Browse Source

Add workaround for NVIDIA HPC

tags/v0.3.14^2
Martin Kroeker GitHub 5 years ago
parent
commit
49959d4f1c
No known key found for this signature in database GPG Key ID: 4AEE18F83AFDEB23
1 changed files with 14 additions and 5 deletions
  1. +14
    -5
      kernel/arm64/KERNEL.CORTEXA53

+ 14
- 5
kernel/arm64/KERNEL.CORTEXA53 View File

@@ -96,11 +96,20 @@ DNRM2KERNEL = nrm2.S
CNRM2KERNEL = znrm2.S
ZNRM2KERNEL = znrm2.S

DDOTKERNEL = dot.S
SDOTKERNEL = ../generic/dot.c
CDOTKERNEL = zdot.S
ZDOTKERNEL = zdot.S
DSDOTKERNEL = dot.S
ifneq ($(C_COMPILER), PGI)
SDOTKERNEL = ../generic/dot.c
else
SDOTKERNEL = dot.S
endif
DDOTKERNEL = dot.S
ifneq ($(C_COMPILER), PGI)
CDOTKERNEL = zdot.S
ZDOTKERNEL = zdot.S
else
CDOTKERNEL = ../arm/zdot.c
ZDOTKERNEL = ../arm/zdot.c
endif
DSDOTKERNEL = dot.S

DGEMM_BETA = dgemm_beta.S
SGEMM_BETA = sgemm_beta.S


Loading…
Cancel
Save