Commits

Commits on Jul 30, 2019

Switch to BLAS for CPU matmul
Sam Hatfield
committedJul 30, 2019
Move CUDA device reset to before cuBLAS handle creation
Sam Hatfield
committedJul 30, 2019
Move GPU memory allocation/deallocation outside tcgemm subroutine
Sam Hatfield
committedJul 30, 2019

Commits on Jul 29, 2019

Split up GPU initialization into separate subroutine
Sam Hatfield
committedJul 29, 2019
Make test matrices different shapes
Sam Hatfield
committedJul 29, 2019

Commits on Jul 24, 2019

Fix bug in Fortran/CUDA interface
Sam Hatfield
committedJul 24, 2019

Commits on Jun 25, 2019

Convert transa and transb parameters to ints
Sam Hatfield
committedJun 25, 2019
Switch to Tensor Core ("tcgemm")
Sam Hatfield
committedJun 25, 2019
Fix whitespace
Sam Hatfield
committedJun 25, 2019
Merge branch 'master' of https://github.com/samhatfield/matmul_fortran_tensor_core
Sam Hatfield
committedJun 25, 2019
Fixed a couple of bugs in cublas_gemm_c.cu
Sam Hatfield
committedJun 25, 2019
Rename trans[a,b]_int variables
samhatfield
committedJun 25, 2019
Fix a few bugs in the cuBLAS TCGEMM
Sam Hatfield
committedJun 25, 2019
Create separate a, b, and c matrices for DGEMM and TCGEMM
Sam Hatfield
committedJun 25, 2019
Implemented cuBLAS DGEMM
Sam Hatfield
committedJun 25, 2019

Commits on Jun 24, 2019

Added skeleton for Tensor Core GEMM
Samuel Hatfield
committedJun 24, 2019

Commits on Jun 21, 2019

samhatfield
committedJun 21, 2019
First working version with cuBLAS
Sam Hatfield
committedJun 21, 2019

Commits on Jun 20, 2019

Remove norm2 intrinsic
Sam Hatfield
committedJun 20, 2019
First working version
Sam Hatfield
committedJun 20, 2019
Add makefile and started matmul program
Sam Hatfield
committedJun 20, 2019
samhatfield
committedJun 20, 2019