search

amd / aocl-sparse

AMD optimized Sparse Linear Algebra library
Other
26 stars 10 forks source link

Adding Atomic operations with AOCL library in custom kernal #5

Open Pavanmahaveer7 opened 6 months ago

Pavanmahaveer7 commented 6 months ago

HI, I been trying to work on a case of utilizing atomic ops as added within kernel to do sparse matrix & vector multplication. how can we improve the performance of this .

vchunchu-amd commented 3 months ago

Atomic operations inside computationally intensive core kernels(hot loop) is generally not a good idea if performance is priority.