Rewrite 1x1 convolutions to gemm

[pfultz2]

This allows us to use rocblas for some gemms.

[shivadbhavsar]

Might be worthwhile to have a 1x1x1 test case too since its allowed here to make sure it generalizes to the 3D case

[pfultz2]

Need to investigate the nasnet failure.

[codecov]

Codecov Report

All modified and coverable lines are covered by tests ✅

Project coverage is 92.19%. Comparing base (aeb2047) to head (0374a01).

Additional details and impacted files

@@             Coverage Diff             @@
##           develop    #3568      +/-   ##
===========================================
+ Coverage    92.18%   92.19%   +0.01%     
===========================================
  Files          513      515       +2     
  Lines        21596    21639      +43     
===========================================
+ Hits         19908    19951      +43     
  Misses        1688     1688              

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

---

🚨 Try these New Features:

• Flaky Tests Detection - Detect and resolve failed and flaky tests

[CharlieL7]

LGTM. Is this the only case where doing a convolution as a GEMM is beneficial? Technically any convolution can be done as a GEMM.

[pfultz2]

LGTM. Is this the only case where doing a convolution as a GEMM is beneficial? Technically any convolution can be done as a GEMM.

For other sizes, we need to do windowing(and possibly some padding) which will require an extra copy when we reshape to a gemm, with tensors that are larger then the original input.

[causten]

I cant merge this with the amount of perf regressions

[migraphx-bot]

Test	Batch	Rate new 0374a0	Rate old 0ad807	Diff	Compare
torchvision-resnet50	64	3,122.29	3,239.44	-3.62%	🔴
torchvision-resnet50_fp16	64	6,651.52	6,989.98	-4.84%	🔴
torchvision-densenet121	32	2,427.28	2,435.40	-0.33%	✅
torchvision-densenet121_fp16	32	4,080.58	4,083.28	-0.07%	✅
torchvision-inceptionv3	32	1,622.52	1,630.36	-0.48%	✅
torchvision-inceptionv3_fp16	32	2,718.13	2,743.98	-0.94%	✅
cadene-inceptionv4	16	749.03	765.72	-2.18%	✅
cadene-resnext64x4	16	678.00	811.01	-16.40%	🔴
slim-mobilenet	64	7,401.44	7,471.20	-0.93%	✅
slim-nasnetalarge	64	182.24	208.50	-12.60%	🔴
slim-resnet50v2	64	3,236.34	3,438.88	-5.89%	🔴
bert-mrpc-onnx	8	1,153.34	1,147.12	0.54%	✅
bert-mrpc-tf	1	494.02	469.98	5.12%	🔆
pytorch-examples-wlang-gru	1	417.42	416.34	0.26%	✅
pytorch-examples-wlang-lstm	1	386.63	463.26	-16.54%	🔴
torchvision-resnet50_1	1	713.48	770.23	-7.37%	🔴
cadene-dpn92_1	1	407.35	399.89	1.87%	✅
cadene-resnext101_1	1	325.37	382.62	-14.96%	🔴
onnx-taau-downsample	1	345.82	345.09	0.21%	✅
dlrm-criteoterabyte	1	33.34	33.34	-0.01%	✅
dlrm-criteoterabyte_fp16	1	52.73	52.72	0.01%	✅
agentmodel	1	8,196.05	10,261.49	-20.13%	🔴
unet_fp16	2	58.72	58.87	-0.26%	✅
resnet50v1_fp16	1	867.75	982.51	-11.68%	🔴
resnet50v1_int8	1	1,008.79	1,007.17	0.16%	✅
bert_base_cased_fp16	64	1,169.18	1,169.34	-0.01%	✅
bert_large_uncased_fp16	32	355.07	363.46	-2.31%	✅
bert_large_fp16	1	192.58	198.15	-2.81%	✅
distilgpt2_fp16	16	2,198.25	2,201.47	-0.15%	✅
yolov5s	1	525.41	536.48	-2.06%	✅
tinyllama	1	43.38	43.38	0.01%	✅
vicuna-fastchat	1	179.54	172.97	3.80%	🔆
whisper-tiny-encoder	1	418.61	418.04	0.14%	✅
whisper-tiny-decoder	1	425.66	423.83	0.43%	✅

This build is not recommended to 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