nakagawa-fj:feature/gemm_divide_rate_for_A64FX - Branch - OpenMathLib/OpenBLAS

Multi-thread Performance Improvement of GEMM with DIVIDE_RATE=1 for A64FX

#5353Merged

Comparing

nakagawa-fj:feature/gemm_divide_rate_for_A64FX

(

5253c8f

) with

develop

(

8f0a1a3

)

Untouched: 62

Benchmarks

62 total

test_dgbmv[1-100-s]

benchmark/pybench/benchmarks/bench_blas.py

36.9 µs36.8 µs

test_gesv[100-d]

benchmark/pybench/benchmarks/bench_blas.py

395.1 µs394.1 µs

test_syrk[100-d]

benchmark/pybench/benchmarks/bench_blas.py

340.8 µs340.1 µs

test_nrm2[1000-dz]

benchmark/pybench/benchmarks/bench_blas.py

35.1 µs35 µs

test_syrk[100-s]

benchmark/pybench/benchmarks/bench_blas.py

213.2 µs213 µs

test_syrk[100-c]

benchmark/pybench/benchmarks/bench_blas.py

473.3 µs473.1 µs

test_dgemv[1000-s]

benchmark/pybench/benchmarks/bench_blas.py

7 ms7 ms

test_gesdd[mn1-d]

benchmark/pybench/benchmarks/bench_blas.py

93.8 ms93.8 ms

test_gesv[1000-d]

benchmark/pybench/benchmarks/bench_blas.py

93.3 ms93.3 ms

test_gesv[100-s]

benchmark/pybench/benchmarks/bench_blas.py

257.3 µs257.2 µs

test_gemm[100-d]

benchmark/pybench/benchmarks/bench_blas.py

470.9 µs470.8 µs

test_dgbmv[1-1000-c]

benchmark/pybench/benchmarks/bench_blas.py

99.1 µs99.1 µs

test_syev[200-s]

benchmark/pybench/benchmarks/bench_blas.py

49.1 ms49.1 ms

test_dgemv[100-c]

benchmark/pybench/benchmarks/bench_blas.py

148.6 µs148.6 µs

test_gemm[1000-s]

benchmark/pybench/benchmarks/bench_blas.py

117.4 ms117.4 ms

test_syrk[1000-s]

benchmark/pybench/benchmarks/bench_blas.py

65.4 ms65.4 ms

test_syrk[1000-c]

benchmark/pybench/benchmarks/bench_blas.py

227.5 ms227.5 ms

test_syrk[1000-d]

benchmark/pybench/benchmarks/bench_blas.py

130.4 ms130.4 ms

test_gesv[1000-c]

benchmark/pybench/benchmarks/bench_blas.py

188.6 ms188.6 ms

test_syrk[1000-z]

benchmark/pybench/benchmarks/bench_blas.py

476.4 ms476.4 ms

test_gemm[1000-d]

benchmark/pybench/benchmarks/bench_blas.py

239.4 ms239.4 ms

test_dgemv[1000-z]

benchmark/pybench/benchmarks/bench_blas.py

26.3 ms26.3 ms

test_gesv[1000-z]

benchmark/pybench/benchmarks/bench_blas.py

353.6 ms353.6 ms

test_gesdd[mn1-s]

benchmark/pybench/benchmarks/bench_blas.py

65.2 ms65.2 ms

test_dgemv[1000-d]

benchmark/pybench/benchmarks/bench_blas.py

13.9 ms13.9 ms

Commits

Click on a commit to change the comparison range

Base

develop

8f0a1a3

-0.13%

Multi-thread Performance Improvement of GEMM with DIVIDE_RATE=1 for

5253c8f

11 months ago

by nakagawa-fj

Home Terms Privacy Docs