Skip to content

ggml : use dynamic thread scheduling for matrix multiplication #6915

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.

Sign up for GitHub

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

Merged
merged 27 commits into from
May 15, 2024
Merged

ggml : use dynamic thread scheduling for matrix multiplication #6915

Changes from all commits
Commits
Show all changes
27 commits
Select commit Hold shift + click to select a range
3024fd6
Just reordering some structs.
kunnis May 9, 2024
5978b6e
Adding in the calls to mm_pause
kunnis May 9, 2024
e098171
Passing around the state
kunnis May 9, 2024
a968553
Renaming and moving a bunch of variables around.
kunnis May 9, 2024
7b932e4
Extracting the logic to it's own function.
kunnis May 9, 2024
4f95478
Moving some variable definitions into the chunk function.
kunnis May 9, 2024
086e5a8
Moving some variables around
kunnis May 10, 2024
209922f
moving src1_cont inside
kunnis May 10, 2024
bb1b1d0
Moving row_size
kunnis May 10, 2024
daa87b1
adding the current_chunk
kunnis May 10, 2024
700c782
Reorg the code.
kunnis May 10, 2024
891d583
Formatting to match the orig patch
kunnis May 10, 2024
9acaec5
starting to setup the chunking variables
kunnis May 10, 2024
c0557fa
Starting the buildup of the loop
kunnis May 10, 2024
4762d79
The yield shouldn't be necessary.
kunnis May 10, 2024
fc7dc51
adding the looping structure based on the chunk configuration.
kunnis May 10, 2024
807c825
Add in the re-chunking code.
kunnis May 10, 2024
974e43b
Making it much more likely to rechunk.
kunnis May 10, 2024
1c68ea8
disable resizing if numa is enabled.
kunnis May 10, 2024
bd80601
Updating comments with what we've learned.
kunnis May 10, 2024
d9ba30a
Fix formatting
kunnis May 14, 2024
163dbfd
Couple more formatting fixes.
kunnis May 14, 2024
6b0c90f
More style fixes.
kunnis May 14, 2024
741a198
Fix Warnings
kunnis May 14, 2024
2dd9f01
Going with unused because there's conditional logic that needs it.
kunnis May 15, 2024
f2aabab
Update ggml.c
slaren May 15, 2024
14c104d
Update ggml.c
slaren May 15, 2024
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
Loading
Loading