Keep output shape from pointwise instruction after fusing with concat

[umangyadav]

fuse_ops.cpp pass is running find_contiguous_pointwise first which folds contiguous inside pointwise.

Later, fuse_concat_pointwise runs which is not maintaining the standard output shape while folding pointwise inside concat.

This leads to errors when trying to use reshape_lazy.

Renaming find_contiguous_pointwise to find_pointwise_contiguous. Many of the matchers have names such that operations appear in order they appear in graph.
E.g.
find_concat_pointwise means output of the concat is feeding into pointwise.
find_layernorm_pointwise means output of layernorm is going into pointwise.
Therefore find_pointwise_contiguous makes more sense.

[codecov]

Codecov Report

All modified and coverable lines are covered by tests ✅

Project coverage is 91.82%. Comparing base (f18b3ae) to head (0cf1444).
Report is 3 commits behind head on develop.

Additional details and impacted files

@@           Coverage Diff            @@
##           develop    #2881   +/-   ##
========================================
  Coverage    91.82%   91.82%           
========================================
  Files          477      477           
  Lines        18112    18112           
========================================
  Hits         16631    16631           
  Misses        1481     1481           

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

[migraphx-bot]

Test	Batch	Rate new 0cf144	Rate old f18b3a	Diff	Compare
torchvision-resnet50	64	2,852.25	2,852.01	0.01%	✅
torchvision-resnet50_fp16	64	6,516.73	6,516.75	-0.00%	✅
torchvision-densenet121	32	2,102.46	2,101.43	0.05%	✅
torchvision-densenet121_fp16	32	3,708.23	3,687.98	0.55%	✅
torchvision-inceptionv3	32	1,604.04	1,604.91	-0.05%	✅
torchvision-inceptionv3_fp16	32	2,572.53	2,574.13	-0.06%	✅
cadene-inceptionv4	16	724.83	724.85	-0.00%	✅
cadene-resnext64x4	16	682.94	683.17	-0.03%	✅
slim-mobilenet	64	5,948.12	5,935.54	0.21%	✅
slim-nasnetalarge	64	152.86	152.89	-0.02%	✅
slim-resnet50v2	64	2,668.26	nan	nan%	❌
bert-mrpc-onnx	8	917.88	919.16	-0.14%	✅
bert-mrpc-tf	1	433.49	436.14	-0.61%	✅
pytorch-examples-wlang-gru	1	430.76	423.38	1.74%	✅
pytorch-examples-wlang-lstm	1	388.35	388.56	-0.05%	✅
torchvision-resnet50_1	1	614.05	602.10	1.98%	✅
cadene-dpn92_1	1	391.92	393.08	-0.29%	✅
cadene-resnext101_1	1	331.37	332.15	-0.24%	✅
onnx-taau-downsample	1	304.18	305.99	-0.59%	✅
dlrm-criteoterabyte	1	28.79	28.74	0.15%	✅
dlrm-criteoterabyte_fp16	1	49.62	49.59	0.05%	✅
agentmodel	1	6,497.56	5,384.71	20.67%	🔆
unet_fp16	2	57.48	57.61	-0.22%	✅
resnet50v1_fp16	1	870.20	884.99	-1.67%	✅
resnet50v1_int8	1	799.18	797.72	0.18%	✅
bert_base_cased_fp16	64	1,055.71	1,055.63	0.01%	✅
bert_large_uncased_fp16	32	311.85	311.88	-0.01%	✅
bert_large_fp16	1	159.38	159.36	0.01%	✅
distilgpt2_fp16	16	1,858.63	1,858.22	0.02%	✅
yolov5s	1	483.42	486.48	-0.63%	✅
tinyllama	1	32.86	32.85	0.01%	✅
vicuna-fastchat	1	159.52	158.00	0.96%	✅
whisper-tiny-encoder	1	350.16	349.19	0.28%	✅
whisper-tiny-decoder	1	401.31	399.52	0.45%	✅

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