Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

fix ell error on small mtx and allow flexible warp_size #375

Merged
merged 1 commit into from
Oct 30, 2019

Conversation

yhmtsai
Copy link
Member

@yhmtsai yhmtsai commented Oct 30, 2019

This PR do the following things

  1. fix ell kernel error on the small matrix.
    the nwarp_per_row calculation is wrong when the needed threads is less than 512.
    Now, it is token from compute_subwarp_size_and_atomicity to make sure using the same nwarp_per_row.
  2. add the corresponding test and missing test.
  3. the ell kernel is depend on config::warp_size not specific 32.

@yhmtsai yhmtsai added reg:testing This is related to testing. mod:cuda This is related to the CUDA module. type:matrix-format This is related to the Matrix formats labels Oct 30, 2019
@yhmtsai yhmtsai self-assigned this Oct 30, 2019
Copy link
Member

@thoasm thoasm left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

LGTM!

@yhmtsai yhmtsai force-pushed the fix_ell_kernel_on_small_matrix branch from 6c1ce1d to 29c6692 Compare October 30, 2019 14:57
@tcojean tcojean added the 1:ST:ready-to-merge This PR is ready to merge. label Oct 30, 2019
@yhmtsai yhmtsai merged commit 3ba4a47 into develop Oct 30, 2019
@yhmtsai yhmtsai deleted the fix_ell_kernel_on_small_matrix branch October 30, 2019 18:59
@tcojean tcojean mentioned this pull request Nov 27, 2019
4 tasks
tcojean added a commit that referenced this pull request Dec 2, 2019
This version of Ginkgo provides a few fixes in Ginkgo's core routines. The
supported systems and requirements are unchanged from version 1.1.0.

### Fixes
+ Improve Ginkgo's installation and fix the `test_install` step ([#406](#406)),
+ Fix some documentation issues ([#406](#406)),
+ Fix multiple code issues reported by sonarqube ([#406](#406)),
+ Update the git-cmake-format repository ([#399](#399)),
+ Improve the global update header script ([#390](#390)),
+ Fix broken bounds checks ([#388](#388)),
+ Fix CSR strategies and improve performance ([#379](#379)),
+ Fix a small typo in the stencil examples ([#381](#381)),
+ Fix ELL error on small matrices ([#375](#375)),
+ Fix SellP read function ([#374](#374)),
+ Add factorization support in `create_new_algorithm.sh`  ([#371](#371)).
tcojean added a commit that referenced this pull request Dec 3, 2019
Minor release v1.1.1

This version of Ginkgo provides a few fixes in Ginkgo's core routines. The
supported systems and requirements are unchanged from version 1.1.0.

### Fixes
+ Fix the `test_install` step with `HIP` ([#409](#409)),
+ Improve Ginkgo's installation and fix the `test_install` step ([#406](#406)),
+ Fix some documentation issues ([#406](#406)),
+ Fix multiple code issues reported by sonarqube ([#406](#406)),
+ Update the git-cmake-format repository ([#399](#399)),
+ Improve the global update header script ([#390](#390)),
+ Fix broken bounds checks ([#388](#388)),
+ Fix CSR strategies and improve performance ([#379](#379)),
+ Fix a small typo in the stencil examples ([#381](#381)),
+ Fix ELL error on small matrices ([#375](#375)),
+ Fix SellP read function ([#374](#374)),
+ Add factorization support in `create_new_algorithm.sh`  ([#371](#371))
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
1:ST:ready-to-merge This PR is ready to merge. mod:cuda This is related to the CUDA module. reg:testing This is related to testing. type:matrix-format This is related to the Matrix formats
Projects
None yet
Development

Successfully merging this pull request may close these issues.

4 participants