search

stanford-ppl / spatial-lang

Spatial: "Specify Parameterized Accelerators Through Inordinately Abstract Language"
MIT License
99 stars 12 forks source link

Added scatter/gather support in Fringe #245

Closed mattvilim closed 6 years ago

mattvilim commented 6 years ago

@mattfel1 @raghup17

Sparse apps are passing again in simulation (no regressions with dense); still need to be tested on the boards.

mattfel1 commented 6 years ago

Looks good to me. One of the SPMV apps is hanging but I need to look more at it to see what is causing that

mattvilim commented 6 years ago

I found the issue for the SPMV hang; I'll put in a fix for it.

mattfel1 commented 6 years ago

Spmv should be fixed already. The last hang I noticed after the pull request was due to a dense request of size 0 due to retiming

On Thu, Nov 30, 2017 at 2:30 PM Matthew Vilim notifications@github.com wrote:

I found the issue for the SPMV hang; I'll put in a fix for it.

— You are receiving this because you modified the open/close state.

Reply to this email directly, view it on GitHub https://github.com/stanford-ppl/spatial-lang/pull/245#issuecomment-348342042, or mute the thread https://github.com/notifications/unsubscribe-auth/AIKq_VLmNNXKLVEKukMlyZT5AHym6Pjbks5s7yx4gaJpZM4QpVKl .

mattvilim commented 6 years ago

Oh, I saw it was timing out on the GitHub regression page. Is the change on a branch other than develop?

mattfel1 commented 6 years ago

I made a few fixes a few days ago after merging your pull request and your changes have made it up to master now. The latest develop has everything passing: https://github.com/stanford-ppl/spatial-lang/wiki/Brnch:develop-Trgt:chisel

There are some new failures on retime branch today but none of these are related to scatter/gather