[llvm] 6c848c2 - Remove redundant environment variable XLA_FLAGS.

Jacob Hegna via llvm-commits llvm-commits at lists.llvm.org
Mon Jun 14 16:58:37 PDT 2021


Author: Jacob Hegna
Date: 2021-06-14T23:58:22Z
New Revision: 6c848c28c2f495f97a8fce879bd0624d61272d24

URL: https://github.com/llvm/llvm-project/commit/6c848c28c2f495f97a8fce879bd0624d61272d24
DIFF: https://github.com/llvm/llvm-project/commit/6c848c28c2f495f97a8fce879bd0624d61272d24.diff

LOG: Remove redundant environment variable XLA_FLAGS.

If the flag is not set, the script saved_model_aot_compile.py in tensorflow will
default it to the correct value. However, in TF 2.5, the way the value is set in
TensorFlowCompile.cmake file triggers a build error.

Reviewed By: mtrofin

Differential Revision: https://reviews.llvm.org/D103972

Added: 
    

Modified: 
    llvm/cmake/modules/TensorFlowCompile.cmake

Removed: 
    


################################################################################
diff  --git a/llvm/cmake/modules/TensorFlowCompile.cmake b/llvm/cmake/modules/TensorFlowCompile.cmake
index c4fae4f021807..3454b5f4ed759 100644
--- a/llvm/cmake/modules/TensorFlowCompile.cmake
+++ b/llvm/cmake/modules/TensorFlowCompile.cmake
@@ -32,7 +32,8 @@ function(tfcompile model tag_set signature_def_key fname cpp_class)
     tfgetmodel(${model} LLVM_ML_MODELS_ABSOLUTE)
     message("Using model at " ${LLVM_ML_MODELS_ABSOLUTE})
     add_custom_command(OUTPUT ${obj_file} ${hdr_file}
-      COMMAND "XLA_FLAGS=\"--xla_cpu_multi_thread_eigen=false\"" ${TENSORFLOW_AOT_COMPILER} aot_compile_cpu
+      COMMAND ${TENSORFLOW_AOT_COMPILER} aot_compile_cpu
+            --multithreading false
             --dir ${LLVM_ML_MODELS_ABSOLUTE}
             --tag_set ${tag_set}
             --signature_def_key ${signature_def_key}


        


More information about the llvm-commits mailing list