PIV Hyperthreading question?

Message boards : Number crunching : PIV Hyperthreading question?
Message board moderation

To post messages, you must log in.

AuthorMessage
Profile AndyW Project Donor
Volunteer tester
Avatar

Send message
Joined: 23 Oct 02
Posts: 5862
Credit: 10,957,677
RAC: 18
United Kingdom
Message 718730 - Posted: 26 Feb 2008, 15:02:25 UTC

If I run a machine that has hyperthreading capability with the hyperthreading edisabled will it crunch a WU twice as fast as it would if hyperthreading was enabled and it was crunching 2 units?

I've tried doing comparisons, but without a set of identically sized WUs to play with it's hard to come up with the answer.
ID: 718730 · Report as offensive
Alinator
Volunteer tester

Send message
Joined: 19 Apr 05
Posts: 4178
Credit: 4,647,982
RAC: 0
United States
Message 718731 - Posted: 26 Feb 2008, 15:07:24 UTC
Last modified: 26 Feb 2008, 15:12:05 UTC

The host will crunch faster with HT disabled, but not twice as fast. IIRC, you gain about 40% or so more total throughput with HT enabled, but I forget what the folks were posting the observed increase in individual task runtime was.

HTH,

Alinator
ID: 718731 · Report as offensive
Profile AndyW Project Donor
Volunteer tester
Avatar

Send message
Joined: 23 Oct 02
Posts: 5862
Credit: 10,957,677
RAC: 18
United Kingdom
Message 718733 - Posted: 26 Feb 2008, 15:10:18 UTC

OK, so there is a gain using hyperthreading for our purposes. Thank you. I can see where it wouldn't always be useful, but for number crunching it seems to fit well. Thanks for your answer.
ID: 718733 · Report as offensive
tfp
Volunteer tester

Send message
Joined: 20 Feb 01
Posts: 104
Credit: 3,137,259
RAC: 0
United States
Message 718734 - Posted: 26 Feb 2008, 15:11:29 UTC

It will not be twice as fast but normally it will add 20-40% vs running just one.
ID: 718734 · Report as offensive
Alinator
Volunteer tester

Send message
Joined: 19 Apr 05
Posts: 4178
Credit: 4,647,982
RAC: 0
United States
Message 718737 - Posted: 26 Feb 2008, 15:13:55 UTC
Last modified: 26 Feb 2008, 15:14:16 UTC

It will be interesting to see how well Intel addressed the shortcomings with HT when they rollout the 2nd generation of it in Nehalem.

Alinator
ID: 718737 · Report as offensive
Profile Keck_Komputers
Volunteer tester
Avatar

Send message
Joined: 4 Jul 99
Posts: 1575
Credit: 4,152,111
RAC: 1
United States
Message 718985 - Posted: 27 Feb 2008, 6:10:08 UTC - in response to Message 718737.  

It will be interesting to see how well Intel addressed the shortcomings with HT when they rollout the 2nd generation of it in Nehalem.

Alinator

I wonder if they even recognise any shortcommings with HT. It was designed to find hidden performance in the CPUs and it did find about 15% to 30%.
BOINC WIKI

BOINCing since 2002/12/8
ID: 718985 · Report as offensive
Profile Misfit
Volunteer tester
Avatar

Send message
Joined: 21 Jun 01
Posts: 21804
Credit: 2,815,091
RAC: 0
United States
Message 718988 - Posted: 27 Feb 2008, 6:32:05 UTC
Last modified: 27 Feb 2008, 6:33:11 UTC

The answers here are correct. Various projects will give different results. For best results always leave H/T on and crunch two different projects at the same time. Having two instances of the same project running will have them competing for the same resources at the same time.
It will be interesting to see how well Intel addressed the shortcomings with HT when they rollout the 2nd generation of it in Nehalem.

I was thinking about that myself. 16 cores anyone?
me@rescam.org
ID: 718988 · Report as offensive

Message boards : Number crunching : PIV Hyperthreading question?


 
©2024 University of California
 
SETI@home and Astropulse are funded by grants from the National Science Foundation, NASA, and donations from SETI@home volunteers. AstroPulse is funded in part by the NSF through grant AST-0307956.