DragonFlyBSD 4.2 vs. Ubuntu 15.10 CPU Performance

[phoronix]

Phoronix: DragonFlyBSD 4.2 vs. Ubuntu 15.10 CPU Performance

In this article are benchmarks comparing the performance of DragonFlyBSD 4.2 to that of Ubuntu 15.10. With these CPU-focused benchmarks, the core scaling performance was also looked at in going from two cores through four cores plus Hyper Threading.

http://www.phoronix.com/vr.php?view=22267

[Dick Palmer]

It's a pity this comparison wasn't done on that 20 core Xeon E5-2687W system which gave those anomalous results a week ago ( http://www.phoronix.com/scan.php?pag...-15-Benchmarks ) - Apart from offering more cores for the comparison, it would have been very interesting to see if those Ubuntu 15.10 results were affected by core-count and see them plotted against a BSD.

Any follow-up to that article Michael? Have you poked around any further?

[Michael]

It's a pity this comparison wasn't done on that 20 core Xeon E5-2687W system which gave those anomalous results a week ago ( http://www.phoronix.com/scan.php?pag...-15-Benchmarks ) - Apart from offering more cores for the comparison, it would have been very interesting to see if those Ubuntu 15.10 results were effected by core-count and see them plotted against a BSD.

Any follow-up to that article Michael? Have you poked around any further?

That Xeon system was busy with other stuff.

No follow-up article yet as not enough interest/support to justify it yet.

[Dick Palmer]

That Xeon system was busy with other stuff.

No follow-up article yet as not enough interest/support to justify it yet.

With fifty-odd comments it looked liked that article had quite a lot of interest... but most of those comments were just people bickering about who's fave distro is the 'ardest.

Do you still have the 15.10 dev snapshot you used in that test? Could the PTS upgrade have an effect? It'd really me (at least) to see that snapshot compared with the current nightly on the new PTS... perhaps as a four-way with Ubuntu 10.04 and a BSD... in a per-core comparison like this one. Hope some more interest bubbles up

[Michael]

With fifty-odd comments it looked liked that article had quite a lot of interest... but most of those comments were just people bickering about who's fave distro is the 'ardest.

Do you still have the 15.10 dev snapshot you used in that test? Could the PTS upgrade have an effect? It'd really me (at least) to see that snapshot compared with the current nightly on the new PTS... perhaps as a four-way with Ubuntu 10.04 and a BSD... in a per-core comparison like this one. Hope some more interest bubbles up

A lot of comments, yes, but mostly people flaming and such with nothing too constructive and no tips.

No, I don't have the 15.10 daily ISO stored. The PTS version really doesn't have any affect, I always use Git master when testing.

[Dick Palmer]

No, I don't have the 15.10 daily ISO stored.

Don't suppose it matters really - if it was just a blip in development and can't be reproduced with something newer then it isn't all that interesting anyway. Oddly, there were some (memory not computation) tests which you didn't highlight which showed almost equally huge improvement. I wonder if the loses and gains are connected..


Thanks for the replies.

[darkcoder]

Numbers look nice. Can you please add FreeBSD comparison later on?

[stiiixy]

I'd definitely like to see a higher-thread count benchy. There seems to be a closing gap there as it goes higher for a few tasks. While some of the smarter beans could probably deduce some reasons behind it (based purely on experience with software on those area's), chuck a few more distro's on top of that higher thread count Michael and see what come's about. A few more tests, besides. Some encryption/decryption really should be a regular in your battery (My opinion, of course).