[opensuse-factory] vanilla Leap 15.1 boosts gaming performance
Hi, Jokes aside the kernel in 15.1 switched to CONFIG_PREEMPT_VOLUNTARY a month ago. That setting is supposed to find a better balance between desktop performance and throughput on servers. So I wonder if beta testers of 15.1 noticed any difference? Was there a change in power consumption for example? Don't wait upgrading to 15.1 until the release, there are only a few weeks left to get fixes in! cu Ludwig -- (o_ Ludwig Nussel //\ V_/_ http://www.suse.com/ SUSE Linux GmbH, GF: Felix Imendörffer, Mary Higgins, Sri Rasiah, HRB 21284 (AG Nürnberg) -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org
Am Dienstag, 2. April 2019, 10:58:16 CEST schrieb Ludwig Nussel:
So I wonder if beta testers of 15.1 noticed any difference? Was there a change in power consumption for example?
Are there any benchmarks around that would help getting a comparison? I could run them on various gaming machines with NVIDIA hardware (GTX10xx, RTX20xx, ...) Regards, vinz. -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org
Hi Vinz, the ususal Phoronix Benchmark [1] suite should be sufficient and create comparable results. Cheers, Bernd [1] - https://www.phoronix-test-suite.com/ Am Di., 2. Apr. 2019 um 11:19 Uhr schrieb Vinzenz Vietzke <vinz@vinzv.de>:
Am Dienstag, 2. April 2019, 10:58:16 CEST schrieb Ludwig Nussel:
So I wonder if beta testers of 15.1 noticed any difference? Was there a change in power consumption for example?
Are there any benchmarks around that would help getting a comparison? I could run them on various gaming machines with NVIDIA hardware (GTX10xx, RTX20xx, ...)
Regards, vinz.
-- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org
-- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org
Am 02.04.19 um 11:36 schrieb John Salvatore Fontanelli:
Hi Vinz,
the ususal Phoronix Benchmark [1] suite should be sufficient and create comparable results.
You are one day late for April's fool jokes ;-) -- Stefan Seyfried "For a successful technology, reality must take precedence over public relations, for nature cannot be fooled." -- Richard Feynman -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org
Hi Stefan, fool's on you, this is about measuring the CONFIG_PREEMPT_VOLUNTARY kernel change :-) Cheers, Bernd Am Di., 2. Apr. 2019 um 14:03 Uhr schrieb Stefan Seyfried <stefan.seyfried@googlemail.com>:
Am 02.04.19 um 11:36 schrieb John Salvatore Fontanelli:
Hi Vinz,
the ususal Phoronix Benchmark [1] suite should be sufficient and create comparable results.
You are one day late for April's fool jokes ;-) -- Stefan Seyfried
"For a successful technology, reality must take precedence over public relations, for nature cannot be fooled." -- Richard Feynman -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org
-- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org
Am 02.04.19 um 14:07 schrieb John Salvatore Fontanelli:
Hi Stefan,
fool's on you, this is about measuring the CONFIG_PREEMPT_VOLUNTARY kernel change :-)
No. The April's fools joke is obviously talking about "PTS being sufficient and creating comparable results".
Cheers, Bernd
Am Di., 2. Apr. 2019 um 14:03 Uhr schrieb Stefan Seyfried <stefan.seyfried@googlemail.com>:
Am 02.04.19 um 11:36 schrieb John Salvatore Fontanelli:
Hi Vinz,
the ususal Phoronix Benchmark [1] suite should be sufficient and create comparable results.--
Stefan Seyfried
"For a successful technology, reality must take precedence over public relations, for nature cannot be fooled." -- Richard Feynman -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org
Am Dienstag, 2. April 2019, 15:34:09 CEST schrieb Stefan Seyfried:
The April's fools joke is obviously talking about "PTS being sufficient and creating comparable results".
If only I could find the detailed blogpost again which explained why PTS is unreliable... That aside: I know PTS but what I meant to ask Ludwig what he (and the Leap release team) would prefer for test results. Regards, vinz. -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org
On Tuesday 2019-04-02 16:02, Vinzenz Vietzke wrote:
Am Dienstag, 2. April 2019, 15:34:09 CEST schrieb Stefan Seyfried:
The April's fools joke is obviously talking about "PTS being sufficient and creating comparable results".
I know PTS but what I meant to ask Ludwig what he (and the Leap release team) would prefer for test results.
Phoronix generally seems to focus on time/work. (Time-to-completion for a fixed task; or number of operations for a fixed time quantum) So if you can make whatever software you are analyzing trade CPU for memory (use less CPU, use more memory), you will score "higher" in most tallies. At least that is my impression of PTS results as published on Phoronix. -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org
On Dienstag, 2. April 2019 17:12:39 CEST Jan Engelhardt wrote:
On Tuesday 2019-04-02 16:02, Vinzenz Vietzke wrote:
Am Dienstag, 2. April 2019, 15:34:09 CEST schrieb Stefan Seyfried:
The April's fools joke is obviously talking about "PTS being sufficient and creating comparable results".
I know PTS but what I meant to ask Ludwig what he (and the Leap release team) would prefer for test results.
Phoronix generally seems to focus on time/work. (Time-to-completion for a fixed task; or number of operations for a fixed time quantum)
So if you can make whatever software you are analyzing trade CPU for memory (use less CPU, use more memory), you will score "higher" in most tallies. At least that is my impression of PTS results as published on Phoronix.
The problem with Phoronix is not the PTS, but the "evaluation" done in the articles. The PTS outputs raw metrics. These metrics include both "speed" and "efficiency" (e.g. frames/second/watt, i.e frames/joule) indicators. Of course the latter depends on ways to e.g. measure power draw. For desktops, this requires external power loggers with realtime data via USB/serial, while for Laptops the internal measurement suffices. So you can get both metrics, time to completion and energy per workload. *Both* are of interest here. If you use the same hardware, the same distribution at a given version and just exchange the kernel itself, you are quite safe to argue any statistical relevant changes are due to the change of kernel. When you manage to pick the same kernel version with just different PREEMPT configs, you can attribute differences to PREEMPT. The last point is where Phoronix regularly fails, it changes to many variables, so you can assess there *is* a difference, but it can not identify a reason for it. The PTS is just a tool. Use it correctly, and you get valid results. Kind regards, Stefan-- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org
Vinzenz Vietzke schrieb:
Am Dienstag, 2. April 2019, 15:34:09 CEST schrieb Stefan Seyfried:
The April's fools joke is obviously talking about "PTS being sufficient and creating comparable results".
If only I could find the detailed blogpost again which explained why PTS is unreliable...
That aside: I know PTS but what I meant to ask Ludwig what he (and the Leap release team) would prefer for test results.
I am not an expert with benchmarks so I can't really answer that question. Looks like several people here have experience in that area though. Maybe some team could be formed to watch over various performance aspects of the release. Raw game rendering speed being just one of many :-) What matters to your customers most for example? cu Ludwig -- (o_ Ludwig Nussel //\ V_/_ http://www.suse.com/ SUSE Linux GmbH, GF: Felix Imendörffer, Mary Higgins, Sri Rasiah, HRB 21284 (AG Nürnberg) -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org
On Tue, 2019-04-02 at 11:19 +0200, Vinzenz Vietzke wrote:
Am Dienstag, 2. April 2019, 10:58:16 CEST schrieb Ludwig Nussel:
So I wonder if beta testers of 15.1 noticed any difference? Was there a change in power consumption for example?
Are there any benchmarks around that would help getting a comparison? I could run them on various gaming machines with NVIDIA hardware (GTX10xx, RTX20xx, ...)
Many games have performance benchmarks, so you might be able to use those instead of synthetic benchmarks. Robert -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org
Hi, On 02.04.19 10:58, Ludwig Nussel wrote:
Hi,
Jokes aside the kernel in 15.1 switched to CONFIG_PREEMPT_VOLUNTARY a month ago. That setting is supposed to find a better balance between desktop performance and throughput on servers.
So I wonder if beta testers of 15.1 noticed any difference? Was there a change in power consumption for example?
Don't wait upgrading to 15.1 until the release, there are only a few weeks left to get fixes in!
cu Ludwig
of course the switch is nice to have, yet i wonder if having a -desktop kernel back (with the distinguished roles in the installer) would be an option, to tune it even more for an improved user experience. Seasoned users can switch between -default and -desktop by choosing the right kernel package (or even install them side by side and boot the right bootloader entry). Just an idea :) Greetings, Tobias -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org
On Tuesday 2019-04-02 16:35, Tobias Klausmann wrote:
of course the switch is nice to have, yet i wonder if having a -desktop kernel back (with the distinguished roles in the installer) would be an option, to tune it even more for an improved user experience. Seasoned users can switch between -default and -desktop by choosing the right kernel package (or even install them side by side and boot the right bootloader entry). Just an idea :)
The current -default is what was once -desktop. Kind of. The "server-ish" k-default does not exist any longer. -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org
On 02.04.19 16:43, Jan Engelhardt wrote:
On Tuesday 2019-04-02 16:35, Tobias Klausmann wrote:
of course the switch is nice to have, yet i wonder if having a -desktop kernel back (with the distinguished roles in the installer) would be an option, to tune it even more for an improved user experience. Seasoned users can switch between -default and -desktop by choosing the right kernel package (or even install them side by side and boot the right bootloader entry). Just an idea :) The current -default is what was once -desktop. Kind of. The "server-ish" k-default does not exist any longer.
Well it has a way lower timer frequency (CONFIG_HZ) and if i am not mistaken it had (until the change by Ludwig) CONFIG_PREEMPT_NONE, so for me the -default always looked more like the *old* -default flavor. Actually if i had the freedom to choose, i would even go with a -desktop-intel and -dektop-amd divided kernel. -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org
On Tue, Apr 02, 2019 at 04:50:47PM +0200, Tobias Klausmann wrote:
On 02.04.19 16:43, Jan Engelhardt wrote:
On Tuesday 2019-04-02 16:35, Tobias Klausmann wrote:
of course the switch is nice to have, yet i wonder if having a -desktop kernel back (with the distinguished roles in the installer) would be an option, to tune it even more for an improved user experience. Seasoned users can switch between -default and -desktop by choosing the right kernel package (or even install them side by side and boot the right bootloader entry). Just an idea :) The current -default is what was once -desktop. Kind of. The "server-ish" k-default does not exist any longer.
Well it has a way lower timer frequency (CONFIG_HZ) and if i am not mistaken it had (until the change by Ludwig) CONFIG_PREEMPT_NONE, so for me the -default always looked more like the *old* -default flavor. Actually if i had the freedom to choose, i would even go with a -desktop-intel and -dektop-amd divided kernel.
I suppose the confusion here is caused by the fact that we switched to PREEMPT_VOLUNTARY from PREEMPT on Tumbleweed and at (almost) the same time we switched openSUSE-15.1 to PREEMPT_VOLUNTARY from PREEMPT_NONE (inherited from SLE15-SP1). Actually, there were also some architectures which had different value for historical reasons. See bsc#1125004 for details. For the record, there is already a request for an alternative kernel flavor with PREEMPT in bsc#1129966 (but with a different motivation than gaming). I can't really say if the difference is big enough to justify reintroducing an alternative flavor; in my use cases, I never observed any visible difference. Michal Kubecek -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org
Am Dienstag, 2. April 2019, 10:58:16 CEST schrieb Ludwig Nussel:
So I wonder if beta testers of 15.1 noticed any difference? Was there a change in power consumption for example?
Don't wait upgrading to 15.1 until the release, there are only a few weeks left to get fixes in!
I just tried the latest build 441.4 but struggled at installing proprietary NVIDIA drivers: "Nothing provides ksym(default:kmem_cache_free) = 5ddb1c65 needed by nvidia- gfxG05-default-418.56_k4.12.14_lp151.22-lp151.9.1.x86_64" (Same issue with G04 and G03) Is there a way besides nouveau driver to get Nvidia graphics to work? Regards, vinz. -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org
On Tue, 2019-04-02 at 17:19 +0200, Vinzenz Vietzke wrote:
I just tried the latest build 441.4 but struggled at installing proprietary NVIDIA drivers:
"Nothing provides ksym(default:kmem_cache_free) = 5ddb1c65 needed by nvidia- gfxG05-default-418.56_k4.12.14_lp151.22-lp151.9.1.x86_64" (Same issue with G04 and G03)
Is there a way besides nouveau driver to get Nvidia graphics to work?
Regards, vinz.
Maybe nvidia-installer instead of the Repo? -- Marcel Kühlhorn Have a lot of fun!
Op dinsdag 2 april 2019 17:19:02 CEST schreef Vinzenz Vietzke:
Am Dienstag, 2. April 2019, 10:58:16 CEST schrieb Ludwig Nussel:
So I wonder if beta testers of 15.1 noticed any difference? Was there a change in power consumption for example?
Don't wait upgrading to 15.1 until the release, there are only a few weeks left to get fixes in!
I just tried the latest build 441.4 but struggled at installing proprietary NVIDIA drivers:
"Nothing provides ksym(default:kmem_cache_free) = 5ddb1c65 needed by nvidia- gfxG05-default-418.56_k4.12.14_lp151.22-lp151.9.1.x86_64" (Same issue with G04 and G03)
Is there a way besides nouveau driver to get Nvidia graphics to work?
Regards, vinz. Most likely the nvidia required kernel version does not match the one installed. Leap 15.1 is still moving heavily, the new nvidia packages may not yet have been built against the newer kernel. NVIDIA's .run installer should work though
-- Gertjan Lettink a.k.a. Knurpht openSUSE Board Member openSUSE Forums Team -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org
Am Dienstag, 2. April 2019, 17:36:08 schrieb Knurpht-openSUSE:
I just tried the latest build 441.4 but struggled at installing proprietary NVIDIA drivers:
"Nothing provides ksym(default:kmem_cache_free) = 5ddb1c65 needed by nvidia- gfxG05-default-418.56_k4.12.14_lp151.22-lp151.9.1.x86_64" (Same issue with G04 and G03)
Is there a way besides nouveau driver to get Nvidia graphics to work?
Regards, vinz.
Most likely the nvidia required kernel version does not match the one installed. Leap 15.1 is still moving heavily, the new nvidia packages may not yet have been built against the newer kernel.
The packages only contain the source code though, the kernel module is compiled on your system when you install the rpm package. So maybe it would work to just ignore that missing dependency. Kind Regards, Wolfgang -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org
Knurpht-openSUSE schrieb:
Op dinsdag 2 april 2019 17:19:02 CEST schreef Vinzenz Vietzke:
Am Dienstag, 2. April 2019, 10:58:16 CEST schrieb Ludwig Nussel:
So I wonder if beta testers of 15.1 noticed any difference? Was there a change in power consumption for example?
Don't wait upgrading to 15.1 until the release, there are only a few weeks left to get fixes in!
I just tried the latest build 441.4 but struggled at installing proprietary NVIDIA drivers:
"Nothing provides ksym(default:kmem_cache_free) = 5ddb1c65 needed by nvidia- gfxG05-default-418.56_k4.12.14_lp151.22-lp151.9.1.x86_64" (Same issue with G04 and G03)
Is there a way besides nouveau driver to get Nvidia graphics to work?
Regards, vinz. Most likely the nvidia required kernel version does not match the one installed. Leap 15.1 is still moving heavily, the new nvidia packages may not yet have been built against the newer kernel. NVIDIA's .run installer should work though
SLE is at RC level and package freeze is two weeks ahead. Leap is not moving heavily anymore. When the nvidia drivers are available in the repo they are expected to work. So if they don't please file a bug. cu Ludwig -- (o_ Ludwig Nussel //\ V_/_ http://www.suse.com/ SUSE Linux GmbH, GF: Felix Imendörffer, Mary Higgins, Sri Rasiah, HRB 21284 (AG Nürnberg) -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org
On 02/04/2019 17.19, Vinzenz Vietzke wrote:
Am Dienstag, 2. April 2019, 10:58:16 CEST schrieb Ludwig Nussel:
So I wonder if beta testers of 15.1 noticed any difference? Was there a change in power consumption for example?
Don't wait upgrading to 15.1 until the release, there are only a few weeks left to get fixes in!
I just tried the latest build 441.4 but struggled at installing proprietary NVIDIA drivers:
"Nothing provides ksym(default:kmem_cache_free) = 5ddb1c65 needed by nvidia- gfxG05-default-418.56_k4.12.14_lp151.22-lp151.9.1.x86_64" (Same issue with G04 and G03)
Is there a way besides nouveau driver to get Nvidia graphics to work?
I installed the 15.1 kernel with nvidia drivers yesterday, I did not hit into that problem. I have a package mix from 15.0 and 15.1, though: cer@Telcontar:/data/...> rpm -qa | grep "kernel\|nvidia" x11-video-nvidiaG03-340.107-lp151.12.2.x86_64 nvidia-uvm-gfxG03-kmp-default-340.107_k4.12.14_lp151.22-lp151.12.25.x86_64 kernel-firmware-20190118-lp151.1.6.noarch kernel-syms-4.12.14-lp151.23.3.x86_64 nfs-kernel-server-2.1.1-lp151.6.3.x86_64 texlive-l3kernel-2017.133.svn44483-lp150.5.4.noarch kernel-source-4.12.14-lp150.12.48.1.noarch kernel-docs-4.12.14-lp151.23.1.noarch texlive-l3kernel-doc-2017.133.svn44483-lp150.5.4.noarch kernel-devel-4.12.14-lp150.12.48.1.noarch nvidia-glG03-340.107-lp151.12.2.x86_64 nvidia-gfxG03-kmp-default-340.107_k4.12.14_lp151.22-lp151.12.26.x86_64 nvidia-computeG03-340.107-lp151.12.2.x86_64 kernel-default-4.12.14-lp151.22.9.x86_64 kernel-source-4.12.14-lp151.23.3.noarch kernel-default-devel-4.12.14-lp151.23.3.x86_64 kernel-default-4.12.14-lp151.23.3.x86_64 kernel-devel-4.12.14-lp151.23.3.noarch kernel-macros-4.12.14-lp151.23.3.noarch kernel-default-4.12.14-lp150.12.48.1.x86_64 -- Cheers / Saludos, Carlos E. R. (from 15.0 x86_64 at Telcontar)
participants (13)
-
Brüns, Stefan -
Carlos E. R. -
Jan Engelhardt -
John Salvatore Fontanelli -
Knurpht-openSUSE -
Ludwig Nussel -
Marcel Kühlhorn -
Michal Kubecek -
Robert Munteanu -
Stefan Seyfried -
Tobias Klausmann -
Vinzenz Vietzke -
Wolfgang Bauer