diff options
author | qiyu8 <qiyu8@foxmail.com> | 2020-06-09 11:41:44 +0800 |
---|---|---|
committer | qiyu8 <qiyu8@foxmail.com> | 2020-06-09 11:41:44 +0800 |
commit | 144919327e360e275a42054e601b3d157d20700d (patch) | |
tree | bc6f0e164630bfded994abdc3c2b9e4d5cd149a3 | |
parent | 4503c2965a6941ecc744a5fc671f3f99d2aab901 (diff) | |
download | numpy-144919327e360e275a42054e601b3d157d20700d.tar.gz |
test sse1 optimization
-rw-r--r-- | numpy/core/src/multiarray/einsum.c.src | 3 |
1 files changed, 2 insertions, 1 deletions
diff --git a/numpy/core/src/multiarray/einsum.c.src b/numpy/core/src/multiarray/einsum.c.src index 9c6913788..1eca7d751 100644 --- a/numpy/core/src/multiarray/einsum.c.src +++ b/numpy/core/src/multiarray/einsum.c.src @@ -546,7 +546,7 @@ finish_after_unrolled_loop: return; } -#if EINSUM_USE_SSE1 && @float32@ +#if EINSUM_USE_SSE3 && @float32@ value1_sse = _mm_set_ps1(value1); /* Use aligned instructions if possible */ @@ -570,6 +570,7 @@ finish_after_unrolled_loop: goto finish_after_unrolled_loop; } #elif EINSUM_USE_SSE2 && @float64@ + printf("using sse2\n"); value1_sse = _mm_set1_pd(value1); /* Use aligned instructions if possible */ |