| Index | Recent Threads | Unanswered Threads | Who's Active | Guidelines | Search |
 |
World Community Grid Forums
Category: Completed Research
Forum: FightAIDS@Home Phase 2
Thread: FightAIDS@Home - Phase 2 AsyncRE work units started |
| No member browsing this thread |
|
Thread Status: Active Total posts in this thread: 47
|
|
| Author |
|
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
I need 2 days to run a work unit, the way I want to run one. At this point I have stopped this project from running. I will change that if the time frame is doubled.
|
||
|
|
Trotador
Senior Cruncher Joined: Mar 26, 2009 Post Count: 154 Status: Offline Project Badges: 
|
If I well understood the WU download limit is 1 wu per thread but I'm getting a maximum of 64 wus in hosts with 72 and 88 threads. Is that way how it is intended to work?
---------------------------------------- |
||
|
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
How many CPUs does BOINC recognize in log. Have you tried setting the NCPUS in cc_config to the number of threads supported by the hardware?
|
||
|
|
Trotador
Senior Cruncher Joined: Mar 26, 2009 Post Count: 154 Status: Offline Project Badges:
|
How many CPUs does BOINC recognize in log. Have you tried setting the NCPUS in cc_config to the number of threads supported by the hardware? It is only with FAH2 that just 64 wus are downloaded and crunched. If for example I also enable SCC I get 100% CPU load or if only SCC is enabled. |
||
|
|
cjslman
Master Cruncher Mexico Joined: Nov 23, 2004 Post Count: 2082 Status: Offline Project Badges: 
|
3. Work units have a short deadline of 24 hours. Quick question: Why do the WUs have a deadline of 24 hrs? I'm not complaining, I'm just curious.Thanks, CJSLMAN Crunching like there's no tomorrow !!! |
||
|
|
NixChix
Veteran Cruncher United States Joined: Apr 29, 2007 Post Count: 1187 Status: Offline Project Badges: 
|
3. Work units have a short deadline of 24 hours. Quick question: Why do the WUs have a deadline of 24 hrs? I'm not complaining, I'm just curious.The short deadline causes me trouble, so I am not presently running FAH2. On some of my crunchers the run time is longer than the deadline.  Cheers   |
||
|
|
cjslman
Master Cruncher Mexico Joined: Nov 23, 2004 Post Count: 2082 Status: Offline Project Badges:
|
The short deadline causes me trouble, so I am not presently running FAH2. On some of my crunchers the run time is longer than the deadline. Yes, it took me a while (goofing with the settings) to get where I could crunch FAAH2. Actually, it's very simple, just only crunch FAAH2, but I can understand were the run time could be longer than the deadline. For me, the short WUs are about 1.5-2 hrs long and the long ones are aprox 8 hrs long.CJSL Crunching for a better world... |
||
|
|
flynryan
Senior Cruncher United States Joined: Aug 15, 2006 Post Count: 235 Status: Offline Project Badges: 
|
Anyone know which O/S and CPU architecture runs FAAH 2 the best/most efficient? Trying to optimize a few crunching machines for it.
|
||
|
|
Jim1348
Veteran Cruncher USA Joined: Jul 13, 2009 Post Count: 1066 Status: Offline Project Badges: 
|
Anyone know which O/S and CPU architecture runs FAAH 2 the best/most efficient? Trying to optimize a few crunching machines for it. I just ran some tests, first on Linux versus Windows on Intel CPUs. In each case, seven cores were running various WCG projects (mostly FAH2), and one core was reserved for supporting a GPU on Folding. In other words, the CPUs were all fully loaded. Also, none of the machines are overclocked, and are running at their default clock speeds of 3.7 GHz. Note that FAH2 falls into two groups: short ones, and long ones, and I averaged them seperately. All times are in minutes. i7-4771 (Windows 7 64-bit): Shorts: (three samples) = 72.3 minutes Longs: (four samples) = 351.5 minutes i7-3770 (Ubuntu 16.04): Shorts: (eight samples) = 76 minutes Longs: (four samples) = 381.5 minutes So the Windows machine is 1.08 (8%) faster on the Longs, and 1.05% (5%) faster on the shorts, but with rather small sample sizes. Since the usual processing speed of the i7-4771 is not much faster than the i7-3770, less than about 10% in most cases due to the architectural differences, the results indicate that Windows and Linux are very closely matched, which is rather unusual. And I also ran a test on my Ryzen 1700 (Ubuntu 17.10) at its default clock speed of 3.0 GHz (as shown by "lscpu" - I don't know if that is correct). In this case, 15 cores were run on FAH2 while the other core supported a GPU on Folding, so this machine was also fully loaded. Ryzen 1700 (Ubuntu 17.10): (fifteen samples, but only the long ones were received) = 349.7 minutes So the Ryzen 1700 is 381.5/349.7 = 1.09 (9%) faster than the i7-3770 per core, though with a rather small sample size for the i7-3770 it could vary a bit from that. All in all, there is a surprisingly small difference between them on a per core basis. [Edit 1 times, last edit by Jim1348 at Jan 3, 2018 4:18:12 PM] |
||
|
|
starlight
Cruncher England Joined: Jun 29, 2005 Post Count: 42 Status: Offline Project Badges: 
|
Terribly sorry, but a day for a WU is too short. Trying to run this project is a waste of my resources. Starting a thread, and getting it bounced before it completes does nobody any good.
----------------------------------------I’ve only got a 1st gen core i7. Please extend a day to something sensible like five days. [Edit 1 times, last edit by starlight at Jan 13, 2018 6:29:52 PM] |
|||
|
||||
|
|