Tensile: add gfx1036 to cmake and add missing gfx1034-1036 to AsmCaps & Common.

[jerenkrantz]

Motivation

Fix TheRock build when gfx103X-all is specified.

rocBLAS build fails with error due to missing gfx1036 in Tensile CMake.

$ cmake -B build -GNinja . -DTHEROCK_AMDGPU_FAMILIES=gfx103X-all
...
[rocBLAS configure] CMake Error at .../TheRock/build/math-libs/BLAS/rocBLAS/build/virtualenv/lib/python3.12/site-packages/Tensile/Source/cmake/TensileSupportedArchitectures.cmake:92 (message):
[rocBLAS configure]   Unsupported GPU target: gfx1036
[rocBLAS configure]
[rocBLAS configure]   Supported targets are:
[rocBLAS configure]   gfx803;gfx900;gfx906;gfx908;gfx90a;gfx942;gfx950;gfx1010;gfx1011;gfx1012;gfx1030;gfx1031;gfx1032;gfx1034;gfx1035;gfx1100;gfx1101;gfx1102;gfx1103;gfx1150;gfx1151;gfx1200;gfx1201;gfx906:xnack+;gfx906:xnack-;gfx908:xnack+;gfx908:xnack-;gfx90a:xnack+;gfx90a:xnack-;gfx942:xnack+;gfx942:xnack-;gfx950:xnack+;gfx950:xnack-

Technical Details

Similar to PR #1477 (where it was for gfx1032), it appears that gfx1036 is defined in
TheRock CMake as a valid target but it isn't in Tensile.

(Of note, this PR doesn't include the test changes like #1477; if this PR looks good, I'm
happy to update the PR!)

Test Plan

Re-run cmake build of TheRock.

Test Result

rocBLAS is able to compile after applying this change.

Submission Checklist

• Look over the contributing guidelines at https://github.com/ROCm/ROCm/blob/develop/CONTRIBUTING.md#pull-requests.

[codecov-commenter]

Codecov Report

✅ All modified and coverable lines are covered by tests.

❌ Your project status has failed because the head coverage (67.07%) is below the target coverage (80.00%). You can increase the head coverage or adjust the target coverage.

Additional details and impacted files

@@             Coverage Diff             @@
##           develop    #2297      +/-   ##
===========================================
- Coverage    67.11%   67.07%   -0.04%     
===========================================
  Files          360      362       +2     
  Lines        50357    51073     +716     
  Branches      5665     5837     +172     
===========================================
+ Hits         33795    34255     +460     
- Misses       13013    13157     +144     
- Partials      3549     3661     +112     

Flag	Coverage Δ
rocBLAS	67.07% <ø> (-0.04%)	⬇️

Flags with carried forward coverage won't be shown. Click here to find out more.
see 8 files with indirect coverage changes

🚀 New features to boost your workflow:

• ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.

[TorreZuk]

@jerenkrantz related work has been merged, conflicts look simple to resolve as this is now a gfx1033 only addition so please revise PR title and after resolving conflicts I can push for quicker review and merge

[TorreZuk]

@jerenkrantz appears all this has been merged already on Tensile side. Thus this can be closed and please go ahead with rocblas side changes.

[TorreZuk]

Work already integrated into develop so closing out, thanks for the contribution.
@bstefanuk