Add unpack_int4 kernel

[music-dino]

• Implement JIT kernel for unpack_int4
• Remove unpack_int4 CPU offloading for GPU target

[codecov]

Codecov Report

All modified and coverable lines are covered by tests ✅

Project coverage is 92.17%. Comparing base (ddc4c0c) to head (fa65b0c).
Report is 136 commits behind head on develop.

Additional details and impacted files

@@           Coverage Diff            @@
##           develop    #3528   +/-   ##
========================================
  Coverage    92.17%   92.17%           
========================================
  Files          512      512           
  Lines        21387    21391    +4     
========================================
+ Hits         19714    19718    +4     
  Misses        1673     1673           

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

[lakhinderwalia]

Left a couple of minor comments. Approved.

[migraphx-bot]

Test	Batch	Rate new fa65b0	Rate old 275f85	Diff	Compare
torchvision-resnet50	64	3,258.03	3,260.33	-0.07%	✅
torchvision-resnet50_fp16	64	6,995.32	6,996.93	-0.02%	✅
torchvision-densenet121	32	2,435.96	2,436.13	-0.01%	✅
torchvision-densenet121_fp16	32	4,086.13	4,080.65	0.13%	✅
torchvision-inceptionv3	32	1,637.39	1,639.01	-0.10%	✅
torchvision-inceptionv3_fp16	32	2,759.10	2,759.60	-0.02%	✅
cadene-inceptionv4	16	776.93	775.88	0.14%	✅
cadene-resnext64x4	16	811.82	811.11	0.09%	✅
slim-mobilenet	64	7,538.22	7,539.27	-0.01%	✅
slim-nasnetalarge	64	211.48	211.54	-0.03%	✅
slim-resnet50v2	64	3,502.95	3,505.47	-0.07%	✅
bert-mrpc-onnx	8	1,148.74	1,151.85	-0.27%	✅
bert-mrpc-tf	1	468.86	463.50	1.16%	✅
pytorch-examples-wlang-gru	1	414.84	411.56	0.80%	✅
pytorch-examples-wlang-lstm	1	379.97	388.27	-2.14%	✅
torchvision-resnet50_1	1	788.03	764.51	3.08%	🔆
cadene-dpn92_1	1	415.37	402.09	3.30%	🔆
cadene-resnext101_1	1	383.97	381.69	0.60%	✅
onnx-taau-downsample	1	343.27	342.54	0.21%	✅
dlrm-criteoterabyte	1	33.35	33.35	0.00%	✅
dlrm-criteoterabyte_fp16	1	52.71	52.74	-0.06%	✅
agentmodel	1	8,509.91	8,186.92	3.95%	🔆
unet_fp16	2	58.97	58.82	0.27%	✅
resnet50v1_fp16	1	941.69	953.22	-1.21%	✅
resnet50v1_int8	1	1,026.06	1,003.11	2.29%	✅
bert_base_cased_fp16	64	1,172.09	1,170.57	0.13%	✅
bert_large_uncased_fp16	32	363.54	363.63	-0.03%	✅
bert_large_fp16	1	198.88	201.70	-1.40%	✅
distilgpt2_fp16	16	2,204.39	2,202.28	0.10%	✅
yolov5s	1	538.65	537.82	0.16%	✅
tinyllama	1	43.45	43.44	0.03%	✅
vicuna-fastchat	1	175.85	173.47	1.37%	✅
whisper-tiny-encoder	1	418.07	418.26	-0.05%	✅
whisper-tiny-decoder	1	425.04	436.61	-2.65%	✅

Check results before merge 🔆

[migraphx-bot]

     ✅ bert-mrpc-onnx: PASSED: MIGraphX meets tolerance

     ✅ bert-mrpc-tf: PASSED: MIGraphX meets tolerance

     ✅ pytorch-examples-wlang-gru: PASSED: MIGraphX meets tolerance

     ✅ pytorch-examples-wlang-lstm: PASSED: MIGraphX meets tolerance

     ✅ torchvision-resnet50_1: PASSED: MIGraphX meets tolerance

     ✅ cadene-dpn92_1: PASSED: MIGraphX meets tolerance

     ✅ cadene-resnext101_1: 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