Geekbench fails on CentOS 7 current

Thomas's Avatar

Thomas

03 Sep, 2019 04:38 PM

Hi, just downloaded GB5 on my up-to-date CentOS7 and it fails to run because evidently it is not statically linked and missing a matching version of the libstdc++ (asks for "GLIBCXX_3.4.20" specifically).

perhaps you want to provide a static variant since CentOS7 isn't entirely uncommon.

  1. 1 Posted by Artem S. Tashki... on 03 Sep, 2019 04:43 PM

    Artem S. Tashkinov's Avatar

    Try installing libstdc++ from CentOS 8 (just this library alone).

  2. 2 Posted by Thomas on 03 Sep, 2019 05:19 PM

    Thomas's Avatar

    just one tiny issue with that: CentOS 8 ain't out yet.

  3. 3 Posted by Artem S. Tashki... on 03 Sep, 2019 08:28 PM

    Artem S. Tashkinov's Avatar

    Grab a package from Fedora 28 then - it'll probably work.

  4. 4 Posted by Thomas on 03 Sep, 2019 10:07 PM

    Thomas's Avatar

    I have Fedora 30 on a bare metal host and RHEL8 in a VM: Both work with Geekbench 5 w/o problems.

    Again, the issue is, that CentOS 7 is one of the most important and popular Linux platforms and will remain so for a couple of years while RHEL8/CentOS8 mature or appear.

    To publish a benchmark that aims at being a "reference" without support for one of the most important platforms is just plain neglegence and must be fixed.

    Do it sooner rather than later.

  5. Support Staff 5 Posted by John on 04 Sep, 2019 07:20 PM

    John's Avatar

    We'll take a look to see if there's anything we can do to improve Geekbench 5 compatibility with CentOS 7. In the meantime, Geekbench 4 should work with CentOS 7 without issue.

  6. 6 Posted by Thomas on 04 Sep, 2019 07:41 PM

    Thomas's Avatar

    Hi John,

    I know Geekbench 4 works on RHEL/CentOS 7 and I have used it extensively to gather profiling information about the hardware I use.

    It's also why I was surprised to discover that Geekbench 5 doesn't work just as well there, given that it represents the vast majority of all corporate Linux workloads globally.

    Since I noticed in the past, that Geekbench 4 CPU results tended to favor Linux vs Windows on identical hardware often by quite a bit, I wanted to find out if that bias continues in GB5 or if newer code and all those stringent optimizations I had just paid for with the new version had solved that issue.

    Knowing that would have made it easier to do the full switch to GB5 and justify me rebuilding my profile results database to a GB5 baseline.

    Otherwise there'd be little reason for me to switch, when relative hardware comparability is all I want.

  7. 7 Posted by scott on 11 Sep, 2019 06:37 PM

    scott's Avatar

    I'd like to add my vote for fixing Geekbench 5 in CentOS 7.

    Using Geekbench 4 is a fine workaround for the moment, but I paid for Geekbench 5 so I'd like to be able to use it on my CentOS machine ;-). It's working great on most of my other systems.

Reply to this discussion

Internal reply

Formatting help / Preview (switch to plain text) No formatting (switch to Markdown)

Attaching KB article:

»

Attached Files

You can attach files up to 10MB

If you don't have an account yet, we need to confirm you're human and not a machine trying to post spam.

Keyboard shortcuts

Generic

? Show this help
ESC Blurs the current field

Comment Form

r Focus the comment reply box
^ + ↩ Submit the comment

You can use Command ⌘ instead of Control ^ on Mac