Coder Social home page Coder Social logo

benchmarks's Introduction

Linux CPU Schedulers Benchmarks and Feedback

ECHO CPU Scheduler

By Array (@A_Array)

stress-ng

By Peter (@ptr1337)

BORE

BORE

ECHO

ECHO

By Skyblueborb (@Skyblueborb)

its pretty amazing not gonna lie, earlier my browser would start to lag when my pc was stuck at 100% compiling. now its pretty responsive, great job

default

default

default

echo

By @Ferdi Scholten (@Ferdi_Scholten)

Finally got to test your new echo scheduler. Had to work around a kernel bug with 6.8.2 first (intel bluetooth on my system). On my laptop (Lenovo P52s) the raw performance of echo is similar or slightly better than eevdf. But when multitasking echo seems to do a better job of keeping things going smoothly.

For the actual benchmarks I did they were video transcoding and code compilation. These are the results:

EEVDF

[out#0/webm @ 0x5629cf176140] video:12301KiB audio:66KiB subtitle:0KiB other streams:0KiB global headers:0KiB muxing overhead: 0.030094% frame= 180 fps=1.1 q=21.0 Lsize= 12371KiB time=00:00:05.96 bitrate=16988.8kbits/s speed=0.0365x test5eevdf.webm created in 164 seconds

[out#0/webm @ 0x557bae01fae0] video:13290KiB audio:539KiB subtitle:0KiB other streams:0KiB global headers:0KiB muxing overhead: 0.185755% frame= 1524 fps=2.6 q=21.0 Lsize= 13855KiB time=00:00:51.98 bitrate=2183.5kbits/s dup=0 drop=39 speed=0.0884x test2eevdf.webm created in 589 seconds

AOM compilation real 1m46,140s, user 3m8,868s, sys 0m15,622s

real 1m46,603s, user 3m9,063s, sys 0m15,676s

FFmpeg compilation real 3m3,995s, user 7m40,606s, sys 0m38,110s

ECHO

[out#0/webm @ 0x562ee23d5140] video:12301KiB audio:66KiB subtitle:0KiB other streams:0KiB global headers:0KiB muxing overhead: 0.030094% frame= 180 fps=1.3 q=21.0 Lsize= 12371KiB time=00:00:05.96 bitrate=16988.8kbits/s speed=0.0446x test5echo.webm created in 134 seconds

[out#0/webm @ 0x563775225ae0] video:13290KiB audio:539KiB subtitle:0KiB other streams:0KiB global headers:0KiB muxing overhead: 0.185755% frame= 1524 fps=2.6 q=21.0 Lsize= 13855KiB time=00:00:51.98 bitrate=2183.5kbits/s dup=0 drop=39 speed=0.0882x test2echo.webm created in 589 seconds

AOM compilation real 1m46,416s, user 3m7,630s, sys 0m15,674s

real 1m47,497s, user 3m7,294s, sys 0m15,846s

FFmpeg compilation real 3m2,495s, user 7m36,955s, sys 0m37,232s

By the way while posting this and making the screenshot, a large transcode is happily progressing which would have caused serious lag when using eevdf but telegram, text editing and browsing, partial screenshotting etc. all function without noticing the cpu and gpu full utilization in the background.

htop

This is happening while posting and without me noticing any delay in typing or anything else....

By Mr. Error (@RiverOnVenus)

My opinion on Linux CPU scheduler

jitterdebugger to measure wake up latency

arch vanilla kernel 6.8.1

arch vanilla kernel 6.8.1

arch vanilla kernel 6.8.1

arch vanilla kernel 6.8.1

arch vanilla kernel 6.8.1 patched ECHO

arch vanilla kernel 6.8.1 patched ECHO

arch vanilla kernel 6.8.1 patched ECHO

arch vanilla kernel 6.8.1 patched ECHO

By Notyou-Drbiguint (@Drbiguintyohaparfum)

default vs bore vs echo

By Hamad Al Marri

WebGL - default

WebGL - echo

Note: The default linux scheduler shows more FPS and while running the test the motion was smooth, however the OBS recording frame rate has went down in the favour of the web page render. In the other hand, ECHO was able to fairly distribute the CPU time between all tasks and the motion was smooth while testing as well as the OBS recording in which the webGL frame rate was lower than default scheduler but that is normal because all running tasks had enough CPU time to work togather. In the default scheduler since OBS didn't have enough CPU time, its frame rate went down which even gives more CPU time to the WebGL. The final findings is that ECHO is more fair and smoother.

phoronix-test-suite

Shared by @marioroy

A Clear Linux user tried my ClearMod repository and compared the Vanilla native kernel (no preemption) and ECHO (XanMod + preemption + ECHO).

I completed testing a demo for the phmap author. Yet, another surprise. :-)

benchmarks's People

Contributors

hamadmarri avatar

Stargazers

 avatar  avatar

Watchers

 avatar  avatar

benchmarks's Issues

Request: make gaming performance comperisions to different schedulers under high and low background system load.

I'd be interested to see how this scheduler does when only running a desktop environment and a CPU intensive game that's Single core limited and a game that's good at using multiple cores.
And i'd be also interested in seeing how the scheduler handles running the same two game tests under heavy background load. For example firefox running in the background with a heavy website like youtube, playing a video, or other taxing processes in the background.

Request: Make the benchmark comparision results more human readable/digestable.

The Benchmark compressions in the README.md are very barren and require prior understanding of the benchmark tools to understand what results mean.

  • include simple short explanations what which test tests.
  • Tests that give a score, mention if lower is better or higher is better.
  • Later on potentially write a proper white paper on the scheduler.

Recommend Projects

  • React photo React

    A declarative, efficient, and flexible JavaScript library for building user interfaces.

  • Vue.js photo Vue.js

    ๐Ÿ–– Vue.js is a progressive, incrementally-adoptable JavaScript framework for building UI on the web.

  • Typescript photo Typescript

    TypeScript is a superset of JavaScript that compiles to clean JavaScript output.

  • TensorFlow photo TensorFlow

    An Open Source Machine Learning Framework for Everyone

  • Django photo Django

    The Web framework for perfectionists with deadlines.

  • D3 photo D3

    Bring data to life with SVG, Canvas and HTML. ๐Ÿ“Š๐Ÿ“ˆ๐ŸŽ‰

Recommend Topics

  • javascript

    JavaScript (JS) is a lightweight interpreted programming language with first-class functions.

  • web

    Some thing interesting about web. New door for the world.

  • server

    A server is a program made to process requests and deliver data to clients.

  • Machine learning

    Machine learning is a way of modeling and interpreting data that allows a piece of software to respond intelligently.

  • Game

    Some thing interesting about game, make everyone happy.

Recommend Org

  • Facebook photo Facebook

    We are working to build community through open source technology. NB: members must have two-factor auth.

  • Microsoft photo Microsoft

    Open source projects and samples from Microsoft.

  • Google photo Google

    Google โค๏ธ Open Source for everyone.

  • D3 photo D3

    Data-Driven Documents codes.