debug

Author: mikebo93

chapter_accelerator/Performance_Optimization_Methods.md

@@ -93,15 +93,4 @@ used to compute the entire matrix $C$.
 Eigen is used to generate data and compute the GEMM result on the CPU.
 In addition, error computing and time profiling code are implemented for
 the GPU computing result. For details, see
-[first_attempt.cu](https://github.com/openmlsys/openmlsys-cuda/blob/main/first_attempt.cu).
-After the program is compiled and executed, output results are as
-follows:
 
-    Average time: 48.961 ms
-    Max error: 0.000092
-
-The peak GPU throughput can be approximated by using the following
-formula: 2 $\times$ Frequency $\times$ Number of single-precision
-compute units. The number of single-precision compute units equals the
-number of SMs in the GPU multiplied by the number of single-precision
-compute units in each SM. The results are as follows: