Balaji V. Iyer. comments

Results 5 comments of


                                            Balaji V. Iyer.

Conv2D with XLA jit_compile=True fails to run

Hello Colin, I see that you are using cuDNN version 8100 and that could be the reason for your issue. Please see here (https://github.com/tensorflow/tensorflow/blob/master/tensorflow/compiler/xla/stream_executor/cuda/cuda_dnn.cc#L5309) . Is it possible for you...

Broadcast_in_Dim e2e test fails to compile in VMVX backend

Note; Same error is found in `stablehlo.reduce` for f16: https://github.com/openxla/iree/issues/16519

[Clang]Check for a valid Index in array before getting it

> -fexperimental-new-constant-interpreter When I have this enabled in the test case, it will evaluate it and find out that the array index is greater than array size and thus cannot...

[Clang]Check for a valid Index in array before getting it

> This change needs a release note. Please add an entry to `clang/docs/ReleaseNotes.rst` in the section the most adapted to the change, and referencing any Github issue this change fixes....

[Clang]Check for a valid Index in array before getting it

NOTE: I had to force push due to a merge issue (Very sorry about it). All the changes are there, just that I squashed two commits into 1.