You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Before uploading the packages on the server they are compressed by conan. The issue is that this compressing it's just taking forever with the machine having a very low cpu usage.
Is there any plan to make the compression more efficiently?
Have you read the CONTRIBUTING guide?
I've read the CONTRIBUTING guide
The text was updated successfully, but these errors were encountered:
For understanding what is happening, it would be necessary to have more data, and something closer to a benchmark that can be reproduced accurately in different machines. We are aware that zipping Conan packages, but it is not that
Coming up with such a benchmark should be relatively easy, a simple recipe that creates some large files, and conan create then time the conan upload --dry-run, and provide the full specifications of the machine running it, including important things like harddrive specifications, memory and CPU
I have been playing a bit with packaging the android_ndk.
It is true that with the default compress_level=9, the speed of zipping the binary in Windows takes around 5 minutes, resulting in a binary around 600Mb, while the system zipping takes only around 1 minute, but creates a larger binary of around 650Mbs. But lowering the compress_level=6 in Conan, will achieve similar results than the system zipping, around 650Mbs in around 1 minute.
Conan uses a higher compress level, even if it is slower, because packages are downloaded more times than zipped, so saving in storage and download transfer is worth taking a bit more time to zip the package when created.
What is your question?
Before uploading the packages on the server they are compressed by conan. The issue is that this compressing it's just taking forever with the machine having a very low cpu usage.
Is there any plan to make the compression more efficiently?
Have you read the CONTRIBUTING guide?
The text was updated successfully, but these errors were encountered: