| Index | Recent Threads | Unanswered Threads | Who's Active | Guidelines | Search |
 |
World Community Grid Forums
Category: Community
Forum: Chat Room
Thread: [FINAL RESULTS HT Test] The Solar System Initiative - SSI |
| No member browsing this thread |
|
Thread Status: Active Total posts in this thread: 81
|
|
| Author |
|
|
Jack007
Master Cruncher CANADA Joined: Feb 25, 2005 Post Count: 1604 Status: Offline Project Badges: 
|
Hypernova, I am Eagerly awaiting the results of your test... I have but one machine with HT, for badge purposes it's great but are we losing some actual processing power due to it having to switch around...? interesting. I love badges but.... it's about the science when it comes down to it so I wanna do what's best there. (not saying you don't!)
----------------------------------------By the way, in honor of SSI... My laptop gets very hot, but i have cooling fan it sits on so ... MERCURY My wife uses my Q6600 so VENUS obviously I was gonna name my monster cruncher after our planet but... I realized it's an Alienware lol, so the I7 at 3.33 ghz is my fav Alien planet... MARS Cheers  |
||
|
|
Hypernova
Master Cruncher Audaces Fortuna Juvat ! Vaud - Switzerland Joined: Dec 16, 2008 Post Count: 1908 Status: Offline Project Badges: 
|
Excellent idea Jack007. The more planets we have around the better it is for WCG.
----------------------------------------  Crunch, Crunch, Crunch, Crunch   |
||
|
|
JmBoullier
Former Community Advisor Normandy - France Joined: Jan 26, 2007 Post Count: 3716 Status: Offline Project Badges:
|
but are we losing some actual processing power due to it having to switch around...? Jack,I was thinking like you at the beginning, seeing HT only as a mean to switch tasks faster when there was a real need to run several tasks at once, but not giving any more global power. Until I decided to measure it seriously in my old P4 HT. And to my surprise this P4 HT was able to crunch about 30 % more HCC WUs per day. It looks like HT is the only mean to pull out all the power that these processors are able to offer. And I think that Movieman has measured a higher improvement ratio when he was testing his first i7. Anyway, with Hypernova's almost ideal test we will soon know for sure. |
||
|
|
Movieman
Veteran Cruncher Joined: Sep 9, 2006 Post Count: 1042 Status: Offline |
HT definately adds to the work done and not in a small way.
----------------------------------------OT but Hypernova if your ever planning a trip to the US get in touch with me. Someone I'd like to meet,shake your hand and buy you dinner. Not for what you do but the mindset behind it. You fit in well with the people I call the "Good Hearts" of this world.  |
||
|
|
Jack007
Master Cruncher CANADA Joined: Feb 25, 2005 Post Count: 1604 Status: Offline Project Badges:
|
And to my surprise this P4 HT was able to crunch about 30 % more HCC WUs per day. It looks like HT is the only mean to pull out all the power that these processors are able to offer. And I think that Movieman has measured a higher improvement ratio when he was testing his first i7. Anyway, with Hypernova's almost ideal test we will soon know for sure. I read somewhere that the I7 has turbo built in, it won't kick in unless needed and maybe it's not needed unless HT is on... but that wouldnt explain your old P4... so since it is definitely good for badges (twice as much time counted?) and it is improvement for actual wunits crunched I shall leave it on til such time as needed off (A type WU's if I get 3 or 4 of them so they get done in half the time and don't have sad wingers) ok ok ok, it's APRIL FIRST NOW HYPERNOVA I can't stand the wait anymore or did I miss the results of your test ima go read more now and see if i did. Edit: can't find it, the suspense is KILLING me. [Edit 1 times, last edit by Jack007 at Apr 1, 2010 3:50:11 PM] |
||
|
|
JmBoullier
Former Community Advisor Normandy - France Joined: Jan 26, 2007 Post Count: 3716 Status: Offline Project Badges:
|
I can't stand the wait anymore Hypernova said "We will wait until Sunday to get a firs...cally sound result." or did I miss the results of your test ima go read more now and see if i did. Edit: can't find it, the suspense is KILLING me. It's not Sunday yet. |
||
|
|
Jack007
Master Cruncher CANADA Joined: Feb 25, 2005 Post Count: 1604 Status: Offline Project Badges:
|
Sigh, I seem to recall that but for some reason
----------------------------------------I was thinking end of the month not end of the week... TY |
||
|
|
Hypernova
Master Cruncher Audaces Fortuna Juvat ! Vaud - Switzerland Joined: Dec 16, 2008 Post Count: 1908 Status: Offline Project Badges:
|
Happy Easter to all of you and your families. Do not exxagerate with eggs and chocolate today.
----------------------------------------So here are the first results of: The Intel i7 950 Hyper Threading International Ultimate Contest. We try to answer here how worth is the HT feature of the core i7 processors, for WCG crunching. A small reminder of the conditions: - Two identical rigs (ERIS and PLUTO) with core i7 950 CPU's. - CPU frequency is 3.66 Ghz (Core temps 60C) - RAM is DDR3 6Gb at 1600 Mhz - Project crunched is the latest HCC version 6.08. - The machines crunched for 6 days but the stats are based on 5 full days. The first day was not counted to eliminate any possible residual transitory effect between processing modes. Results: ERIS (HT OFF) 106'167 points PLUTO (HT ON) 159'261 points PLUTO crunched exactly 50% MORE points than ERIS. In other words the HT feature is equivalent to have a Hexacore I7 950. Conclusion: The Hyperthreading feature is a MUST for WCG crunchers. This test will continue for another 5 days and I will make an update over 10 days to check consistency and then stop. What about the 980X. Probably even a better result as I suppose the HT feature has been improved. That means an 980X with HT is equivalent to a nine core CPU. When crunching at 4 Ghz (easy) that is a total of 36 GHz. It will be difficult for AMD to beat that. They need at least a 12 core running at 3 Ghz which they do not have yet. |
||
|
|
JmBoullier
Former Community Advisor Normandy - France Joined: Jan 26, 2007 Post Count: 3716 Status: Offline Project Badges:
|
Hypernova,
----------------------------------------The points claimed and granted for hyperthreaded threads tend to be higher than what they should be, at least that is the case in my P4 HT. Also they can be affected by various factors external to the experiment. So, in order to have not arguable measurements could you please also publish the number of HCC WUs which have been returned by each machine during this period, whether validated or pending validation. Since HCC WUs duration is rather stable this is is a better measurement of real work done, not of points earned. Cheers. Jean. Edit: Added "whether validated or pending validation" ---------------------------------------- [Edit 1 times, last edit by JmBoullier at Apr 4, 2010 8:11:24 AM] |
||
|
|
Hypernova
Master Cruncher Audaces Fortuna Juvat ! Vaud - Switzerland Joined: Dec 16, 2008 Post Count: 1908 Status: Offline Project Badges:
|
Here are some numbers corresponding to a full 7 days period. This are taken from the daily device statistics.
----------------------------------------Just to remember Eris is without HT, and Pluto with HT. Points: Eris 171'523 Pluto 247'767 ( +44% ) Runtime in seconds: Eris 2'628'723 Pluto 5'539'205 ( +110% ) Returned Valid Results Eris 637 Pluto 722 ( +13% ) According to what is the most important criterion for you you are very happy or just happy. For Points and Runtime HT remains a must. For Returned Results HT is definitively better but not a killer. Jean you were right on that point. For crunchers HT is a boon, as Badges are based on Runtime and Ranks on Points. For the Scientists valid WU's are the most important criteria, and even for them HT remains a plus. The test will continue with a new and last update for a 10 days period. After that Eris will be back wit HT. This update will be on Sunday 11th April. |
|||
|
||||
|
|