Commit 04a06545 authored by Andreas Marek's avatar Andreas Marek

Correct stripe_width in complex case

parent 24921892
...@@ -3273,7 +3273,7 @@ ...@@ -3273,7 +3273,7 @@
#endif /* WITH_OPENMP */ #endif /* WITH_OPENMP */
if (.not.(useGPU)) then if (.not.(useGPU)) then
#ifdef DOUBLE_PRECISION_COMPLEX #ifdef DOUBLE_PRECISION_COMPLEX
stripe_width = ((stripe_width+1)/2)*2 ! Must be a multiple of 2 because of AVX/SSE memory alignment of 32 bytes stripe_width = ((stripe_width+3)/4)*4 ! Must be a multiple of 2 because of AVX/SSE memory alignment of 32 bytes
! (2 * sizeof(double complex) == 32) ! (2 * sizeof(double complex) == 32)
#else #else
stripe_width = ((stripe_width+3)/4)*4 ! Must be a multiple of 4 because of AVX/SSE memory alignment of 32 bytes stripe_width = ((stripe_width+3)/4)*4 ! Must be a multiple of 4 because of AVX/SSE memory alignment of 32 bytes
......
Markdown is supported
0% or .
You are about to add 0 people to the discussion. Proceed with caution.
Finish editing this message first!
Please register or to comment