Move handling of integer signedness to the backend conversions #2597
Conversation
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
The function `getTypeForScalarType` currently takes an argument to specify the signedness of integer types. This is leakage of backend specific requirements into the torch dialect world. Because `getTypeForScalarType` is a utility function for the torch dialect, it should only produce types that match the sign conventions used by PyTorch (regular integers are signed and unsigned integers are unsigned). This commit removes the signedness argument from `getTypeForScalarType`, and moves the backend specific handling of integer types to the backend code.
|
With this patch, we get the following transformation: func.func private @forward(%arg0: !torch.vtensor<[20,100,35,45],f32>) -> !torch.vtensor<[20,100,35,45],f32> {
%int0 = torch.constant.int 0
%0 = torch.prim.ListConstruct %int0 : (!torch.int) -> !torch.list<int>
%int0_0 = torch.constant.int 0
%int0_1 = torch.constant.int 0
%cpu = torch.constant.device "cpu"
%none = torch.constant.none
%none_2 = torch.constant.none
%1 = torch.aten.empty.memory_format %0, %int0_0, %int0_1, %cpu, %none, %none_2 : !torch.list<int>, !torch.int, !torch.int, !torch.Device, !torch.none, !torch.none -> !torch.vtensor<[0],ui8>
return %arg0 : !torch.vtensor<[20,100,35,45],f32>
}
// -----// IR Dump After ConvertTorchToLinalg (convert-torch-to-linalg) //----- //
func.func private @forward(%arg0: !torch.vtensor<[20,100,35,45],f32>) -> !torch.vtensor<[20,100,35,45],f32> {
%int0 = torch.constant.int 0
%0 = torch.prim.ListConstruct %int0 : (!torch.int) -> !torch.list<int>
%int0_0 = torch.constant.int 0
%int0_1 = torch.constant.int 0
%cpu = torch.constant.device "cpu"
%none = torch.constant.none
%none_2 = torch.constant.none
%1 = torch_c.to_i64 %int0
%2 = arith.index_cast %1 : i64 to index
%3 = tensor.empty(%2) : tensor<?xi8>
%cast = tensor.cast %3 : tensor<?xi8> to tensor<0xi8>
return %arg0 : !torch.vtensor<[20,100,35,45],f32>
} |
AmosLewis
approved these changes
Nov 29, 2023
There was a problem hiding this comment.
LGTM. I also test the iree and shark-turbine where the issue first found, everything looks good now. nod-ai/SHARK-ModelDev#110
qingyunqu
approved these changes
Nov 29, 2023
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
The function
getTypeForScalarTypecurrently takes an argument to specify the signedness of integer types. This is leakage of backend specific requirements into the torch dialect world. BecausegetTypeForScalarTypeis a utility function for the torch dialect, it should only produce types that match the sign conventions used by PyTorch (regular integers are signed and unsigned integers are unsigned).This commit removes the signedness argument from
getTypeForScalarType, and moves the backend specific handling of integer types to the backend code.