[SparseBlas]Initial design of Sparse Coo data #496
Closed
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
Description
Motivation
Currently from one of the priority app CUDA - SYCL migration perspective (CUDA-SYCL Task Force) (bitsandbytes) , there is a requirement to support COO format. Since currently the scope is through CSR, the code migration is routed through CSR pathway. However for an efficient migration of CUDA applications invoking COO, I propose a design for COO format. I have taken reference from the existing CSR implementation as most of the function args are similar. The current application requires
cusparseCreateCoo
api, and I had taken reference from this api invokation:cusparseCreateCoo(cusparseSpMatDescr_t* spMatDescr, int64_t rows, int64_t cols, int64_t nnz, void* cooRowInd, void* cooColInd, void* cooValues, cusparseIndexType_t cooIdxType, cusparseIndexBase_t idxBase, cudaDataType valueType)
It would entail development of some COO auxiliary methods , getters ,aos and const variants in general. Since currently COO is not supported on iGPU yet , it would be nice to have this feature for future application migrations.
Tagging @Rbiessy from discussion.