Kun Lu
Kun Lu
@mscherer
Hi @amarts , could you please take a look at this issue? Thank you very much!
Hi @amarts , Thank you so much for looking into this issue. We also tried to run the regression tests from devel branch on local s390x VMs (8 vCPU, 16g...
Hi @amarts @mscherer , Just now I checked the CI status and found that the whole regression test suite could complete after increasing the timeout, but it took nearly 2...
Cool. I'll try it. Thanks @mscherer !
Hi all, I've followed the steps in scripts `build.sh` and `regression.sh` from https://github.com/gluster/glusterfs-patch-acceptance-tests to run the regression tests on a local s390x VM (similar quota to our s390x builder), but...
Hi @mscherer @rakshithakamath94 , could you please take a look when you have some time? Thank you!
Thanks @cheshire , Sure. Test case `//tensorflow/compiler/xla/tests:bitcast_convert_test_cpu` uses two XLA backends (HOST/CPU and Interpreter) to execute the `BitcastConvertHloTest`. However, on BE machines these two backends got different results. The HOST...
Hi @sherhut , Thanks! From my understanding, the rationale behind this issue is that some tensor buffers may contain endian specific data when performing model input/output or serialization/deserialization in TensorFlow...
@Cheshire We are porting TensorFlow code on s390x for each major release and will keep testing on this code. We will raise PRs to update them if necessary. You can...