Lightgbm: Performing Test MM_MALLOC - Failed

Created on 25 May 2020  路  3Comments  路  Source: microsoft/LightGBM

Hi!

Unable to build GPU support. Please help.

I had tried VS 2017 with CUDA 8.0, 9.0, 10.0, 10.1 and 10.2 and also VS 2019 with CUDA 10.2. I am installing on Windows 10 inside Anaconda Prompt.

Below reports show details of my software and failure report.

image

image

image

image

Most helpful comment

@raphay3l The MM_MALLOC macro failing is not an issue as it is related to choosing the correct header library containing malloc (unrelated to your GPU issue - also if it "fails", it does not mean it is an issue, it just means it chooses the correct header library).

Unable to build GPU support.

What is the error you encountered while compiling LightGBM with GPU support?

All 3 comments

@raphay3l thanks for using LightGBM and for the report! Performing test MM_MALLOC - failed does not necessarily mean the installation is broken. It is just a check to see how the library should approach malloc-type operations (https://github.com/microsoft/LightGBM/blob/5c0baf6fe008a9d59d604d0dd2742d4be4f9d6e1/include/LightGBM/utils/common.h#L29-L41)

Did you experience some other issue running LightGBM code? If so can you please share the code you ran and the error it produced?

@raphay3l The MM_MALLOC macro failing is not an issue as it is related to choosing the correct header library containing malloc (unrelated to your GPU issue - also if it "fails", it does not mean it is an issue, it just means it chooses the correct header library).

Unable to build GPU support.

What is the error you encountered while compiling LightGBM with GPU support?

@Laurae2 @jameslamb thank you for swift response. Noted. I am closing this issue - resolved.

Was this page helpful?
0 / 5 - 0 ratings