tensorflow：gcc：error：無法識別的命令行選項'--copt = -msse4.2'

我導致official tensorflow guide從源代碼安裝它，爲了解決警告，如The TensorFlow library wasn't compiled to use SSE4.1 instructions, but these are available on your machine and could speed up CPU computations，好像他們已經省略了配置期間如何指定優化標誌的最重要部分（./configure）。tensorflow：gcc：error：無法識別的命令行選項'--copt = -msse4.2'

... 
gcc: error: unrecognized command line option '--copt=-mavx' 
gcc: error: unrecognized command line option '--copt=-mavx2' 
gcc: error: unrecognized command line option '--copt=-mfma' 
gcc: error: unrecognized command line option '--copt=-msse4.1' 
gcc: error: unrecognized command line option '--copt=-msse4.2' 
gcc: error: unrecognized command line option '--copt=-mavx' 
gcc: error: unrecognized command line option '--copt=-mavx2' 
gcc: error: unrecognized command line option '--copt=-mfma' 
gcc: error: unrecognized command line option '--copt=-msse4.1' 
gcc: error: unrecognized command line option '--copt=-msse4.2' 
Target //tensorflow/tools/pip_package:build_pip_package failed to build

什麼是指定的優化參數的正確方法：用無效的配置（bazel build --config=opt //tensorflow/tools/pip_package:build_pip_package）建設時

餵養它們錯誤地導致了下面的錯誤導致了這樣的錯誤？

來源

2017-08-27 reubenjohn

機會是你可能已經錯過了部分地方the guide規定：

We recommend accepting the default (-march=native), which will optimize the generated code for your local machine's CPU type. However, if you are building TensorFlow on one CPU type but will run TensorFlow on a different CPU type, then consider specifying a more specific optimization flag as described in the gcc documentation .

而在gcc documentation，你會發現列表，如：

-mmmx -mno-mmx -msse -mno-sse -msse2 -mno-sse2 -msse3 -mno-sse3 -mssse3 -mno-ssse3 -msse4.1 -mno-sse4.1 -msse4.2 -mno-sse4.2 -msse4 -mno-sse4 -mavx -mno-avx -maes -mno-aes -mpclmul -mno-pclmul -msse4a -mno-sse4a -mfma4 -mno-fma4 -mxop -mno-xop -mlwp -mno-lwp -m3dnow -mno-3dnow -mpopcnt -mno-popcnt -mabm -mno-abm

來源

2017-08-27 13:20:06 reubenjohn

也有類似的問題，把下面的在我的「./configure」中：

Please specify optimization flags to use during compilation when bazel option "--config=opt" is specified [Default is -march=native]: 
--copt=-mavx --copt=-mavx2 --copt=-mfma --copt=-mfpmath=both --copt=-msse4.1 --copt=-msse4.2

As lin由gub doc中的reubenjohn ked'-march = whatever'與'-msse4.1'處於同一級別，因此'--copt ='必須被移除，因爲這些選項實際上直接饋送給GCC。

解決通過將改爲：

Please specify optimization flags to use during compilation when bazel option "--config=opt" is specified [Default is -march=native]: 
-mavx -mavx2 -mfma -mfpmath=both -msse4.1 -msse4.2

來源

2017-09-15 01:58:53

tensorflow：gcc：error：無法識別的命令行選項'--copt = -msse4.2'

回答

相關問題