Add rock prefix for rocMLIR attributes

[dhernandez0]

Motivation

rocMLIR is going to use rock. prefix for attributes (ROCm/rocMLIR#2334), just updating them on MIGraphX side as well.

We probably want to sync this PR merge with rocMLIR PR. And bump rocMLIR commit hash.

Technical Details

• docs/dev/triage-rocmlir.rst: update example, added prefix
• src/targets/gpu/mlir.cpp: set attributes with prefix
• test/gpu/mlir.cpp: update get_attrs() with prefix as well

Changelog Category

Add a CHANGELOG.md entry for any option other than Not Applicable

• • Added: New functionality.
• • Changed: Changes to existing functionality.
• • Removed: Functionality or support that has been removed. (Compared to a previous release)
• • Optimized: Component performance that has been optimized or improved.
• • Resolved Issues: Known issues from a previous version that have been resolved.
• • Not Applicable: This PR is not to be included in the changelog.

[Copilot]

Pull request overview

Updates MIGraphX’s rocMLIR integration to use the new rock.-prefixed function attributes expected by rocMLIR, and aligns tests/docs with the updated MLIR printing.

Changes:

• Switch MLIR func.func attributes from arch/kernel/num_* to rock.arch/rock.kernel/rock.num_* (and rock.enable_splitk_for_tuning).
• Update GPU MLIR dump tests to expect the new attribute names.
• Refresh rocMLIR triage documentation snippets to show the new attribute names.

Reviewed changes

Copilot reviewed 3 out of 3 changed files in this pull request and generated 2 comments.

File	Description
src/targets/gpu/mlir.cpp	Emits rock.* attributes on generated func.func ops (including SplitK unit attr).
test/gpu/mlir.cpp	Updates expected MLIR attribute dictionaries to match rock.* naming.
docs/dev/triage-rocmlir.rst	Updates example MLIR module snippets to use rock.* attribute names.

---

💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.

[Copilot]

The example shows rock.enable_splitk_for_tuning = true, but MIGraphX emits this as a unit attribute (see mlirUnitAttrGet(...) usage in src/targets/gpu/mlir.cpp), which prints as just rock.enable_splitk_for_tuning (no = true) in MLIR assembly. Update the snippet to match the actual printed MLIR so users can copy/paste it reliably during triage.

Suggested change

	func.func @mlir_convolution(%arg0: !migraphx.shaped<2x8x3x3xf32, 72x9x3x1>, %arg1: !migraphx.shaped<1x8x4x4xf32, 128x16x4x1>) -> !migraphx.shaped<1x2x2x2xf32, 8x4x2x1> attributes {rock.arch = "gfx90a:sramecc+:xnack-", rock.enable_splitk_for_tuning = true, rock.kernel = "mixr", rock.num_cu = 110 : i64, rock.num_chiplets = 1 : i64} {
	func.func @mlir_convolution(%arg0: !migraphx.shaped<2x8x3x3xf32, 72x9x3x1>, %arg1: !migraphx.shaped<1x8x4x4xf32, 128x16x4x1>) -> !migraphx.shaped<1x2x2x2xf32, 8x4x2x1> attributes {rock.arch = "gfx90a:sramecc+:xnack-", rock.enable_splitk_for_tuning, rock.kernel = "mixr", rock.num_cu = 110 : i64, rock.num_chiplets = 1 : i64} {

[Copilot]

Same as above: rock.enable_splitk_for_tuning should be shown as a unit attribute (no = true) to match MIGraphX’s emitted/printed MLIR and avoid confusing copy/paste during triage.

Suggested change

	func.func @mlir_dot_add(%arg0: !migraphx.shaped<1x5x4xf32, 20x4x1>, %arg1: !migraphx.shaped<1x4x3xf32, 12x3x1>, %arg2: !migraphx.shaped<1x5x3xf32, 15x3x1>) -> !migraphx.shaped<1x5x3xf32, 15x3x1> attributes {rock.arch = "gfx90a:sramecc+:xnack-", rock.enable_splitk_for_tuning = true, rock.kernel = "mixr", rock.num_cu = 110 : i64, rock.num_chiplets = 1 : i64} {
	func.func @mlir_dot_add(%arg0: !migraphx.shaped<1x5x4xf32, 20x4x1>, %arg1: !migraphx.shaped<1x4x3xf32, 12x3x1>, %arg2: !migraphx.shaped<1x5x3xf32, 15x3x1>) -> !migraphx.shaped<1x5x3xf32, 15x3x1> attributes {rock.arch = "gfx90a:sramecc+:xnack-", rock.enable_splitk_for_tuning, rock.kernel = "mixr", rock.num_cu = 110 : i64, rock.num_chiplets = 1 : i64} {

[umangyadav]

you would need to bump rocMLIR SHA too

[bdevorem]

assuming the rocmlir hash gets updated after that PR is merged, lgtm