Remove the hard coded path in MIGRAPHX_CXX_COMPILER

[raramakr]

amdclang++/amdclang can be invoked without providing absolute path, since update-alternatives command has been run for these binaries This will also help to remove the hard coded paths in cmake target files for these compilers

[codecov]

Codecov Report

All modified and coverable lines are covered by tests ✅

Project coverage is 92.02%. Comparing base (22df950) to head (f34d90c).
Report is 3 commits behind head on develop.

Additional details and impacted files

@@           Coverage Diff            @@
##           develop    #3377   +/-   ##
========================================
  Coverage    92.02%   92.02%           
========================================
  Files          509      509           
  Lines        21014    21014           
========================================
  Hits         19339    19339           
  Misses        1675     1675           

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

[pfultz2]

This should still use the absolute path so we can be sure its using the same compiler the developer set migraphx to build with. I dont see any reason why this should be changed either as it doesnt affect any users.

[raramakr]

This should still use the absolute path so we can be sure its using the same compiler the developer set migraphx to build with. I dont see any reason why this should be changed either as it doesnt affect any users.

As part of our efforts to make ROCm relocatable , these hard coded paths need to be removed.

[pfultz2]

There shouldn't be any changes to cmake. The compiler_src.hpp should be updated to only use MIGRAPHX_CXX_COMPILER macro when its not windows:

struct MIGRAPHX_EXPORT src_compiler
{
#ifdef _WIN32
    fs::path compiler                         = MIGRAPHX_CXX_COMPILER;
#else
    fs::path compiler                         = "c++";
#endif
    std::vector<std::string> flags            = {};
    fs::path output                           = {};
    fs::path launcher                         = {};
    std::string out_ext                       = ".o";
    std::function<fs::path(fs::path)> process = nullptr;
    std::vector<char> compile(const std::vector<src_file>& srcs) const;
};

[raramakr]

Thank you @pfultz2
Updated the PR as suggested

[migraphx-bot]

Test	Batch	Rate new d68f73	Rate old fc26f0	Diff	Compare
torchvision-resnet50	64	3,240.31	3,261.06	-0.64%	✅
torchvision-resnet50_fp16	64	6,881.20	6,993.61	-1.61%	✅
torchvision-densenet121	32	2,428.60	2,434.05	-0.22%	✅
torchvision-densenet121_fp16	32	4,063.94	4,095.98	-0.78%	✅
torchvision-inceptionv3	32	1,634.23	1,638.20	-0.24%	✅
torchvision-inceptionv3_fp16	32	2,736.94	2,759.48	-0.82%	✅
cadene-inceptionv4	16	772.35	775.34	-0.39%	✅
cadene-resnext64x4	16	807.91	808.43	-0.06%	✅
slim-mobilenet	64	7,443.76	7,536.82	-1.23%	✅
slim-nasnetalarge	64	207.54	211.47	-1.86%	✅
slim-resnet50v2	64	3,343.96	3,501.50	-4.50%	🔴
bert-mrpc-onnx	8	1,149.97	1,149.92	0.00%	✅
bert-mrpc-tf	1	310.27	465.24	-33.31%	🔴
pytorch-examples-wlang-gru	1	436.80	413.68	5.59%	🔆
pytorch-examples-wlang-lstm	1	413.93	379.94	8.95%	🔆
torchvision-resnet50_1	1	758.89	786.00	-3.45%	🔴
cadene-dpn92_1	1	396.62	400.42	-0.95%	✅
cadene-resnext101_1	1	379.09	381.26	-0.57%	✅
onnx-taau-downsample	1	344.32	343.07	0.36%	✅
dlrm-criteoterabyte	1	35.09	33.36	5.17%	🔆
dlrm-criteoterabyte_fp16	1	57.41	52.73	8.88%	🔆
agentmodel	1	9,998.70	8,550.08	16.94%	🔆
unet_fp16	2	57.87	58.85	-1.66%	✅
resnet50v1_fp16	1	932.86	941.58	-0.93%	✅
resnet50v1_int8	1	946.32	970.00	-2.44%	✅
bert_base_cased_fp16	64	1,142.38	1,171.45	-2.48%	✅
bert_large_uncased_fp16	32	351.83	363.62	-3.24%	🔴
bert_large_fp16	1	207.09	200.57	3.25%	🔆
distilgpt2_fp16	16	2,155.18	2,203.25	-2.18%	✅
yolov5s	1	502.12	540.96	-7.18%	🔴
tinyllama	1	43.34	43.52	-0.41%	✅
vicuna-fastchat	1	172.37	174.66	-1.31%	✅
whisper-tiny-encoder	1	410.79	418.66	-1.88%	✅
whisper-tiny-decoder	1	425.47	428.20	-0.64%	✅

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