Skip to content

[AMDGPU] Add image_atomic_fmin/fmax as aliases for GFX12 #107242

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

Merged
merged 1 commit into from
Sep 5, 2024
Merged

[AMDGPU] Add image_atomic_fmin/fmax as aliases for GFX12 #107242

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
1 commit
Select commit
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
4 changes: 4 additions & 0 deletions llvm/lib/Target/AMDGPU/MIMGInstructions.td
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -1675,6 +1675,10 @@ defm IMAGE_ATOMIC_PK_ADD_BF16 : MIMG_Atomic <mimgopc<0x87, MIMG.NOP, MIMG.NOP,
defm IMAGE_ATOMIC_ADD_FLT : MIMG_Atomic <mimgopc<0x83, MIMG.NOP, MIMG.NOP, MIMG.NOP>, "image_atomic_add_flt", 0, 1>;
defm IMAGE_ATOMIC_MIN_FLT : MIMG_Atomic <mimgopc<0x84, MIMG.NOP, MIMG.NOP, MIMG.NOP>, "image_atomic_min_num_flt", 0, 1, "image_atomic_min_flt">;
defm IMAGE_ATOMIC_MAX_FLT : MIMG_Atomic <mimgopc<0x85, MIMG.NOP, MIMG.NOP, MIMG.NOP>, "image_atomic_max_num_flt", 0, 1, "image_atomic_max_flt">;
let AssemblerPredicate = isGFX12Plus in {
def : AMDGPUMnemonicAlias<"image_atomic_fmin", "image_atomic_min_flt">;
Copy link
Contributor

Choose a reason for hiding this comment

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

Do these actually have the same behavior? I've been meaning to check, I think these switched from 2008 minnum to the 2019 behavior

Copy link
Collaborator

Choose a reason for hiding this comment

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

It's the same opcode 0x84.

Copy link
Contributor

Choose a reason for hiding this comment

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

I don't doubt it's the same opcode, I'm questioning whether this is another surprise opcode behavior change

Copy link
Contributor Author

Choose a reason for hiding this comment

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

Yes I think the behaviour changed (but I can't find a good spec for the pre-GFX12 behaviour).

def : AMDGPUMnemonicAlias<"image_atomic_fmax", "image_atomic_max_flt">;
}

defm IMAGE_SAMPLE : MIMG_Sampler_WQM <mimgopc<0x1b, 0x1b, 0x20>, AMDGPUSample>;
let OtherPredicates = [HasImageInsts, HasExtendedImageInsts] in {
Expand Down
6 changes: 6 additions & 0 deletions llvm/test/MC/AMDGPU/gfx12_asm_vimage_alias.s
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -29,3 +29,9 @@ image_atomic_min_num_flt v0, v0, s[0:7] dmask:0x1 dim:SQ_RSRC_IMG_1D

image_atomic_max_num_flt v0, v0, s[0:7] dmask:0x1 dim:SQ_RSRC_IMG_1D
// GFX12: image_atomic_max_flt v0, v0, s[0:7] dmask:0x1 dim:SQ_RSRC_IMG_1D ; encoding: [0x00,0x40,0x61,0xd0,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00]

image_atomic_fmin v0, v0, s[0:7] dmask:0x1 dim:SQ_RSRC_IMG_1D
// GFX12: image_atomic_min_flt v0, v0, s[0:7] dmask:0x1 dim:SQ_RSRC_IMG_1D ; encoding: [0x00,0x00,0x61,0xd0,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00]

image_atomic_fmax v0, v0, s[0:7] dmask:0x1 dim:SQ_RSRC_IMG_1D
// GFX12: image_atomic_max_flt v0, v0, s[0:7] dmask:0x1 dim:SQ_RSRC_IMG_1D ; encoding: [0x00,0x40,0x61,0xd0,0x00,0x00,0x00,0x00,0x00,0x00,0x00,0x00]
Loading