Linux 4.6-rc1 Kernel Benchmarks On A Xeon E5 Haswell

[phoronix]

Phoronix: Linux 4.6-rc1 Kernel Benchmarks On A Xeon E5 Haswell

Here are some of my earliest Linux 4.6 kernel benchmarks out there from a few different subsystems with the tests happening from a Xeon E5 Haswell system with AMD Radeon graphics...

http://www.phoronix.com/scan.php?page=news_item&px=Linux-4.6-rc1-Xeon-E5-v3

[Dick Palmer]

Well that all looked startlingly serendipitous for a moment! Especially with that sneaky CPU test hiding so effortlessly among all the block device triumphs... but...

"Here are some of my earliest..." bothered me
"Here are the best of my..."?

Was it a genuinely arbitrary set of tests, or was it cherry-picked from a larger set?.. so off I trotted to attempt to evaluate another of those fiendishly obfuscated number blobs.

Yup. Cherry-picked. It seems.

Any chance of actually labelling the rows in those horrid PTS result tables with a little clue to the identity of the test to which their values are meant to pertain?

Just mouseover labels would be better than NOTHING if you're trying to save space.

Please?

[Michael]

Well that all looked startlingly serendipitous for a moment! Especially with that sneaky CPU test hiding so effortlessly among all the block device triumphs... but...

"Here are some of my earliest..." bothered me
"Here are the best of my..."?

Was it a genuinely arbitrary set of tests, or was it cherry-picked from a larger set?.. so off I trotted to attempt to evaluate another of those fiendishly obfuscated number blobs.

Yup. Cherry-picked. It seems.

Any chance of actually labelling the rows in those horrid PTS result tables with a little clue to the identity of the test to which their values are meant to pertain?

Just mouseover labels would be better than NOTHING if you're trying to save space.

Please?

Obviously the results I show on the main page are of the more interesting data and then from the OpenBenchmarking.org results linked to from the article are all of the results run. I wasn't seggregating by best or worst results to show on the main page, just the results with most change to look at.

There should be mouseover support on that OB overview table with more details. Otherwise likely some browser compatibility issue and in that case patches are welcome.