Open discussion for MAGMA library (Matrix Algebra on GPU and Multicore Architectures)
This question is about the BLAS functions implemented in MAGMA (not interfaced to CUDA). Inspecting the magmablas_d.h file and the source code I understand:
are implemented in MAGMA
- geocmagmabi_dtrsm (only for Tesla, not Fermi)
are declared in magmablas_d.h but not implemented. Or am I wrong and these are implemented? I have not found the code.
- Posts: 30
- Joined: Tue Nov 15, 2011 12:38 pm
This is correct. We had at some point implementation of all routines in the header but as CUBLAS improved they were not needed and removed. In general, we prefer not to maintain a complete BLAS implementation; only certain routines that are important for magma and we see how to accelerate (and they stay until they find their way to CUBLAS).
- Posts: 254
- Joined: Fri Aug 21, 2009 10:39 pm
Return to User discussion
Who is online
Users browsing this forum: Bing [Bot], fletchjp, Google [Bot] and 1 guest