Comment out qlinear reuse matcher

[CharlieL7]

Comment out the qlinear_reused matcher because of an accuracy error for quantized resnet50

[codecov]

Codecov Report

All modified and coverable lines are covered by tests ✅

Project coverage is 92.00%. Comparing base (d15a23a) to head (280a304).
Report is 9 commits behind head on release/rocm-rel-6.2.

Additional details and impacted files

@@                  Coverage Diff                  @@
##           release/rocm-rel-6.2    #3264   +/-   ##
=====================================================
  Coverage                 92.00%   92.00%           
=====================================================
  Files                       489      489           
  Lines                     19368    19344   -24     
=====================================================
- Hits                      17819    17797   -22     
+ Misses                     1549     1547    -2     

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

[umangyadav]

Is this required if we were to pull in changes from MLIR ?

[CharlieL7]

Is this required if we were to pull in changes from MLIR ?

We won't hit the verify "accuracy" bug for the quantized resnet50 model with the change from MLIR. However, the matcher itself is questionable. I'm proposing to remove it permanently here in #3269.

[pfultz2]

We won't hit the verify "accuracy" bug for the quantized resnet50 model with the change from MLIR.

If we dont hit an accuracy error then we shouldn't remove it.

[CharlieL7]

We won't hit the verify "accuracy" bug for the quantized resnet50 model with the change from MLIR.

If we dont hit an accuracy error then we shouldn't remove it.

Matcher accuracy is brittle, if we change the randomized values slightly it will even fail without MLIR.

[CharlieL7]

Closing: Should be able to pass verify accuracy with this compiler pass enable with the MLIR fix ROCm/rocMLIR#1567