-
Notifications
You must be signed in to change notification settings - Fork 14
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] Add support for multislot FIFO loads #330
Conversation
# (c) Copyright 2025 Advanced Micro Devices, Inc. or its affiliates | ||
# RUN: llc -march=aie2p -run-pass=postmisched %s -o - | FileCheck %s | ||
|
||
# This test checks that we are scheduling mulislot loads from differnt banks in on VLIW bundle. |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
nit: multislot ; different; one
# This test checks that we are scheduling mulislot loads from differnt banks in on VLIW bundle. | ||
|
||
--- | ||
name: VLD_FILL_512_pseudo_same |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
nit: same memory bank, not same bundle. That's confusing with the comment right above.
; CHECK-NEXT: [[VLDB_FILL_512_:%[0-9]+]]:eps, [[VLDB_FILL_512_1:%[0-9]+]]:eldfiforeg, [[VLDB_FILL_512_2:%[0-9]+]]:erf2 = VLDB_FILL_512 [[DEF]], [[DEF1]], [[DEF2]] | ||
; CHECK-NEXT: PseudoRET implicit $lr, implicit [[VLDB_FILL_512_]], implicit [[VLDB_FILL_512_1]], implicit [[VLDB_FILL_512_2]] | ||
; CHECK-NEXT: [[VLD_FILL_512_pseudo:%[0-9]+]]:eps, [[VLD_FILL_512_pseudo1:%[0-9]+]]:eldfiforeg, [[VLD_FILL_512_pseudo2:%[0-9]+]]:erf2 = VLD_FILL_512_pseudo [[DEF]], [[DEF1]], [[DEF2]] | ||
; CHECK-NEXT: PseudoRET implicit $lr, implicit [[VLD_FILL_512_pseudo]], implicit [[VLD_FILL_512_pseudo1]], implicit [[VLD_FILL_512_pseudo2]] |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Would be nice if we would have a pass 'expand-pseudo-to-first-alternative' that we could run in this test.
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
But we are testing the instruction-select
pass, so I think it makes sense to test the new opcodes
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
I see no problems. @F-Stuckmann may be eager to approve...
d78e1ff
to
e4c971a
Compare
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
LGTM
All remaining VLD operations will be handled in a follow-up PR.