Bump CAPI version number for LDS usage estimation function

[justinrosner]

Motivation

rocMLIR bumped its MIGraphX CAPI from version 5 to 6 to add mlirMIGraphXLDSUsageFitsArch(), so this PR bumps that API version and the rocMLIR commit hash to go with that.

This PR should only be merged once ROCm/rocMLIR#2400 get's merged into rocMLIR's develop.

Technical Details

• Update MLIR_MIGRAPHX_DIALECT_API_VERSION from version 5 -> 6
• Bumps rocMLIR commit hash

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

This PR updates MIGraphX’s GPU MLIR integration to match rocMLIR’s bumped MIGraphX CAPI version (to pick up the newly added mlirMIGraphXLDSUsageFitsArch() function on the rocMLIR side), and pins the dependency to a rocMLIR commit that provides that API.

Changes:

• Update the compile-time compatibility check to require MLIR_MIGRAPHX_DIALECT_API_VERSION == 6.
• Bump the pinned ROCm/rocMLIR commit in requirements.txt.

Reviewed changes

Copilot reviewed 1 out of 2 changed files in this pull request and generated no comments.

File	Description
src/targets/gpu/mlir.cpp	Updates the rocMLIR CAPI version gate from 5 to 6 to ensure ABI/API compatibility.
requirements.txt	Pins rocMLIR to a newer commit hash expected to expose CAPI version 6.

[pfultz2]

Converted to draft until the rocmlir PR gets merged.

[gh-app-migraphx-bot-pr-write]

Test	Batch	New Rate (aeb063)	Old Rate (edb3a3)	Diff	Status
torchvision-resnet50	64	738.27	1,534.68	-51.89%	🔴
torchvision-resnet50_fp16	64	711.39	5,810.63	-87.76%	🔴
torchvision-densenet121	32	1,831.92	2,602.81	-29.62%	🔴
torchvision-densenet121_fp16	32	801.33	4,538.31	-82.34%	🔴
torchvision-inceptionv3	32	1,616.62	1,592.40	1.52%	✅
torchvision-inceptionv3_fp16	32	2,708.40	2,759.48	-1.85%	✅
cadene-inceptionv4	16	625.53	737.33	-15.16%	🔴
cadene-resnext64x4	16	355.45	753.02	-52.80%	🔴
slim-mobilenet	64	8,024.61	5,697.60	40.84%	🔆
slim-nasnetalarge	64	176.12	223.38	-21.16%	🔴
slim-resnet50v2	64	3,152.86	2,218.69	42.11%	🔆
bert-mrpc-onnx	8	518.70	613.90	-15.51%	🔴
bert-mrpc-tf	1	153.75	484.65	-68.28%	🔴
pytorch-examples-wlang-gru	1	338.88	85.60	295.88%	🔆
pytorch-examples-wlang-lstm	1	458.78	179.15	156.09%	🔆
torchvision-resnet50_1	1	748.20	176.74	323.33%	🔆
cadene-dpn92_1	1	438.42	475.89	-7.88%	🔴
cadene-resnext101_1	1	333.62	363.74	-8.28%	🔴
onnx-taau-downsample	1	389.97	392.61	-0.67%	✅
dlrm-criteoterabyte	1	1.05	1.12	-5.82%	🔴
dlrm-criteoterabyte_fp16	1	nan	52.45	nan	❌
agentmodel	1	9,999.40	9,609.66	4.06%	✅
unet_fp16	2	19.21	57.31	-66.48%	🔴
resnet50v1_fp16	1	922.98	932.06	-0.97%	✅
resnet50v1_int8	1	929.64	934.40	-0.51%	✅
bert_base_cased_fp16	64	1,031.08	1,103.70	-6.58%	🔴
bert_large_uncased_fp16	32	347.03	347.14	-0.03%	✅
bert_large_fp16	1	203.42	205.05	-0.80%	✅
distilgpt2_fp16	16	645.71	2,099.92	-69.25%	🔴
yolov5s	1	565.41	568.38	-0.52%	✅
tinyllama	1	45.97	45.65	0.70%	✅
vicuna-fastchat	1	44.01	44.00	0.02%	✅
whisper-tiny-encoder	1	419.99	420.50	-0.12%	✅
whisper-tiny-decoder	1	413.77	417.67	-0.93%	✅
llama2_7b	1	20.47	20.50	-0.15%	✅
qwen1.5-7b	1	14.88	23.56	-36.85%	🔴
phi3-3.8b	1	26.99	26.79	0.74%	✅
llama3-8b	1	21.74	21.70	0.20%	✅
whisper-large-encoder	1	10.29	10.33	-0.39%	✅
whisper-large-decoder	1	35.97	106.29	-66.16%	🔴
mistral-7b	1	19.72	23.62	-16.54%	🔴
FLUX.1-schnell	1	785.84	764.63	2.77%	✅

Regressions detected 🔴

[gh-app-migraphx-bot-pr-write]

Test	Status	Result
bert-mrpc-onnx	✅	PASSED: MIGraphX meets tolerance
bert-mrpc-tf	❌	ERROR - check error output traceback Traceback (most recent call last): File "/src/AMDMIGraphX/tools/accuracy/accuracy_checker.py", line 377, in main() File "/src/AMDMIGraphX/tools/accuracy/accuracy_checker.py", line 313, in main import tensorflow as tf File "/usr/local/lib/python3.10/dist-packages/tensorflow/init.py", line 38, in from tensorflow.python.tools import module_util as _module_util File "/usr/local/lib/python3.10/dist-packages/tensorflow/python/init.py", line 36, in from tensorflow.python import pywrap_tensorflow as _pywrap_tensorflow File "/usr/local/lib/python3.10/dist-packages/tensorflow/python/pywrap_tensorflow.py", line 26, in self_check.preload_check() File "/usr/local/lib/python3.10/dist-packages/tensorflow/python/platform/self_check.py", line 63, in preload_check from tensorflow.python.platform import _pywrap_cpu_feature_guard ImportError: libamdhip64.so.6: cannot open shared object file: No such file or directory
pytorch-examples-wlang-gru	✅	PASSED: MIGraphX meets tolerance
pytorch-examples-wlang-lstm	✅	PASSED: MIGraphX meets tolerance
dlrm-criteoterabyte	✅	PASSED: MIGraphX meets tolerance
agentmodel	✅	PASSED: MIGraphX meets tolerance
unet	✅	PASSED: MIGraphX meets tolerance
resnet50v1	✅	PASSED: MIGraphX meets tolerance
bert_base_cased_fp16	✅	PASSED: MIGraphX meets tolerance
bert_large_uncased_fp16	🔴	FAILED: MIGraphX is not within tolerance - check verbose output
bert_large	✅	PASSED: MIGraphX meets tolerance
yolov5s	✅	PASSED: MIGraphX meets tolerance
tinyllama	✅	PASSED: MIGraphX meets tolerance
vicuna-fastchat	✅	PASSED: MIGraphX meets tolerance
whisper-tiny-encoder	✅	PASSED: MIGraphX meets tolerance
whisper-tiny-decoder	✅	PASSED: MIGraphX meets tolerance
distilgpt2_fp16	✅	PASSED: MIGraphX meets tolerance
llama2_7b	✅	PASSED: MIGraphX meets tolerance
qwen1.5-7b	✅	PASSED: MIGraphX meets tolerance
phi3-3.8b	✅	PASSED: MIGraphX meets tolerance
llama3-8b	✅	PASSED: MIGraphX meets tolerance
whisper-large-encoder	❌	ERROR - check error output traceback 2026-06-19 18:32:19.985335 [WARN] [/data/src/onnx/onnx_parser.cpp:282] Model has unbound symbolic dimension(s): batch_size, encoder_sequence_length, feature_size. These default to 1 and may cause unexpected behavior. Try setting --dim-param @<name> <value> or --input-dim @<input> <dims> if program compilation fails. Traceback (most recent call last): File "/src/AMDMIGraphX/tools/accuracy/accuracy_checker.py", line 377, in main() File "/src/AMDMIGraphX/tools/accuracy/accuracy_checker.py", line 224, in main model = migraphx.parse_onnx(model_name, default_dim_value=batch) RuntimeError: /data/src/include/migraphx/op/convolution.hpp:113: normalize_compute_shape: CONVOLUTION: mismatched channel numbers: input channels (1) != weights channels (80) * group (1)
whisper-large-decoder	✅	PASSED: MIGraphX meets tolerance
mistral-7b	✅	PASSED: MIGraphX meets tolerance
FLUX.1-schnell	✅	PASSED: MIGraphX meets tolerance