fix bug in find_concat_op when input broadcasts are on different axes

[shivadbhavsar]

Fixes issue #3224
Which in turn fixes one of the issues exposed by #3104

[codecov]

Codecov Report

All modified and coverable lines are covered by tests ✅

Project coverage is 92.20%. Comparing base (1b7653d) to head (0150cbc).
Report is 149 commits behind head on develop.

Additional details and impacted files

@@             Coverage Diff             @@
##           develop    #3242      +/-   ##
===========================================
+ Coverage    92.16%   92.20%   +0.03%     
===========================================
  Files          493      493              
  Lines        19690    19700      +10     
===========================================
+ Hits         18148    18164      +16     
+ Misses        1542     1536       -6     

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

[migraphx-bot]

Test	Batch	Rate new 0150cb	Rate old e2d6ef	Diff	Compare
torchvision-resnet50	64	1,741.30	1,749.23	-0.45%	✅
torchvision-resnet50_fp16	64	4,064.50	4,078.43	-0.34%	✅
torchvision-densenet121	32	1,459.79	1,467.35	-0.52%	✅
torchvision-densenet121_fp16	32	2,520.24	2,532.63	-0.49%	✅
torchvision-inceptionv3	32	885.25	888.93	-0.41%	✅
torchvision-inceptionv3_fp16	32	1,478.60	1,483.54	-0.33%	✅
cadene-inceptionv4	16	410.51	411.78	-0.31%	✅
cadene-resnext64x4	16	417.53	419.22	-0.40%	✅
slim-mobilenet	64	3,987.61	4,003.58	-0.40%	✅
slim-nasnetalarge	64	100.56	100.96	-0.40%	✅
slim-resnet50v2	64	1,672.18	1,678.56	-0.38%	✅
bert-mrpc-onnx	8	612.15	616.93	-0.77%	✅
bert-mrpc-tf	1	276.91	279.35	-0.87%	✅
pytorch-examples-wlang-gru	1	321.83	364.12	-11.61%	🔴
pytorch-examples-wlang-lstm	1	291.65	294.38	-0.93%	✅
torchvision-resnet50_1	1	467.29	465.00	0.49%	✅
cadene-dpn92_1	1	246.51	247.39	-0.36%	✅
cadene-resnext101_1	1	203.22	204.15	-0.46%	✅
onnx-taau-downsample	1	205.50	206.13	-0.31%	✅
dlrm-criteoterabyte	1	22.83	22.91	-0.36%	✅
dlrm-criteoterabyte_fp16	1	42.57	42.71	-0.32%	✅
agentmodel	1	6,342.24	6,228.16	1.83%	✅
unet_fp16	2	34.19	34.30	-0.31%	✅
resnet50v1_fp16	1	584.69	595.94	-1.89%	✅
resnet50v1_int8	1	579.90	570.46	1.66%	✅
bert_base_cased_fp16	64	642.42	645.74	-0.51%	✅
bert_large_uncased_fp16	32	197.82	198.84	-0.51%	✅
bert_large_fp16	1	117.07	117.52	-0.38%	✅
distilgpt2_fp16	16	1,205.09	1,209.51	-0.37%	✅
yolov5s	1	301.13	300.77	0.12%	✅
tinyllama	1	23.21	23.32	-0.46%	✅
vicuna-fastchat	1	133.20	133.25	-0.04%	✅
whisper-tiny-encoder	1	243.31	244.46	-0.47%	✅
whisper-tiny-decoder	1	255.30	256.38	-0.42%	✅

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

[umangyadav]

Looks like there is another bug on line 886

   auto pred = [](auto i, auto j) {
            return i->get_operator() == j->get_operator() and
                   i->inputs().size() == i->inputs().size() and
                   i->outputs().size() == i->outputs().size();
        };

[shivadbhavsar]

Tracking in #3245