Update lib target soversion

[arvindcheru]

Summary of proposed changes:

• Update lib target so version format to recommended format - (ex: libmigraphx.so.2.7.0 updated to libmigraphx.so.2.7)

[pfultz2]

Why are we dropping the patch version?

[arvindcheru]

Why are we dropping the patch version?
Looking latest versions majority of modules are having soversion with major,minor based link names. there are 3 other modules including migraphx were only showing difference by having patch version also. these 4 modules are requested to update using these changes proposed to the recommended soversion format.

[codecov]

Codecov Report

Merging #2267 (70e72b1) into develop (e748657) will increase coverage by 0.09%.
The diff coverage is n/a.

❗ Current head 70e72b1 differs from pull request most recent head d988f21. Consider uploading reports for the commit d988f21 to get more accurate results

@@             Coverage Diff             @@
##           develop    #2267      +/-   ##
===========================================
+ Coverage    91.37%   91.46%   +0.09%     
===========================================
  Files          439      433       -6     
  Lines        16493    16174     -319     
===========================================
- Hits         15069    14792     -277     
+ Misses        1424     1382      -42     

see 13 files with indirect coverage changes

[migraphx-bot]

Test	Batch	Rate new d51632	Rate old 883732	Diff	Compare
torchvision-resnet50	64	2,325.71	2,323.99	0.07%	✅
torchvision-resnet50_fp16	64	5,350.90	5,358.70	-0.15%	✅
torchvision-densenet121	32	1,849.08	1,847.03	0.11%	✅
torchvision-densenet121_fp16	32	3,417.04	3,410.48	0.19%	✅
torchvision-inceptionv3	32	1,295.60	1,295.72	-0.01%	✅
torchvision-inceptionv3_fp16	32	2,537.61	2,536.60	0.04%	✅
cadene-inceptionv4	16	620.49	620.02	0.08%	✅
cadene-resnext64x4	16	589.13	588.80	0.06%	✅
slim-mobilenet	64	7,219.21	7,219.83	-0.01%	✅
slim-nasnetalarge	64	236.43	236.41	0.01%	✅
slim-resnet50v2	64	2,557.41	2,559.45	-0.08%	✅
bert-mrpc-onnx	8	824.88	824.99	-0.01%	✅
bert-mrpc-tf	1	389.80	387.19	0.67%	✅
pytorch-examples-wlang-gru	1	299.90	299.70	0.07%	✅
pytorch-examples-wlang-lstm	1	311.71	316.36	-1.47%	✅
torchvision-resnet50_1	1	548.38	543.92	0.82%	✅
torchvision-inceptionv3_1	1	307.60	305.91	0.55%	✅
cadene-dpn92_1	1	351.64	353.11	-0.42%	✅
cadene-resnext101_1	1	219.68	219.09	0.27%	✅
slim-vgg16_1	1	224.20	224.68	-0.21%	✅
slim-mobilenet_1	1	1,532.56	1,495.14	2.50%	✅
slim-inceptionv4_1	1	218.26	216.37	0.87%	✅
onnx-taau-downsample	1	304.50	306.35	-0.60%	✅
dlrm-criteoterabyte	1	21.71	21.68	0.13%	✅
dlrm-criteoterabyte_fp16	1	40.75	40.75	0.02%	✅
agentmodel	1	5,861.80	5,869.23	-0.13%	✅
unet_fp16	2	55.19	55.16	0.04%	✅
resnet50v1_fp16	1	771.56	751.35	2.69%	✅
bert_base_cased_fp16	64	971.31	970.93	0.04%	✅
bert_large_uncased_fp16	32	305.07	305.32	-0.08%	✅
bert_large_fp16	1	166.89	167.18	-0.17%	✅
distilgpt2_fp16	16	1,278.83	1,278.19	0.05%	✅

This build is OK for merge ✅

[migraphx-bot]

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

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

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

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

    :white_check_mark:torchvision-resnet50_1: PASSED: MIGraphX meets tolerance

    :white_check_mark:torchvision-inceptionv3_1: PASSED: MIGraphX meets tolerance

    :white_check_mark:cadene-dpn92_1: PASSED: MIGraphX meets tolerance

    :white_check_mark:cadene-resnext101_1: PASSED: MIGraphX meets tolerance

    :white_check_mark:slim-vgg16_1: PASSED: MIGraphX meets tolerance

    :white_check_mark:slim-mobilenet_1: PASSED: MIGraphX meets tolerance

    :white_check_mark:slim-inceptionv4_1: PASSED: MIGraphX meets tolerance

    :white_check_mark:dlrm-criteoterabyte: PASSED: MIGraphX meets tolerance

    :white_check_mark:agentmodel: PASSED: MIGraphX meets tolerance

    :white_check_mark:unet: PASSED: MIGraphX meets tolerance

    :white_check_mark:resnet50v1: PASSED: MIGraphX meets tolerance

🔴bert_base_cased_fp16: FAILED: MIGraphX is not within tolerance - check verbose output

🔴bert_large_uncased_fp16: FAILED: MIGraphX is not within tolerance - check verbose output

    :white_check_mark:bert_large: PASSED: MIGraphX meets tolerance

🔴distilgpt2_fp16: FAILED: MIGraphX is not within tolerance - check verbose output

[pfultz2]

The major, minor and patch should be part of the SO major version:

math(EXPR MIGRAPHX_SO_MAJOR_VERSION "(${PROJECT_VERSION_MAJOR} * 1000 * 1000) + (${PROJECT_VERSION_MINOR} * 1000) + ${PROJECT_VERSION_PATCH}")
set(MIGRAPHX_SO_VERSION ${MIGRAPHX_SO_MAJOR_VERSION}.0)

Or we if we can drop the minor version it could just be:

math(EXPR MIGRAPHX_SO_VERSION "(${PROJECT_VERSION_MAJOR} * 1000 * 1000) + (${PROJECT_VERSION_MINOR} * 1000) + ${PROJECT_VERSION_PATCH}")

[umangyadav]

The major, minor and patch should be part of the SO major version:

math(EXPR MIGRAPHX_SO_MAJOR_VERSION "(${PROJECT_VERSION_MAJOR} * 1000 * 1000) + (${PROJECT_VERSION_MINOR} * 1000) + ${PROJECT_VERSION_PATCH}")
set(MIGRAPHX_SO_VERSION ${MIGRAPHX_SO_MAJOR_VERSION}.0)

Or we if we can drop the minor version it could just be:

math(EXPR MIGRAPHX_SO_VERSION "(${PROJECT_VERSION_MAJOR} * 1000 * 1000) + (${PROJECT_VERSION_MINOR} * 1000) + ${PROJECT_VERSION_PATCH}")

I don't see any difference between the two EXPRs that you mentioned.

[pfultz2]

I don't see any difference between the two EXPRs that you mentioned.

One has the version as ${VERSION}.0 and the other is just ${VERSION}.

[nunnikri]

Are we good to approve this change? Or need further changes here? All other modules are using same format for lib version

[causten]

Merge conflict and you have not addressed Pauls comments

[frepaul]

I don't see any difference between the two EXPRs that you mentioned.

One has the version as ${VERSION}.0 and the other is just ${VERSION}.

The convention for actual library files that we have in ROCm across all the component is to have major.minor.{encoded ROCm version}.
rocm_set_soversion need to pass only major.minor. {encoded ROCm version} is handled by the function and attached as the patch number and ROCm version is passed by the build CI. Hope this clarify why we do not need the patch number passed here.