CentOS Stream 9 Improves Performance For Intel Xeon, AMD EPYC

[Paradigm Shifter]

Its faster because its new and they said so doesnt make it any more attractive to anyone.

My thinking also.

Improvements in speed can come from external projects (see Michaels' regular updates regarding improvements to Zstd, for example) which would translate across all distributions if the latest version was used... as the OS defaults are used for those tests. CentOS8 now uses quite an old version of Zstd.

[cynic]

what happened to all the comments in this discussion? were they censored?

[mSparks]

Really dont like the changes they made to CentOS, So migrated mostly everything over to the Oracle Linux fork.

Its faster because its new and they said so doesnt make it any more attractive to anyone.

Not sure I will ever go back to Redhat or IBM for servers after that debacle. No particular "salt" as some seem to call it, just as much respect for them as they had for their users when they pulled the rug out from everyone.

[mether]

just wanted to let you know that I'm not a pimply script kiddie that discovered Linux yesterday.

Who you are is irrelevant to the topic and certainly doesn't justify flippant responses.

[mether]

I'd really like to know more about why Centos Stream is faster. Here's my motivation.
...
If the performance improvements are due to, for example, kernel scheduling improvements then we could likely benefit from upgrading our host OSes to Centos Stream. But if the benefits are due to newer glibc versions, then we're unlikely to see any benefit. And obviously if they 're due to newer compilers we won't see any benefit, as we don't get to recompile the vendor tools...

There has certainly been a good amount of kernel performance improvements between releases but it is not possible to say whether a particular vendor tool would benefit from it. You are going to have to benchmark that anyway before you schedule any upgrades.

[igxqrrl]

I'd really like to know more about why Centos Stream is faster. Here's my motivation.

We run vendor software on hundreds of servers. This software requires Centos6 or Centos7. Because of that mixed environment we run everything in containers.

If the performance improvements are due to, for example, kernel scheduling improvements then we could likely benefit from upgrading our host OSes to Centos Stream. But if the benefits are due to newer glibc versions, then we're unlikely to see any benefit. And obviously if they 're due to newer compilers we won't see any benefit, as we don't get to recompile the vendor tools...

[mether]

<sarcasm> oh, thanks for letting me know, very kind of you! I wasn't aware of that </sarcasm>

No sarcasm warranted here. You should have just gone with the actual reply and skipped this part

[cynic]

If it is just a dozen, https://www.redhat.com/en/blog/new-y...ys-access-rhel would cover you completely, just FYI.

<sarcasm> oh, thanks for letting me know, very kind of you! I wasn't aware of that </sarcasm>

seriously speaking, that dozen were just those already migrated to CentOS 8.
also, we fire tens of VM for tests and some of use use it as a workstation OS.

[skeevy420]

To add to what cynic said, most all the salt is over the name of the project because it went from being a free RHEL clone to being an intermediate between Fedora and Red Hat. IMHO, it should have been called anything else other than CentOS. I'd have gone with Cowboy, Trilby, or Bowler to keep with the hat theme.

Cowboy Stream would have been an apt name because you've wrangled up all the Rawhide packages from the Fedora Ranch and you're in the process of taking them to the Red Hat market for mass consumption.

[mether]

I had a dozen of newly installed server just when they announced the drop of the CentOS support and I think that is enough reason

If it is just a dozen, https://www.redhat.com/en/blog/new-y...ys-access-rhel would cover you completely, just FYI.