224 |     226 |           8480 |       32 |
NvSinf_1                 |           1023 |        6 |    1018 |    1036 |           2880 |       32 |
NvSinf_128               |            786 |        0 |     786 |     788 |            576 |       32 |
NvSinf_1024              |            975 |        1 |     970 |     975 |           2880 |       32 |
NvSinf_4096              |           1010 |        0 |    1010 |    1010 |            352 |       32 |
NvSinfTwoPi_1            |            162 |        3 |     162 |     509 |         362464 |       32 |
NvSinfTwoPi_128          |            141 |        0 |     141 |     143 |           2880 |       32 |
NvSinfTwoPi_1024         |            331 |        0 |     330 |     331 |            896 |       32 |
NvSinfTwoPi_4096         |            366 |        0 |     366 |     366 |            352 |       32 |
NvSinfTwoPow30_1         |           1025 |       10 |    1018 |    1280 |          47616 |       32 |
NvSinfTwoPow30_128       |            776 |        0 |     776 |     777 |            352 |       32 |
NvSinfTwoPow30_1024      |            966 |        0 |     966 |     967 |            352 |       32 |
NvSinfTwoPow30_4096      |           1000 |        0 |    1000 |    1000 |            352 |       32 |
NvSinfVeryLarge_1        |           1002 |        2 |    1000 |    1023 |           4160 |       32 |
NvSinfVeryLarge_128      |            757 |        0 |     757 |     758 |            576 |       32 |
NvSinfVeryLarge_1024     |            948 |        0 |     948 |     948 |            352 |       32 |
NvSinfVeryLarge_4096     |            981 |        0 |     981 |     981 |            352 |       32 |
[4/4] Running hermetic test libc.benchmarks.gpu.src.math.atan2_benchmark
Running Suite: LlvmLibcAtan2GpuBenchmark
Benchmark                |  Cycles (Mean) |   Stddev |     Min |     Max |     Iterations |  Threads |
------------------------------------------------------------------------------------------------------
Atan2_1                  |           4081 |      954 |    1892 |    5264 |          24032 |       32 |
Atan2_128                |           2505 |       37 |    2346 |    2601 |          33856 |       32 |
Atan2_1024               |           2747 |       13 |    2708 |    2783 |           2880 |       32 |
Atan2_4096               |           2749 |        6 |    2737 |    2759 |            576 |       32 |
Atan2TwoPi_1             |           2737 |       16 |    2728 |    3156 |          24032 |       32 |
Atan2TwoPi_128           |           1068 |        3 |    1064 |    1109 |           8480 |       32 |
Atan2TwoPi_1024          |           1301 |        1 |    1300 |    1303 |            576 |       32 |
Atan2TwoPi_4096          |           1302 |        0 |    1302 |    1303 |            576 |       32 |
Atan2TwoPow30_1          |           2733 |       16 |    2722 |    3157 |          24032 |       32 |
Atan2TwoPow30_128        |           1072 |        3 |    1065 |    1108 |           5952 |       32 |
Atan2TwoPow30_1024       |           1302 |        1 |    1301 |    1304 |            896 |       32 |
Atan2TwoPow30_4096       |           1303 |        0 |    1302 |    1303 |            352 |       32 |
Atan2Large_1             |           3569 |      382 |    1125 |    3878 |         131648 |       32 |
Atan2Large_128           |           1811 |       17 |    1766 |    1839 |           1984 |       32 |
Atan2Large_1024          |           2050 |        4 |    2043 |    2059 |            576 |       32 |
Atan2Large_4096          |           2049 |        2 |    2046 |    2053 |            352 |       32 |
NvAtan2_1                |           2909 |       38 |    2866 |    3341 |          17024 |       32 |
NvAtan2_128              |           2836 |        3 |    2830 |    2853 |           2880 |       32 |
NvAtan2_1024             |           3075 |        1 |    3074 |    3076 |            352 |       32 |
NvAtan2_4096             |           3075 |        1 |    3074 |    3076 |            352 |       32 |
NvAtan2TwoPi_1           |           2032 |       13 |    2032 |    2383 |          24032 |       32 |
NvAtan2TwoPi_128         |           1979 |        3 |    1979 |    1999 |           1984 |       32 |
NvAtan2TwoPi_1024        |           2218 |        0 |    2217 |    2218 |            352 |       32 |
NvAtan2TwoPi_4096        |           2218 |        0 |    2218 |    2218 |            352 |       32 |
NvAtan2TwoPow30_1        |           2032 |        8 |    2032 |    2184 |          12032 |       32 |
NvAtan2TwoPow30_128      |           1979 |        2 |    1979 |    1998 |           1984 |       32 |
NvAtan2TwoPow30_1024     |           2218 |        0 |    2217 |    2218 |            352 |       32 |
NvAtan2TwoPow30_4096     |           2218 |        0 |    2218 |    2219 |            576 |       32 |
NvAtan2Large_1           |           2032 |       12 |    2032 |    2354 |          24032 |       32 |
NvAtan2Large_128         |           1979 |        2 |    1979 |    1997 |           1984 |       32 |
NvAtan2Large_1024        |           2218 |        0 |    2217 |    2218 |            352 |       32 |
NvAtan2Large_4096        |           2219 |        0 |    2218 |    2219 |            352 |       32 |

jhuber6 · 2025-08-15T23:22:00Z

libc/benchmarks/gpu/timing/nvptx/timing.h

@@ -66,7 +64,7 @@ template <typename F, typename T>
  uint64_t stop = gpu::processor_clock();
  cpp::atomic_thread_fence(cpp::MemoryOrder::ACQ_REL);
  asm("" ::"r"(stop));
-  volatile T output = result;
+  volatile auto output = result;


Weird that we need this here and not in the AMDGPU version, seems weird that these wouldn't be the same type. What were you seeing here?

The reason I had to touch the NVPTX version (and not the AMDGPU one) is simply that only the NVPTX latency() had this instruction at the end:

volatile T output = result;

In the ctype benches, we instantiate latency<int (*)(int), char>. The function returns int, but the template parameter T (the input type) is char. That line therefore assigns an int to a volatile char, which produces:

[4/15] Building CXX object libc/benchmarks/gpu/src/ctype/CMakeFiles/libc.benchmarks.gpu.src.ctype.isalnum_benchmark.__build__.dir/isalnum_benchmark.cpp.o In file included from /home/leandro/llvm-project/libc/benchmarks/gpu/src/ctype/isalnum_benchmark.cpp:1: In file included from /home/leandro/llvm-project/libc/benchmarks/gpu/LibcGpuBenchmark.h:4: In file included from /home/leandro/llvm-project/libc/benchmarks/gpu/timing/timing.h:17: /home/leandro/llvm-project/libc/benchmarks/gpu/timing/nvptx/timing.h:67:23: warning: implicit conversion loses integer precision: 'int' to 'volatile char' [-Wimplicit-int-conversion] 67 | volatile T output = result; | ~~~~~~ ^~~~~~ /home/leandro/llvm-project/libc/benchmarks/gpu/src/ctype/isalnum_benchmark.cpp:7:26: note: in instantiation of function template specialization '__llvm_libc_22_0_0_git::latency<int (*)(int), char>' requested here 7 | return LIBC_NAMESPACE::latency(LIBC_NAMESPACE::isalnum, x); | ^ 1 warning generated. [6/15] Building CXX object libc/benchmarks/gpu/src/ctype/CMakeFiles/libc.benchmarks.gpu.src.ctype.isalpha_benchmark.__build__.dir/isalpha_benchmark.cpp.o In file included from /home/leandro/llvm-project/libc/benchmarks/gpu/src/ctype/isalpha_benchmark.cpp:1: In file included from /home/leandro/llvm-project/libc/benchmarks/gpu/LibcGpuBenchmark.h:4: In file included from /home/leandro/llvm-project/libc/benchmarks/gpu/timing/timing.h:17: /home/leandro/llvm-project/libc/benchmarks/gpu/timing/nvptx/timing.h:67:23: warning: implicit conversion loses integer precision: 'int' to 'volatile char' [-Wimplicit-int-conversion] 67 | volatile T output = result; | ~~~~~~ ^~~~~~ /home/leandro/llvm-project/libc/benchmarks/gpu/src/ctype/isalpha_benchmark.cpp:7:26: note: in instantiation of function template specialization '__llvm_libc_22_0_0_git::latency<int (*)(int), char>' requested here 7 | return LIBC_NAMESPACE::latency(LIBC_NAMESPACE::isalpha, x); | ^ 1 warning generated.

On AMDGPU, there is no such instruction in latency() (it seems to rely on the v_or_b32 asm to use the value), and it also has a small-type special-case for the asm operand (char/bool), so there’s no narrowing assignment there.

I didn’t dive deeper here because latency() isn’t used by the math benchmarks; I only wanted to make the ctype benches warning-free.

I vaguely remember needing that because the PTX optimizer would defy the inline assembly so I still needed it. That's fine it needed.

leandrolcampos added 4 commits August 15, 2025 19:42

Clean up headers and CMake deps

45e974b

Remove dead BenchmarkLogger

607f4e7

Fix precision-loss warning in NVPTX version of latency()

6642c9a

Round up scaled iteration count to ensure growth

71dd568

llvmbot added backend:AMDGPU libc labels Aug 15, 2025

jhuber6 reviewed Aug 15, 2025

View reviewed changes

jhuber6 approved these changes Aug 15, 2025

View reviewed changes

jhuber6 merged commit cf5f311 into llvm:main Aug 16, 2025
22 checks passed

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Uh oh!

[libc] Polish GPU benchmarking #153900

[libc] Polish GPU benchmarking #153900

Uh oh!

leandrolcampos commented Aug 15, 2025

Uh oh!

llvmbot commented Aug 15, 2025 •

edited

Loading

Uh oh!

leandrolcampos commented Aug 15, 2025 •

edited

Loading

Uh oh!

jhuber6 Aug 15, 2025

Uh oh!

leandrolcampos Aug 15, 2025

Uh oh!

jhuber6 Aug 15, 2025

Uh oh!

Uh oh!

Uh oh!

[libc] Polish GPU benchmarking #153900

[libc] Polish GPU benchmarking #153900

Uh oh!

Conversation

leandrolcampos commented Aug 15, 2025

Uh oh!

llvmbot commented Aug 15, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

leandrolcampos commented Aug 15, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

jhuber6 Aug 15, 2025

Choose a reason for hiding this comment

Uh oh!

leandrolcampos Aug 15, 2025

Choose a reason for hiding this comment

Uh oh!

jhuber6 Aug 15, 2025

Choose a reason for hiding this comment

Uh oh!

Uh oh!

Uh oh!

llvmbot commented Aug 15, 2025 •

edited

Loading

leandrolcampos commented Aug 15, 2025 •

edited

Loading