-
Notifications
You must be signed in to change notification settings - Fork 152
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
[5.2.X] TRMM functions do not have correct correspondence in rocBLAS #1265
Comments
Hi @emankov, For We have, however, implemented rocblas_xtrmm_outofplace functions which have the 2 extra parameters being requested. This function supports out-of-place and in-place functionality for trmm. We do not plan on changing the current I hope this helps, |
+ Mark rocBLAS TRMM functions rocblas_(s|d|c|z)trmm_outofplace, as supported only for TRMM v2 CUDA analogues + Mark hipBLAS TRMM functions hipblas(S|D|C|Z)trmm as HIP_UNSUPPORTED + Regenerate and update docs and hipify-perl accordingly [Reasons] + hipBLAS TRMM functions hipblas(S|D|C|Z)trmm, actually, do not match neither cublas TRMM functions, nor cublas TRMM _v2 functions: ROCm/hipBLAS#524 + There is a correspondence between cuBLAS cublas_(s|d|c|z)trmm and rocBLAS TRMM rocblas_(s|d|c|z)trmm_outofplace, not rocblas_(s|d|c|z)trmm: fixed it [ToDo] + Close ROCm/rocBLAS#1265 as erroneous + Remove HIP_UNSUPPORTED mark from hipblas(S|D|C|Z)trmm functions after merging ROCm/hipBLAS#504 + Add cublas2rocblas and update cublas2hipblas synthetic tests
Thank you, @daineAMD! I've renamed So closing this issue as ERRONEOUS. |
…rformance further. Related ticket SWDEV-325544 and LWPTENSILE-51. (ROCm#1265)
rocBLAS TRMM functions
rocblas_strmm
,rocblas_dtrmm
,rocblas_ctrmm
,rocblas_ztrmm
do not match neither cublas TRMM functions, nor cublas TRMM _v2 functions.For instance:
cublasStrmm:
cublasStrmm_v2:
rocblas_strmm:
The same goes for hipBLAS analogues
hipblasStrmm
,hipblasDtrmm
,hipblasCtrmm
,hipblasZtrmm
(ROCm/hipBLAS#524).So, the above 4 rocBLAS and 4 hipBLAS functions are marked as
HIP UNSUPPORTED
.[Solution]
As far as rocBLAS doesn't support
v1
BLAS functions, populate rocblas TRMM functions with two missing arguments:float* C
andint ldc
and revise functions' logic.The text was updated successfully, but these errors were encountered: