| Index | Recent Threads | Unanswered Threads | Who's Active | Guidelines | Search |
 |
World Community Grid Forums
Category: Community
Forum: Chat Room
Thread: Hardware - Technology - news |
| No member browsing this thread |
|
Thread Status: Active Total posts in this thread: 269
|
|
| Author |
|
|
Jack007
Master Cruncher CANADA Joined: Feb 25, 2005 Post Count: 1604 Status: Offline Project Badges: 
|
"As an example my Boinc benchmark gives a total rounded of 144 GFLOP/sec on my i7 980X at 4Ghz"
----------------------------------------Seriously? My I7 920 @ 3.33 is 3GFLOPS/core so 24GFLOPS total (24.328) If I had 6 cores that would be 36.492 plus the ghz dif is 43.790 Your getting 3.29 TIMES the output of my I7? (adjusted even!) (to clarify this is the BOINC manager/advanced tab/run cpu benchmark/WHETSTONE) I know the 980x is a great chip and all but... a little extra cache a little faster RAM (mine is 1333) what can account for such a large difference...? Edit: Working backwards from my Score (around 22000 WCG PPD) thats 31 GFLOPS. Boinc benchmarks put me at 24 GFLOPS which would be 17000 WCG PPD, not too far off. This is using the 700 WCG PPD = GFLOP/S formula. So I do believe that the 400 Teraflop(s) WCG is doing is per second not day, but thats my 2 cents One more thing, using your formula (where you got 1 teraflop=4.4 WCG points) I used same formula on my single I7 and get 1 TFLOP = 1.2 wcg I wonder if your using whetstone... thats the Floating point operation per second as opposed to the drystone/mips which is about 3.27 times more points suspiciously like the 3.29 times faster your 980x is over my 920 when adjusted for cores and ghz... I'm not guaranteeing this just musing.  [Edit 1 times, last edit by Jack007 at Jun 1, 2010 3:15:23 PM] |
||
|
|
Hypernova
Master Cruncher Audaces Fortuna Juvat ! Vaud - Switzerland Joined: Dec 16, 2008 Post Count: 1908 Status: Offline Project Badges: 
|
(to clarify this is the BOINC manager/advanced tab/run cpu benchmark) Yes that is correct Jack. But I am unable now to retest. As usual I am travelling abroad. I will be back on friday and retest/recheck the values.  I wonder if your using whetstone... thats the Floating point operation per second as opposed to the drystone/mips which is about 3.27 times more points I used the highest value of both. I do not remember now what it was but I think it was whetstone. It is the last value coming out of the benchmark.  [Edit 1 times, last edit by Hypernova at Jun 1, 2010 3:19:45 PM] |
||
|
|
Jack007
Master Cruncher CANADA Joined: Feb 25, 2005 Post Count: 1604 Status: Offline Project Badges:
|
the last value and highest is drystone...
----------------------------------------ok, that makes sense for your numbers, The first value is whetstone floating point operations per second the second drystone is instructions per second. Enjoy your trip! |
||
|
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
Yes enjoy your trip !
Any other Hardware - Technology - news I wonder  |
||
|
|
mikaok
Senior Cruncher Finland Joined: Aug 8, 2006 Post Count: 489 Status: Offline Project Badges: 
|
China supercomputer named world's second-fastest
----------------------------------------"The Nebulae is capable of sustained computing of 1.271 petaflops — or 1,271 trillion calculations — per second, according to TOP500. It said the Jaguar was capable of sustained computing of 1.75 petafl.." wait! Did I say that out loud  edit. quote marks
to infinity and beyond
---------------------------------------- [Edit 1 times, last edit by mikaok at Jun 1, 2010 5:56:34 PM] |
||
|
|
Hypernova
Master Cruncher Audaces Fortuna Juvat ! Vaud - Switzerland Joined: Dec 16, 2008 Post Count: 1908 Status: Offline Project Badges:
|
ok, that makes sense for your numbers, Thanks Jack. But it means that I calculated a ratio for GIPS (Giga Instructions Per Second) and not GFLOPS. If I use now the Whetstone value that I rememeber, that is 45 GFLOPS per device. I will retest this anyway but as order of magnitude it should be ok. Then we have the new ratio: 1 TFLOP = 14.2 WCG points This means that on this base my total contribution up to today that is 52.6 Million points boils down to 3'704'225 TFLOP or 3'704 PETAFLOP. If we look at my daily contribution we are 42.2 PETAFLOP. If I take now the ratio above and I apply it to WCG last daily point total (1st June) I have: WCG total yesterday 278'963'231 points. This is equivalent to a 227.3 TERAFLOP/SECOND It is true that 980X processors are an infinitesimal quantity in WCG and cannot be used as reference. But still the ratio itself may not be so different. Even if we double it or triple it we are still orders of magnitude far from the initial one of 1 TFLOP = 700'000 WCG points. This is dead wrong. In view of all said above, Sekerob calculated WCG crunching power, is indeed in TERAFLOP/SECOND or TERAFLOPS. This can also bee seen in the title legend of his daily table. Now that we have clarified everything let's calculate the true WCG ratio TFLOP/WCG Points. This is for yesterday 1st June 395.8 TERAFLOPS or for the day 34'197'120 TERFLOP total Floating Points Operations. WCG in the same period has produced 278'963'231 WCG Points. The WCG ratio is in simple and rounded form: 1 TFLOP = 8 WCG Points We should use that ratio to have a comparable base of reference to benchmark the daily output of all machines in WCG. We see that we are, give and take at: 1 TFLOP = 1 Boinc Credit. On that base, my 600'000 points daily contribution, will relate to 75 PETAFLOP worth of calculations. Now I feel better, as all the numbers have now logical relations that fit all in. So my target is not anymore 1 TFLOP/day but 100 PFLOP/day. Much more rewarding. Sek the title of your thread should have be "Building the 400 TERAFLOPS Machine...". Otherwise the legend of your daily updated table is correct as it is in TERAFLOPS. I know it is just a small letter difference but It has a world difference in meaning. |
||
|
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
Hello Hypernova
----------------------------------------Could kindly please keep your posts with not so many wide spaces between lines , and not so big writing .. and as Sek suggested take it to a different/own thread read Many thanks !, no need to quote me and have a good day ! [Edit 5 times, last edit by Former Member at Jun 2, 2010 2:59:28 PM] |
||
|
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
VIA reveals 1.6GHz Nano DC processor at Computex, shows it handling 720p (video)
----------------------------------------[Edit 1 times, last edit by Former Member at Jun 2, 2010 3:35:29 PM] |
||
|
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
Jean this may be up your alley. :D A water cleaning chip!
http://www.trendhunter.com/trends/desalination-chip |
||
|
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
---------------------------------------- [Edit 3 times, last edit by Former Member at Jun 3, 2010 1:52:36 PM] |
||
|
|
|