openSUSE Leap 15.3 Beta Performance Is Looking Good

[phoronix]

Phoronix: openSUSE Leap 15.3 Beta Performance Is Looking Good

Given the beta phase of openSUSE Leap 15.3 having started earlier this month, here are some preliminary benchmarks looking at the performance for openSUSE Leap 15.3 with the initial beta snapshot against Leap 15.2 with all stable release updates against the rolling-release openSUSE Tumbleweed.

https://www.phoronix.com/vr.php?view=29999

[onlyLinuxLuvUBack]

Please review, alma linux they already have rc1

[smartalgorithm]

wow, thank you very much for these insights... these are very helpful for me!

[Charlie68]

I'm more interested in the comparison between Leap 15.2 vs Leap 15.3 and it seems to me there are significant improvements.
By the way this is a service pack, not a new version of Leap, so I see major improvements, this release also marks alignment with SLE-SUSE so I was curious to see how it goes, it looks in good shape.

[Linuxxx]

Michael
I just noticed that Your PTS reports the setting used for "Transparent Hugepage" now, which is really nice!
Could You also consider doing something similar with "uname -a"?
This way, one could easily see whether the benchmarked kernel had PREEMPT enabled, now that with the upcoming Linux 5.12 anybody can easily set this option for themselves without recompiling the whole kernel.
Thanks in advance if You are going to implement this addition to Your already great software!

And while being on the topic of transparent hugepages:
Why on earth does openSUSE still default to "always", when everyone else (including upstream) have already moved to the much saner default of "madvise"? (No pun intended - it's really called this way!)

Those Germans really can be quite stubborn...

[Ground0]

I did a fresh beta Installation on my very old MSI M635 Notebook and wow it works :-) 390MB RAM Usage for default KDE Plasma Desktop and Web browser (As long as only one Tab) everything works and is "fast"/usable. Short spec : AMD MT-37 @ 2.00GHz (1 Core), 1GB RAM, Radeon X700-128MB RAM and a IDE SSD in.

[lsatenstein]

I write C code. The C compiler from 15.2 produces a much smaller code size than does 15.3 version. The software I write is used to parse, edit and format a /etc/fstab. One would use it to create a pretty fstab, a validated formatted fstab that would be needed after insertions and removals of partitions.

Code speed is not my requirement.

[Go_Vulkan]

These results are great, most notably because Leap 15.3 is not Tumbleweed, though the performance is quite close to the latter. Can anyone assume which new software / version helped to accomplish these results?

Those Germans really can be quite stubborn...

A stupid imputation. If you call this "stubborn", how would you call all the voters of Donald Trump? Or Boris Johnson with his idiotic Brexit?

Maybe they have a reason, maybe they just overlooked it. Whatever.