Skip to content

Raisetolinalg #412

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

Jump to bottom
Draft
arpitj1 wants to merge 83 commits into
base: raisetolinalg
Choose a base branch
from
Draft

Raisetolinalg #412

arpitj1 wants to merge 83 commits into from

Conversation

@arpitj1
Copy link
Collaborator

@arpitj1 arpitj1 commented Jun 6, 2024

Some modifications to fuse linalg.generic op with for op

arpitj1 added 30 commits June 6, 2024 08:41
@arpitj1
Unfinished changes with prototype function
0b97da9
@arpitj1
Loop over linalg.generic's input and output ops
69ef423
@arpitj1
Some comments
7678a05
@arpitj1
Partial changes from coding session to implement fusion of linalg.gen…
0e88095 
…eric and for op
@arpitj1
Incremental changes to fuse linalg and for loop- Logic for shifted op…
b57c0b8 
…erands and map for linalg.generic
@arpitj1
ran clang format
f54c33d
@arpitj1
some compile time fixes
56e2c54
@arpitj1
Some compile fixes
e253040
@arpitj1
Fixed all the compilation issues. Sample MLIR not raised
e99b8a5
@arpitj1
Bug fixes, generating some output at getLinalgArgMap
34f595c
@arpitj1
Almost implementated remap in affine dim for multi idx
05bad97
@arpitj1
Added submap op support and refactored the code to use submap
5bbf5ef
@arpitj1
bunch of fixes. Now able to generate raise linalg code
9018d92
@arpitj1
Now almost working second loop raising to linalg
ec041a0
@arpitj1
Fixes to correctly raise 2 level for loops to linalg.generic
23138fc
@arpitj1
Missed file update to enable linalg dialect in polygeist
5f20bd7
@arpitj1
Fix for syms and dims calculation
b0e96aa
@arpitj1
More tests added to cover different loop cases
ea76f0a
@arpitj1
Now able to compile 3/any number of loops with parallel iter type; Ad…
591c84e 
…ded extra tests in lit test
@arpitj1
Non iter-arg variant of matrix-mul and conv are now raised to linalg.…
b0108e3 
…generic
@arpitj1
submap canonicalizer implemented
4362c80
@arpitj1
Added reduction loops for linalg
77c8168
@arpitj1
Fix for incorrect for loop dims
98f0119
@arpitj1
Linalg.generic 4 loop cases raised- todo: reduction and some if-else …
59eec0b 
…cases failing
@arpitj1
Adding test case for all passing raising and lowering, example case o…
a363f13 
…f debufferizing added which works for tiling and fusion
@arpitj1
Added pass remove iter args from scf; Added psuedo code for submap ca…
814ca51 
…nonicalize
@arpitj1
Added removal of iter_args for affine loops
701f25a
@arpitj1
Temporary reverted pass registeration as the code was failing
d285fb5
@arpitj1
WIP commit
c40e7a9
@arpitj1
Added submap of submap canonicalizer with test- failing
788a3c4
arjaiswal and others added 30 commits June 11, 2025 13:12
Enabled kernel dialect correctly running on sample IR with kernel def…
c0f36d3 
…n collection
@arpitj1
Added linalgToKernel pass- compile failure
6a67379
@arpitj1
Working pattern matching and replacement for linalg generics
7f9d00f
@arpitj1
Partial changes for different files for kernel and input
d765bb9
@arpitj1
Crash fix
15ef84e
@arpitj1
Improved lib
44fed6c
@arpitj1
Removing redundant file
4a95c7f
@arpitj1
Renamed kernel lib
f1e5f02
@arpitj1
Added min_abs_index test
e941c5e
@arpitj1
Fixed a bunch of bugs in raiseToLinalg while raising polybench
a99fad9
@arpitj1
Fixed raise to linalg and canonicalizer to generate subview
4e782d5
@arpitj1
Fixed submap simplification, improved raisedToLinalg to work with non…
bd15b6d 
… constant bounds, added debufferization to work with allocs
@arpitj1
Added parallel fission pass
cb34836
@arpitj1
Added pattern for parallel to seq for loops
53c5d14
@arpitj1
Added raise-to-linalg-pipeline
60b81d2
@arpitj1
Added linalgGenericEliminateSubmaps and commented out submapToSubviewOp
7b2f5d9
@arpitj1
Canonicalization fix
71e441f
@arpitj1
bug fix for non nullptr in submap creation
e421a86
@arpitj1
Fix in linalg debufferizer - failure return and only insert memref.co…
56724a5 
…py if current!=totensor
@arpitj1
improved matcher to create a dependency graph and use it for matching
c3c2700
@arpitj1
Runtime failure but match happening correctly to kernel dialect
ca12291
@arpitj1
Working match for linalg kernel match for gemm
7c204f2
@arpitj1
Added debug prints
37dd847
@arpitj1
Able to raise gemv
7e3f0d0
@arpitj1
blas C codes- for raising to linalg
3b56eb3
@arpitj1
Debug prints for RaiseTolinalg and 2. SelectFunc pass to process just…
fa99aa8 
… a given function
@arpitj1
Update RemoveIterArgs to work with chain ops before store for affine.for
ed30a14
@arpitj1
Added int op support
a816708
@arpitj1
remote iter args improved and test added
0edd38e
@arpitj1
Implemented improvement in linalg debufferize to work through inverse…
3b8c43b 
…submap, and a more shophisticated approach taken for tracking current tensors via a tree of regions
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Reviewers

@chelini chelini Awaiting requested review from chelini

@wsmoses wsmoses Awaiting requested review from wsmoses

@ivanradanov ivanradanov Awaiting requested review from ivanradanov

Assignees

No one assigned

Labels

None yet

Projects

None yet

Milestone

No milestone

Development

Successfully merging this pull request may close these issues.

2 participants

@arpitj1 @chelini