Ubuntu vs. Fedora vs. openSUSE vs. Manjaro vs. Clear Linux On Intel's Core i9 7900X

[phoronix]

Phoronix: Ubuntu vs. Fedora vs. openSUSE vs. Manjaro vs. Clear Linux On Intel's Core i9 7900X

Continuing on with our Core i9 7900X Linux benchmarking this week are some numbers when testing this ten-core Skylake-X processor on various Linux distributions under an array of different workloads. Tested for this roundup was Ubuntu 16.04.2 LTS, Ubuntu 17.04, Fedora 26, openSUSE Tumbleweed, Manjaro 17.0.2, and Clear Linux 16160.

http://www.phoronix.com/vr.php?view=24879

[OlafLostViking]

Would it be possible to look into these strange differences and find a reason for *why* the distributions are diverging by such a huge amount? Thanks.

[DebianXFCE Jr]

As for me Intel is dead. For now it is rather good. Let them suffer and think about releasing something that will be both good in technical and economical terms - performance2price.
For to long there was no big revolution on consumer desktop market.

[Michael]

Would it be possible to look into these strange differences and find a reason for *why* the distributions are diverging by such a huge amount? Thanks.

Unfortunately, not with the limited/time resources I have available... The time needed to thoroughly do such would likely never be recouped in any notable increase in ad revenue or many new premium subscribers.

[chuckula]

Thanks for the additional numbers. Clearly the compiler is quite important for tuning high-performance workloads just right. That SciKit benchmark is very interesting as well since it might be the first real workload that's actually utilizing AVX-512 that I've seen tested anywhere for these new chips.

[tigerroast]

Ubuntu and Manjaro have older kernel versions. The tests were done using a solid-state drive, and Manjaro comes with BFQ pre-installed (at least, it did when I tried Manjaro last). I believe BFQ can cause performance issues when on flash memory as opposed to spinning disks.

That's not to discredit the validity of the tests; this is on Manjaro's head.

[bug77]

Would it be possible to reverse the sort order of the results? Maybe it's just me, but having the winner at the bottom just doesn't look right.

[Michael]

Would it be possible to reverse the sort order of the results? Maybe it's just me, but having the winner at the bottom just doesn't look right.

Others have commented in the past they like best on bottom... But its all personal pref. But OpenBenchmarking.org makes it easy either way. Here you go: http://openbenchmarking.org/result/1...X158&obr_sor=y

[bug77]

Others have commented in the past they like best on bottom... But its all personal pref. But OpenBenchmarking.org makes it easy either way. Here you go: http://openbenchmarking.org/result/1...X158&obr_sor=y

Best at the bottom, go figure. But ok.

[chuckula]

Hey Mike, thanks for all the work you've been doing with this new platform and all your work around the RyZen launch too.
Just did a $35 tip since the subscription requires a full paypal account, but we appreciate the effort and this is really the only place on the web where you can see how Linux runs on different hardware in great detail.