| 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 |
|
|
JmBoullier
Former Community Advisor Normandy - France Joined: Jan 26, 2007 Post Count: 3716 Status: Offline Project Badges: 
|
Over a twenty eight day average MERCURY produced 2.6% more points and 8.5% more results than CERES. Since you set your personal objectives in points per day I can understand your disappointment, but if you consider real work done 8.5 % more WUs simply by improving RAM settings or RAM management is not "nothing", particularly when one knows that RAM was already very good in your pool.The impact is much less than expected. By the way this is also confirming that your processors are not often waiting for data and thus HT has not much to recover (ref your previous experiment). Now if you are looking for more ideas of experiments to run I would suggest installing Ubuntu 64-bit (in dual boot) in one of your planets. I am curious to see how many more HCC WUs it would return per day.  If you are hesitating, look back at my numbers in my previous post. I cannot promise a tremendous increase in points because credit granting under Linux is not always marvellous, but with the increased throughput at least you should not lose points. |
||
|
|
darth_vader
Veteran Cruncher A galaxy far, far away... Joined: Jul 13, 2005 Post Count: 514 Status: Offline Project Badges: 
|
I decided to check the influence of DDR3 RAM speed and characteristics to see if there is a difference. <snip> There are plenty of benchmark results available that show more or less the same thing you saw. Once you've got relatively fast memory, even faster memory only helps a little. You would need to have multiple applications that have large memory working set sizes and very random access patterns for both code and data (i.e., you basically have to defeat the processor caches) before you would see an improvement that is commensurate with the faster memory speed. Even then, there is some question as to whether you might hit other bottlenecks, such as contention for the memory bus or DRAM refresh delays.Over a twenty eight day average MERCURY produced 2.6% more points and 8.5% more results than CERES. The impact is much less than expected. - D |
||
|
|
Hypernova
Master Cruncher Audaces Fortuna Juvat ! Vaud - Switzerland Joined: Dec 16, 2008 Post Count: 1908 Status: Offline Project Badges:
|
but if you consider real work done 8.5 % more WUs simply by improving RAM settings or RAM management is not "nothing", You are perfectly right. Science is measured in WU's returned and that is what counts, this is also why I did not bother comparing runtimes. My "disappointment" was 8.5% versus 16%. I would have expected a more proportional improvement around 12% to be true. This is why I will test even faster memory with agressive timings in the coming days. I will also monitor power consumption to check that an improvement in speed does not relate in a ruinous running cost increase. Results not expected before a run of at least two weeks to be meaningful. After that we should know were is the acceptable limit.  |
||
|
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
Glad to have you on the team for this upcoming event Hypernova...Welcome!
|
||
|
|
Hypernova
Master Cruncher Audaces Fortuna Juvat ! Vaud - Switzerland Joined: Dec 16, 2008 Post Count: 1908 Status: Offline Project Badges:
|
Yesterday was my best score ever for Returned Results with 1072.
---------------------------------------- I passed my previous record of 1020 on the 22nd of March. I have still two personal records to beat: Points: 518'283 on the 15th of March Runtime: 150 days on the 15th of March Yesterday Points were at 463'326 (second best 54'957 points short). Runtime 114 days (second best 36 days short). There is a difference however. When the absolute points record was made my daily average was around 230'000 points. Now through various system upgrades I have a daily average that is going over 400'000 points. This means that if I try to replicate the events that lead to the absolute record day I may be able to reach a peak on a given day of 1 Million points. The strategy is simple. I switch synchronously all my solar system devices to crunch HPFP2. This will start to generate a backlog of thousands of WU's due to the high quorum. When the backlog peak is reached (I estimate it could reach 10'000 WU) I then switch all devices to let's say HFCC (quorum 1). So you have now a daily points total which in a very short time reaches the daily average and ofter some days the backlog starts to be released and adds up. This could generate a rogue wave of 1Mpoint height. (see my other post https://secure.worldcommunitygrid.org/forums/...ead,28596_offset,0#274022 ) But I will not do that without an Ok from CA's. |
||
|
|
Sekerob
Ace Cruncher Joined: Jul 24, 2005 Post Count: 20043 Status: Offline |
You may want to acquaint yourself with the DPC patented MEGAFLUSH technique. You could get a day with your farm of 4-4.5 million and still be within deadline.
----------------------------------------Wait at least till v 6.17 is launched for HPF2. You don't want a cache filled and then having them go at 60 seconds each. HPF2 is good as the run times don't vary too much and nobody would be waiting on you for validation. They'd just almost all be probably #19 to return result. edit: Time it so the flush happens to coincide with May 7 or 8, whatever the last date of the XS "Experiment" ... that'd be fun :O)
WCG
----------------------------------------Please help to make the Forums an enjoyable experience for All! [Edit 1 times, last edit by Sekerob at Apr 22, 2010 9:04:07 AM] |
||
|
|
Hypernova
Master Cruncher Audaces Fortuna Juvat ! Vaud - Switzerland Joined: Dec 16, 2008 Post Count: 1908 Status: Offline Project Badges:
|
Wait at least till v 6.17 is launched for HPF2 Sek when is this upgrade planned? |
||
|
|
Jack007
Master Cruncher CANADA Joined: Feb 25, 2005 Post Count: 1604 Status: Offline Project Badges: 
|
Ok, So I can't remember where I mentioned this, but I had found some of the moons of Saturn in Wiki. There was a nice collection of them in one combined jpg. I took that 4 meg file downloaded it to a thumb drive and took it to a printer and had them print up a 12" x 36" Poster (for lack of a better term) on a foam backing. It will now reside above my computer desks and i can look up and dream about Dione (the most beautiful one in my opinion). I miss my astronomy, since I live more in the country now, perhaps I'll break out the spotter scope and start checking out the night sky as it gets warmer. How does this fit in with this thread? Well Hypernova has the planets locked down, so I'ma gonna go after the moons of Saturn (there are like 70 of them so plenty to go around) I'm taking Dione (densest moon covered with ice) Rhea largest airless moon, and Hyperion because it's shaped like a potato. Next comp, perhaps a I7980x, or I've seen a quad socket motherboard (for $679 just for the moboard), will be TITAN. Well some friends of ours need a computer (theirs is like 7+ years old) So I'm going to gut theirs and Rhea (Q6600) and basically 'move' Rhea into a new orbit (their home). Which I then decided to upgrade my now almost empty box to an I7 930 (will overclock to around 3.3/3.5 ghz depending on heat), with 6 Gigs of RAM @ 1600 mhz. So now my wife's non gaming computer will be faster than my gaming computer (I7 920 OC 3.3ghz @ 1333mhz RAM. So Helene is now added to my system, a smaller moon sharing the same orbit as Dione just out in front a bit. My system now contains 5 moons. Still waiting on a big enuf system to qualify as titan.  |
||
|
|
Hypernova
Master Cruncher Audaces Fortuna Juvat ! Vaud - Switzerland Joined: Dec 16, 2008 Post Count: 1908 Status: Offline Project Badges:
|
Jack007 excellent to see that the Solar System is developing at WCG.
----------------------------------------Could you tell us what are the names of the five moons that are in orbit and active radiating HEGP, and also how much HEGPPD they emit. I am wondering if I should launch a Solar System Team. What do you think  Reminder HEGP: High Energy Graviton Particles. HEGPPD: High Energy Graviton Particles Per Day |
||
|
|
Jack007
Master Cruncher CANADA Joined: Feb 25, 2005 Post Count: 1604 Status: Offline Project Badges:
|
I think a Solar System team would be excellent!
----------------------------------------I've always been an astronomy buff (or would be if I had roots) As to my Moons... First Dione Uhm the I7 920 (3.3ghz & 12 gigs of 1333 mhz RAM) around 22,000 HEGPPD but it varies wildly due to being my gaming RIG. Next Rhea The Q6600 @ 2.4 with 4 gigs at 800(I think?) around 10K but that will go to my friends and not sure how/if they will run BOINC. It might be in a bedroom noise etc. I will try and instill the urgency of our work. Next is Helene (my wife picked the moon since it will be hers, parts should arrive on the 26th of May) It's the new I7 930 not sure what clock speed it will be at 2.8 min 6 gigs of 1600 Mhz RAM (I was inspired by your RAM test to not stay with 1333) so another 20Kish. next Hyperion This is my oldest running comp at 1.8 ghz (a laptop) single core with ram upgraded twice all the way to a whopping 2 GIGS maybe 1000 PPD Finally Calypso, A refurbished emachine bought for my parents and rejected running @ 1.6 ghz w/2 gigs or RAM, best guess 800-1000 ppd however it's so small(way low energy) and runs so cool (21 C) that I dont have the heart to box it or sell it for 50 to 100 bucks. (I tried to give it to my friends for use, and they rejected it lol) Lastly I'm planning Titan But that will have to wait. (dreaming of 4 socket moboard w/8 core each socket) I'm hoping to break 50K a Day. Edit: So no team eh... [Edit 1 times, last edit by Jack007 at May 27, 2010 3:41:02 AM] |
||
|
|
|