Re: [I] [VL] Compiling 1.4.0 on X86 fails [incubator-gluten]

Tue, 29 Jul 2025 01:42:04 -0700 


kevinw66 commented on issue #10102:
URL: 
https://github.com/apache/incubator-gluten/issues/10102#issuecomment-3131319563

   > > > > > [@A-little-bit-of-data](https://github.com/A-little-bit-of-data) 
Have you tried export `CPU_TARGET="aarch64"` and 
`VCPKG_FORCE_SYSTEM_BINARIES=1`? Also, If all CPU cores are used for 
compilation, ARM is more likely to crash than x86. The reason is still unclear 
for now, but you can try reducing the value of `NUM_THREADS`.
   > > > > 
   > > > > 
   > > > > [@kevinw66](https://github.com/kevinw66) Hello, I have a question. I 
have used the images of `vcpkg-centos-8, centos-8-jdk11, and 
1.4-vcpkg-centos-8`, but there are these errors.
   > > > > ```
   > > > > [137/647] Building CXX object 
velox/buffer/CMakeFiles/velox.dir/__/vector/VectorStream.cpp.o
   > > > > In file included from 
/incubator-gluten/dev/vcpkg/vcpkg_installed/arm64-linux-release/include/xsimd/xsimd.hpp:55,
   > > > >                  from 
/incubator-gluten/ep/build-velox/build/velox_ep/./velox/common/base/SimdUtil.h:24,
   > > > >                  from 
/incubator-gluten/ep/build-velox/build/velox_ep/./velox/common/memory/RawVector.h:21,
   > > > >                  from 
/incubator-gluten/ep/build-velox/build/velox_ep/./velox/common/memory/Scratch.h:19,
   > > > >                  from 
/incubator-gluten/ep/build-velox/build/velox_ep/./velox/common/memory/ByteStream.h:18,
   > > > >                  from 
/incubator-gluten/ep/build-velox/build/velox_ep/./velox/vector/VectorStream.h:22,
   > > > >                  from 
/incubator-gluten/ep/build-velox/build/velox_ep/velox/vector/VectorStream.cpp:16:
   > > > > 
/incubator-gluten/dev/vcpkg/vcpkg_installed/arm64-linux-release/include/xsimd/arch/xsimd_scalar.hpp:
 In function ‘std::pair<float, float> xsimd::sincos(float)’:
   > > > > 
/incubator-gluten/dev/vcpkg/vcpkg_installed/arm64-linux-release/include/xsimd/arch/xsimd_scalar.hpp:981:54:
 note: parameter passing for argument of type ‘std::pair<float, float>’ when 
C++17 is enabled changed to match C++14 in GCC 10.1
   > > > >   981 |     inline std::pair<float, float> sincos(float val) noexcept
   > > > >       |                                                      ^~~~~~~~
   > > > > 
/incubator-gluten/dev/vcpkg/vcpkg_installed/arm64-linux-release/include/xsimd/arch/xsimd_scalar.hpp:
 In function ‘std::pair<double, double> xsimd::sincos(double)’:
   > > > > 
/incubator-gluten/dev/vcpkg/vcpkg_installed/arm64-linux-release/include/xsimd/arch/xsimd_scalar.hpp:988:57:
 note: parameter passing for argument of type ‘std::pair<double, double>’ when 
C++17 is enabled changed to match C++14 in GCC 10.1
   > > > >   988 |     inline std::pair<double, double> sincos(double val) 
noexcept
   > > > > ```
   > > > > 
   > > > > 
   > > > >     
   > > > >       
   > > > >     
   > > > > 
   > > > >       
   > > > >     
   > > > > 
   > > > >     
   > > > >   
   > > > > What needs to be adjusted to compile properly? The following 
variables have been used,
   > > > > ```
   > > > > export VCPKG_FORCE_SYSTEM_BINARIES=1
   > > > > export CPU_TARGET="aarch64"
   > > > > export NUM_THREADS=4
   > > > > ```
   > > > > 
   > > > > 
   > > > >     
   > > > >       
   > > > >     
   > > > > 
   > > > >       
   > > > >     
   > > > > 
   > > > >     
   > > > >   
   > > > > My compiling machine is 8C
   > > > 
   > > > 
   > > > [@A-little-bit-of-data](https://github.com/A-little-bit-of-data) I 
think these are just warnings and based on my current testing, they do not 
affect our main functions, so I ignored them. These warnings also appears in 
our CI environment: 
https://github.com/apache/incubator-gluten/actions/runs/16491898666/job/46628360332#step:5:1799.
   > > > <img alt="Image" width="1889" height="952" 
src="https://private-user-images.githubusercontent.com/31196226/470230088-1eb86399-1d6b-4c15-8a1e-6919f9d01032.png?jwt=eyJhbGciOiJIUzI1NiIsInR5cCI6IkpXVCJ9.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.nQLGIcq6Q3Znw3OceowyqSorqSAFG8bQVCbppQTfaXU";>
   > 
   > [@kevinw66](https://github.com/kevinw66) Okay, thanks for the tip. My ARM 
machine currently doesn't have enough memory to compile. I also have a question 
about the minimum memory required for this compilation. I can compile with less 
than 32GB on an x86 machine using the same image, but there are only 16GB 
machines for the ARM architecture. Can you give me a more accurate memory 
reference so I can apply for a machine? Thanks!
   
   @A-little-bit-of-data Actually 16GB is enough for me


-- 
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.

To unsubscribe, e-mail: [email protected]

For queries about this service, please contact Infrastructure at:
[email protected]


---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]

Reply via email to