Phoronix Test Suite 3.0 "Iveland" Alpha 1

[phoronix]

Phoronix: Phoronix Test Suite 3.0 "Iveland" Alpha 1

For those that don't subscribe to the Phoronix Test Suite mailing list, there is a new announcement: Phoronix Test Suite 3.0 "Iveland" Alpha 1 has been released...

http://www.phoronix.com/vr.php?view=ODc2OQ

[jakubo]

i know its a bit off topic... but:
what happened to the phoromatic performance tracker? its stuck at 28th dec. Nothing up to date (i.e. no 2.6.37 kernel)

[Michael]

i know its a bit off topic... but:
what happened to the phoromatic performance tracker? its stuck at 28th dec. Nothing up to date (i.e. no 2.6.37 kernel)

Kernel hung the systems and finally got back to my office this morning so I can manually restart them.

[jakubo]

Kernel hung the systems and finally got back to my office this morning so I can manually restart them.

sounds promising...

[Stebs]

BTW, for inspiration, here are some dynamic (SVG+JS) benchmark graphs that a german site uses since 3 years or so (+ png Fallback):

http://www.computerbase.de/artikel/g...frameverlaeufe

Click on the lines (or the names of the GPUs) to see the added value of dynamic graphs.

Their horizontal bar graphs are also simple but well done, hover over GPU Names to set 100% at the current selection:

http://www.computerbase.de/artikel/g...ating_mit_aaaf

[Michael]

BTW, for inspiration, here are some dynamic (SVG+JS) benchmark graphs that a german site uses since 3 years or so (+ png Fallback):

http://www.computerbase.de/artikel/g...frameverlaeufe

Click on the lines (or the names of the GPUs) to see the added value of dynamic graphs.

Their horizontal bar graphs are also simple but well done, hover over GPU Names to set 100% at the current selection:

http://www.computerbase.de/artikel/g...ating_mit_aaaf

Americans may not be able to make better beer than the Germans, but when OpenBenchmarking.org is done I am confident the graphs will be better than theirs

[Stebs]

when OpenBenchmarking.org is done I am confident the graphs will be better than theirs

Great! I was hoping so... but I had to post the example, preventing blaming myself later - if you would miss the opportunity to do nice dynamic graphs

[Stebs]

While I'm at it: Pleease add cairo-perf-trace benchs to Iveland!

I know you said since over a year that you were waiting for some recorded traces from Intel? (WTF?)
But you DO know that there is even a cairo-traces repository (besides that you can easily record a trace yourself), do you?
Relevant information: http://cairographics.org/FAQ/#profiling

Sure, PTS is THE benchmarking tool for Linux, but i am sure that you heard the complain of questionable benchmark relevance often enough.

Benchs with cairo traces from common apps like Firefox, Evolution, poppler, swfdec etc. would be WAY more useful than those syntetic tests like GTKperf. It's all there (and to quote the Duke)
What are you waiting for, Christmas?

[Michael]

While I'm at it: Pleease add cairo-perf-trace benchs to Iveland!

I know you said since over a year that you were waiting for some recorded traces from Intel? (WTF?)
But you DO know that there is even a cairo-traces repository (besides that you can easily record a trace yourself), do you?
Relevant information: http://cairographics.org/FAQ/#profiling

Sure, PTS is THE benchmarking tool for Linux, but i am sure that you heard the complain of questionable benchmark relevance often enough.

Benchs with cairo traces from common apps like Firefox, Evolution, poppler, swfdec etc. would be WAY more useful than those syntetic tests like GTKperf. It's all there (and to quote the Duke)
What are you waiting for, Christmas?

It's not traces I am waiting on but Intel has a PTS test profile for cairo-perf-trace and have been waiting for them to send it over so I can save a few minutes rather than writing one myself...

[Stebs]

It's not traces I am waiting on but Intel has a PTS test profile for cairo-perf-trace and have been waiting for them to send it over so I can save a few minutes rather than writing one myself...

Ok, but then, maybe after this long time of waiting (and perhaps a last wake-up mail), those "few minutes" of work (even when reinventing the wheel) could be well worth it?