From e4cabb977d038688688aca39b366f98f9c36b7eb Mon Sep 17 00:00:00 2001 From: RuQing Xu Date: Thu, 16 Sep 2021 01:34:26 +0900 Subject: [PATCH] Arm SVE Typo Fix ZGEMM/CGEMM C Prefetch Reg --- .../3/bli_gemm_armsve_asm_c2vx10_unindexed.c | 18 +++++++++--------- .../3/bli_gemm_armsve_asm_z2vx10_unindexed.c | 18 +++++++++--------- .../3/bli_gemm_armsve_asm_z2vx7_unindexed.c | 12 ++++++------ .../3/bli_gemm_armsve_asm_z2vx8_unindexed.c | 14 +++++++------- 4 files changed, 31 insertions(+), 31 deletions(-) diff --git a/kernels/armsve/3/bli_gemm_armsve_asm_c2vx10_unindexed.c b/kernels/armsve/3/bli_gemm_armsve_asm_c2vx10_unindexed.c index ffaf13992..2f29075ab 100644 --- a/kernels/armsve/3/bli_gemm_armsve_asm_c2vx10_unindexed.c +++ b/kernels/armsve/3/bli_gemm_armsve_asm_c2vx10_unindexed.c @@ -116,23 +116,23 @@ GEMM_ACOLCMPLX_CONTIGUOUS_LOAD_FWD(z28,z29,p0,%0,x2) " b.ne END_CCOL_PRFM \n\t" // Do not prefetch for generic C storage. " mov x16, %2 \n\t" " prfm PLDL1KEEP, [x16] \n\t" -" add x16, x16, %5 \n\t" +" add x16, x16, %4 \n\t" " prfm PLDL1KEEP, [x16] \n\t" -" add x16, x16, %5 \n\t" +" add x16, x16, %4 \n\t" " prfm PLDL1KEEP, [x16] \n\t" -" add x16, x16, %5 \n\t" +" add x16, x16, %4 \n\t" " prfm PLDL1KEEP, [x16] \n\t" -" add x16, x16, %5 \n\t" +" add x16, x16, %4 \n\t" " prfm PLDL1KEEP, [x16] \n\t" -" add x16, x16, %5 \n\t" +" add x16, x16, %4 \n\t" " prfm PLDL1KEEP, [x16] \n\t" -" add x16, x16, %5 \n\t" +" add x16, x16, %4 \n\t" " prfm PLDL1KEEP, [x16] \n\t" -" add x16, x16, %5 \n\t" +" add x16, x16, %4 \n\t" " prfm PLDL1KEEP, [x16] \n\t" -" add x16, x16, %5 \n\t" +" add x16, x16, %4 \n\t" " prfm PLDL1KEEP, [x16] \n\t" -" add x16, x16, %5 \n\t" +" add x16, x16, %4 \n\t" " prfm PLDL1KEEP, [x16] \n\t" " END_CCOL_PRFM: \n\t" " \n\t" diff --git a/kernels/armsve/3/bli_gemm_armsve_asm_z2vx10_unindexed.c b/kernels/armsve/3/bli_gemm_armsve_asm_z2vx10_unindexed.c index 4b48e317d..70a3ca482 100644 --- a/kernels/armsve/3/bli_gemm_armsve_asm_z2vx10_unindexed.c +++ b/kernels/armsve/3/bli_gemm_armsve_asm_z2vx10_unindexed.c @@ -116,23 +116,23 @@ GEMM_ACOLCMPLX_CONTIGUOUS_LOAD_FWD(z28,z29,p0,%0,x2) " b.ne END_CCOL_PRFM \n\t" // Do not prefetch for generic C storage. " mov x16, %2 \n\t" " prfm PLDL1KEEP, [x16] \n\t" -" add x16, x16, %5 \n\t" +" add x16, x16, %4 \n\t" " prfm PLDL1KEEP, [x16] \n\t" -" add x16, x16, %5 \n\t" +" add x16, x16, %4 \n\t" " prfm PLDL1KEEP, [x16] \n\t" -" add x16, x16, %5 \n\t" +" add x16, x16, %4 \n\t" " prfm PLDL1KEEP, [x16] \n\t" -" add x16, x16, %5 \n\t" +" add x16, x16, %4 \n\t" " prfm PLDL1KEEP, [x16] \n\t" -" add x16, x16, %5 \n\t" +" add x16, x16, %4 \n\t" " prfm PLDL1KEEP, [x16] \n\t" -" add x16, x16, %5 \n\t" +" add x16, x16, %4 \n\t" " prfm PLDL1KEEP, [x16] \n\t" -" add x16, x16, %5 \n\t" +" add x16, x16, %4 \n\t" " prfm PLDL1KEEP, [x16] \n\t" -" add x16, x16, %5 \n\t" +" add x16, x16, %4 \n\t" " prfm PLDL1KEEP, [x16] \n\t" -" add x16, x16, %5 \n\t" +" add x16, x16, %4 \n\t" " prfm PLDL1KEEP, [x16] \n\t" " END_CCOL_PRFM: \n\t" " \n\t" diff --git a/kernels/armsve/3/bli_gemm_armsve_asm_z2vx7_unindexed.c b/kernels/armsve/3/bli_gemm_armsve_asm_z2vx7_unindexed.c index 4e5ef17e5..3d25719d9 100644 --- a/kernels/armsve/3/bli_gemm_armsve_asm_z2vx7_unindexed.c +++ b/kernels/armsve/3/bli_gemm_armsve_asm_z2vx7_unindexed.c @@ -123,17 +123,17 @@ GEMM_ACOLCMPLX_CONTIGUOUS_LOAD_FWD(z28,z29,p0,%0,x2) " b.ne END_CCOL_PRFM \n\t" // Do not prefetch for generic C storage. " mov x16, %2 \n\t" " prfm PLDL1KEEP, [x16] \n\t" -" add x16, x16, %5 \n\t" +" add x16, x16, %4 \n\t" " prfm PLDL1KEEP, [x16] \n\t" -" add x16, x16, %5 \n\t" +" add x16, x16, %4 \n\t" " prfm PLDL1KEEP, [x16] \n\t" -" add x16, x16, %5 \n\t" +" add x16, x16, %4 \n\t" " prfm PLDL1KEEP, [x16] \n\t" -" add x16, x16, %5 \n\t" +" add x16, x16, %4 \n\t" " prfm PLDL1KEEP, [x16] \n\t" -" add x16, x16, %5 \n\t" +" add x16, x16, %4 \n\t" " prfm PLDL1KEEP, [x16] \n\t" -" add x16, x16, %5 \n\t" +" add x16, x16, %4 \n\t" " prfm PLDL1KEEP, [x16] \n\t" " END_CCOL_PRFM: \n\t" " \n\t" diff --git a/kernels/armsve/3/bli_gemm_armsve_asm_z2vx8_unindexed.c b/kernels/armsve/3/bli_gemm_armsve_asm_z2vx8_unindexed.c index 305b99261..d0eef4a8c 100644 --- a/kernels/armsve/3/bli_gemm_armsve_asm_z2vx8_unindexed.c +++ b/kernels/armsve/3/bli_gemm_armsve_asm_z2vx8_unindexed.c @@ -120,19 +120,19 @@ GEMM_ACOLCMPLX_CONTIGUOUS_LOAD_FWD(z16,z17,p0,%0,x2) " b.ne END_CCOL_PRFM \n\t" // Do not prefetch for generic C storage. " mov x16, %2 \n\t" " prfm PLDL1KEEP, [x16] \n\t" -" add x16, x16, %5 \n\t" +" add x16, x16, %4 \n\t" " prfm PLDL1KEEP, [x16] \n\t" -" add x16, x16, %5 \n\t" +" add x16, x16, %4 \n\t" " prfm PLDL1KEEP, [x16] \n\t" -" add x16, x16, %5 \n\t" +" add x16, x16, %4 \n\t" " prfm PLDL1KEEP, [x16] \n\t" -" add x16, x16, %5 \n\t" +" add x16, x16, %4 \n\t" " prfm PLDL1KEEP, [x16] \n\t" -" add x16, x16, %5 \n\t" +" add x16, x16, %4 \n\t" " prfm PLDL1KEEP, [x16] \n\t" -" add x16, x16, %5 \n\t" +" add x16, x16, %4 \n\t" " prfm PLDL1KEEP, [x16] \n\t" -" add x16, x16, %5 \n\t" +" add x16, x16, %4 \n\t" " prfm PLDL1KEEP, [x16] \n\t" " END_CCOL_PRFM: \n\t" " \n\t"