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

[AIE2P] Support fifo loads with extra storage #332

Open
wants to merge 4 commits into
base: aie-public
Choose a base branch
from

Conversation

khallouh
Copy link
Collaborator

@khallouh khallouh commented Feb 4, 2025

No description provided.

; CHECK-NEXT: nop
; CHECK-NEXT: nop
; CHECK-NEXT: vlda lfl0, [p1, #0]
; CHECK-NEXT: vlda x0, [p1, #192]; mov p2, p0
Copy link
Collaborator Author

Choose a reason for hiding this comment

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

We select the vector load because of the mapping to the vector bank instead of the fifo bank. This is because we introduce a copy during InstructionSelect to map the extra fifo input and output operands of the llvm intrinsic to the implicit use and def of lfe register in the instruction. The reason being we don't currently look into the uses beyond copies, bitcasts and phi nodes when determining the register bank assignments.

@gbossu
Copy link
Collaborator

gbossu commented Feb 4, 2025

I'd be happy if you could look at the changes I've made in https://github.com/Xilinx/llvm-aie/pull/331/files and already apply them here for the new frontend intrinsics, builtins, and LLVM intrinsics. LMK if you need more details :)

@khallouh
Copy link
Collaborator Author

khallouh commented Feb 5, 2025

I'd be happy if you could look at the changes I've made in https://github.com/Xilinx/llvm-aie/pull/331/files and already apply them here for the new frontend intrinsics, builtins, and LLVM intrinsics. LMK if you need more details :)

Updated

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

Successfully merging this pull request may close these issues.

2 participants