A Look At The Clear Linux Performance For July 2018

[phoronix]

Phoronix: A Look At The Clear Linux Performance For July 2018

Given our fascination with Intel's Clear Linux performance in the plethora of performance benchmarks we frequently run at Phoronix and this open-source operating system being maintained in a rolling-release style, here's a look at how the performance for this x86_64 Linux distribution evolved over the past month...

http://www.phoronix.com/scan.php?page=news_item&px=Clear-Linux-July-2018-Perform

[Herem]

The fastest just got even faster (mostly) impressive stuff.

Michael even though this distribution is tuned for Intel processors, it would have been nice to see some AMD processors as well. That would have shown if the improvements were general or Intel specific.

[perpetually high]

Long-time linux user who just recently became serious about compiling a kernel tailored to their machine, and really appreciating it. There's lots of goodies in there.

By switching Processor type -> Processor family in kernel config from "Generic x86-64" to "Core 2 or newer" it adds 5 new config settings. For reference:

CONFIG_X86_INTERNODE_CACHE_SHIFT=6

CONFIG_X86_L1_CACHE_SHIFT=6

CONFIG_X86_INTEL_USERCOPY=y

CONFIG_X86_USE_PPRO_CHECKSUM=y

CONFIG_X86_P6_NOP=y

So the jig up Clear Linux. That, along with some common-sense tweaks, plus CFLAGS set to -O2 -march=native of course, and voila! You've got Clear Linux. Mystery solved on the serious gains.

[andyprough]

Even with some performance pullbacks in a small portion of the tests, the performance is still likely well ahead of the likes of Fedora and Ubuntu for out-of-the-box performance.

And, it's very likely that ClearLinux can't perform 1/10th of the actual functions of Fedora or Ubuntu.

The question is, if ClearLinux had the overhead of an actual general purpose distribution, would it still test so well?

[perpetually high]

And, it's very likely that ClearLinux can't perform 1/10th of the actual functions of Fedora or Ubuntu.

The question is, if ClearLinux had the overhead of an actual general purpose distribution, would it still test so well?

Not to mention, it's using all the latest versions of core components, pretty much taking all the glory on the soon-to-come gains when other distributions eventually adopt them. Not hating on Clear Linux, I like what they're doing. I'm just annoyed and bitter as an Ubuntu user, that's all.

[AsuMagic]

Not to mention, it's using all the latest versions of core components, pretty much taking all the glory on the soon-to-come gains when other distributions eventually adopt them. Not hating on Clear Linux, I like what they're doing. I'm just annoyed and bitter as an Ubuntu user, that's all.

Arch testing might be the most bleeding edge aside of Clear as of common distros, and I can bet you're never going to get any close to Clear's performance with it. Plus you're claiming the kernel can be tweaked to run as fast as Clear's but there are Clear linux packages on some distros and from what I've seen they're not really this much faster. The only real comparison would be gentoo with -O2 -march=native but it's still going to get stomped.

[nomadewolf]

Arch testing might be the most bleeding edge aside of Clear as of common distros, and I can bet you're never going to get any close to Clear's performance with it. Plus you're claiming the kernel can be tweaked to run as fast as Clear's but there are Clear linux packages on some distros and from what I've seen they're not really this much faster. The only real comparison would be gentoo with -O2 -march=native but it's still going to get stomped.

openSUSE Factory is always the most bleeding edge.

[Venemo]

GraphicsMagick's image resizing was another test case running slightly slower across the tested systems

What? The graph above this quote clearly shows that "End July" is faster than "Start July".

[geearf]

Long-time linux user who just recently became serious about compiling a kernel tailored to their machine, and really appreciating it. There's lots of goodies in there.

By switching Processor type -> Processor family in kernel config from "Generic x86-64" to "Core 2 or newer" it adds 5 new config settings. For reference:

CONFIG_X86_INTERNODE_CACHE_SHIFT=6

CONFIG_X86_L1_CACHE_SHIFT=6

CONFIG_X86_INTEL_USERCOPY=y

CONFIG_X86_USE_PPRO_CHECKSUM=y

CONFIG_X86_P6_NOP=y

So the jig up Clear Linux. That, along with some common-sense tweaks, plus CFLAGS set to -O2 -march=native of course, and voila! You've got Clear Linux. Mystery solved on the serious gains.

Nope

Not to mention, it's using all the latest versions of core components, pretty much taking all the glory on the soon-to-come gains when other distributions eventually adopt them. Not hating on Clear Linux, I like what they're doing. I'm just annoyed and bitter as an Ubuntu user, that's all.

And Nope.

3rd try's the charm?

[andyprough]

Not to mention, it's using all the latest versions of core components, pretty much taking all the glory on the soon-to-come gains when other distributions eventually adopt them. Not hating on Clear Linux, I like what they're doing. I'm just annoyed and bitter as an Ubuntu user, that's all.

When you don't have to support a large universe of software, or really anything at all besides flatpaks and "bundles", you can probably get away with running the latest/greatest of everything.