From 209b63197ec752f43c1bb0141a95f0dffb42254d Mon Sep 17 00:00:00 2001 From: Werner Saar Date: Wed, 11 Jan 2017 07:29:17 +0100 Subject: [PATCH] prepared lapack/lauum for UNROLL values, that are not a power of two --- lapack/lauum/lauum_L_parallel.c | 2 +- lapack/lauum/lauum_U_parallel.c | 2 +- 2 files changed, 2 insertions(+), 2 deletions(-) diff --git a/lapack/lauum/lauum_L_parallel.c b/lapack/lauum/lauum_L_parallel.c index c93c4a8..0ebe3f0 100644 --- a/lapack/lauum/lauum_L_parallel.c +++ b/lapack/lauum/lauum_L_parallel.c @@ -88,7 +88,7 @@ blasint CNAME(blas_arg_t *args, BLASLONG *range_m, BLASLONG *range_n, FLOAT *sa, newarg.beta = NULL; newarg.nthreads = args -> nthreads; - blocking = (n / 2 + GEMM_UNROLL_N - 1) & ~(GEMM_UNROLL_N - 1); + blocking = ((n / 2 + GEMM_UNROLL_N - 1)/GEMM_UNROLL_N) * GEMM_UNROLL_N; if (blocking > GEMM_Q) blocking = GEMM_Q; for (i = 0; i < n; i += blocking) { diff --git a/lapack/lauum/lauum_U_parallel.c b/lapack/lauum/lauum_U_parallel.c index e4a2792..7214c97 100644 --- a/lapack/lauum/lauum_U_parallel.c +++ b/lapack/lauum/lauum_U_parallel.c @@ -88,7 +88,7 @@ blasint CNAME(blas_arg_t *args, BLASLONG *range_m, BLASLONG *range_n, FLOAT *sa, newarg.beta = NULL; newarg.nthreads = args -> nthreads; - blocking = (n / 2 + GEMM_UNROLL_N - 1) & ~(GEMM_UNROLL_N - 1); + blocking = ((n / 2 + GEMM_UNROLL_N - 1)/GEMM_UNROLL_N) * GEMM_UNROLL_N; if (blocking > GEMM_Q) blocking = GEMM_Q; for (i = 0; i < n; i += blocking) { -- 2.7.4