-
Notifications
You must be signed in to change notification settings - Fork 769
[SYCL][FPGA] Changing the attribute max_private_copies to private_copes #994
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
Changes from all commits
File filter
Filter by extension
Conversations
Jump to
Diff view
Diff view
There are no files selected for viewing
Original file line number | Diff line number | Diff line change | ||||
---|---|---|---|---|---|---|
|
@@ -3975,9 +3975,9 @@ void CodeGenModule::generateIntelFPGAAnnotation( | |||||
llvm::APSInt BWAInt = BWA->getValue()->EvaluateKnownConstInt(getContext()); | ||||||
Out << '{' << BWA->getSpelling() << ':' << BWAInt << '}'; | ||||||
} | ||||||
if (const auto *MCA = D->getAttr<IntelFPGAMaxPrivateCopiesAttr>()) { | ||||||
if (const auto *MCA = D->getAttr<IntelFPGAPrivateCopiesAttr>()) { | ||||||
llvm::APSInt MCAInt = MCA->getValue()->EvaluateKnownConstInt(getContext()); | ||||||
Out << '{' << MCA->getSpelling() << ':' << MCAInt << '}'; | ||||||
Out << "{max_private_copies:" << MCAInt << '}'; | ||||||
There was a problem hiding this comment. Choose a reason for hiding this commentThe reason will be displayed to describe this comment to others. Learn more. If this change is required, then:
Suggested change
There was a problem hiding this comment. Choose a reason for hiding this commentThe reason will be displayed to describe this comment to others. Learn more. Ah, I got it. In that way we don't want to change LLVM IR output. It's looking hacky, isn't it? :) There was a problem hiding this comment. Choose a reason for hiding this commentThe reason will be displayed to describe this comment to others. Learn more. Exactly.. As you mentioned, the "max" here is redundant, and is in fact misleading to customers. That's why we're getting rid of it. However, I'm keeping the IR the same for now because otherwise, we would have to change the llvm-spirv translator and the SPIRV spec to do things the right way. I should open a JIRA to track that effort, but this does not affect the customer flow. There was a problem hiding this comment. Choose a reason for hiding this commentThe reason will be displayed to describe this comment to others. Learn more.
I'd argue that no changes would have to be made to the SPIR-V specification, as it doesn't specify the LLVM IR layout. Would it be better to accompany this patch with a correspondent adjustment to the translator? I believe the fix would be trivial. There was a problem hiding this comment. Choose a reason for hiding this commentThe reason will be displayed to describe this comment to others. Learn more. @mohammadfawaz please address @AGindinson comment. I'm OK to go either way. There was a problem hiding this comment. Choose a reason for hiding this commentThe reason will be displayed to describe this comment to others. Learn more. @MrSidims @AGindinson I will address the translator changes asap! |
||||||
} | ||||||
if (const auto *NBA = D->getAttr<IntelFPGANumBanksAttr>()) { | ||||||
llvm::APSInt NBAInt = NBA->getValue()->EvaluateKnownConstInt(getContext()); | ||||||
|
Uh oh!
There was an error while loading. Please reload this page.