[ https://issues.apache.org/jira/browse/SYSTEMML-1436?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15944137#comment-15944137 ]
Nakul Jindal edited comment on SYSTEMML-1436 at 3/27/17 10:21 PM: ------------------------------------------------------------------ Thats great [~KrishnaKalyan3]!. I am not very familiar with the process, but as I understand it, you'll have to write a proposal and submit it . Use the description of this JIRA to get started. I suspect you've already done this, but [here|http://write.flossmanuals.net/gsocstudentguide/writing-a-proposal/] is the student manual for GSoC. To dig deeper, most of the code for the GPU backend is in [this|https://github.com/apache/incubator-systemml/tree/master/src/main/java/org/apache/sysml/runtime/instructions/gpu] directory. To dig even deeped, familiarize yourself with the code, look at SYSTEMML jiras which work on the GPU. was (Author: nakul02): Thats great. I am not very familiar with the process, but as I understand it, you'll have to write a proposal and submit it . Use the description of this JIRA to get started. I suspect you've already done this, but [here|http://write.flossmanuals.net/gsocstudentguide/writing-a-proposal/] is the student manual for GSoC. To dig deeper, most of the code for the GPU backend is in [this|https://github.com/apache/incubator-systemml/tree/master/src/main/java/org/apache/sysml/runtime/instructions/gpu] directory. To dig even deeped, familiarize yourself with the code, look at SYSTEMML jiras which work on the GPU. > Improve Sparse matrix support for GPU operations > ------------------------------------------------ > > Key: SYSTEMML-1436 > URL: https://issues.apache.org/jira/browse/SYSTEMML-1436 > Project: SystemML > Issue Type: Task > Components: Runtime > Reporter: Nakul Jindal > Labels: cuda, deeplearning, gpu, gsoc2017, mentor > > SystemML has a preliminary set of GPU implementation for its primitive > operations (Matmult, reductions, neural net operations among others). > Currently, these GPU operations work when SystemML is run on a single machine > (either using the Standalone mode or Spark mode). Programs written in the > external DSLs (DML & PyDML) and internal DSLs (Python and Scala) can enable > the use of these GPUs. > SystemML is aware of sparsity in matrix blocks and encodes them differently. > It has 3 different types of Sparse formats (CSR, COO & a custom MCSR). A lot > of the GPU operations are implemented for dense matrix blocks; for some GPU > operations, when sparse matrices are encountered, they are first converted to > dense and then sent to the GPU. > - This project is to implement CUDA kernels for Sparse Matrix blocks > - Operations to be implemented include reductions, element-wise operations, > neural network operations among others > This project is fairly isolated from the internal compiler & optimizer, > therefore a thorough knowledge of the entire system will not be needed. > Knowledge of CUDA programming is preferred. > For a initial implementation, the most efficient CUDA kernel is not required. > Rating - Medium > Mentors - [~nakul02], (optionally [~niketanpansare]) -- This message was sent by Atlassian JIRA (v6.3.15#6346)