| Index | Recent Threads | Unanswered Threads | Who's Active | Guidelines | Search |
 |
World Community Grid Forums
Category: Support
Forum: BOINC Agent Support
Thread: New Phenom Help |
| No member browsing this thread |
|
Thread Status: Active Total posts in this thread: 8
|
|
| Author |
|
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
I just picked up a Gateway Phenom based pc. It comes with Vista 64 bit. I loaded Boinc 64 bit and my WU processing times were slower than my Athlon XP 4000 based computer. A lot slower. I then removed Boinc 64 and loaded Boinc 32 bit. Better but still really bad. I just loaded Ubuntu onto this machine and loaded Boinc. So far the numbers are crunching as fast as my Intel Quad running Vista 32 bit. What is going on? I can crunch in Ubuntu at night, but would love to just leave it in Vista all the time. Could it be Vista 64? Has anyone heard anything about trouble crunching in Vista 64?
|
||
|
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
You should get almost identical performance.
What criteria are you using to judge the relative speed? |
||
|
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
You should get almost identical performance. What criteria are you using to judge the relative speed? The time it takes to crunch a WU. I know they vary but there seems to be a range. Say with Help Conquer Cancer. My Quad averages about 4-5 hours per unit. My old AMD 6-7. The new Phenom 7-8 on Vista, I have one crunching in linux now that looks to be coning in at about 6 though I should wait to see it's final time. Numbers get odder with Fightaids. Quad and AMD about 5 hours. Phenom 15 hours!!! [Edit 1 times, last edit by Former Member at May 2, 2008 2:23:19 PM] |
||
|
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
The techs have been adjusting work unit sizes recently. I don't think you can rely on work unit sizes being constant, or even similar.
But we can check some other things. What do the benchmarks report? If you rerun the benchmark, does it remain roughly the same? Are you using all 4 (or 3) cores? Are all cores running at 100%? |
||
|
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
I think I found the problem reading through other project boards, AMD Cool'n'Quiet. Boinc doesn't run at a high enough priority and the CPU speed is cut in half at some point when there is no high priority activity. I disabled it in bios this afternoon. I have a bunch of WUs running in Linux 32 bit (what I had on hand) which are showing much much better speeds. My cache in Linux will be empty by tomorrow and I'll run in Vista again to see the numbers I get. I'll check all the benchmarks in each version of Boinc sometime this weekend.
----------------------------------------[Edit 1 times, last edit by Former Member at May 3, 2008 2:39:49 AM] |
||
|
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
Out of curiosity what are the CPU benchmarks for the Phenom?
You can click Advanced in BOINC and click Run CPU Benchmarks. After the benchmarking is complete click on the Messages tab. That will show the results of the benchmarks. |
||
|
|
RAID Xero
Cruncher Joined: Mar 17, 2006 Post Count: 22 Status: Offline Project Badges: 
|
You have to remember that the AMD Phenom DOES NOT preform as well as the Intel quad cores. The AMD is slower and I think that's why it takes an hour longer on average. Look on the bright side, you still have 4 cores crunching all the time.
 |
||
|
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
My Quad Benchmarks:
Phenom 9600 2328 Whetstone 4997 Dhrystone Intel Q6600 2338 Whetstone 5134 Dhrystone The Phenom is still averaging about double the time it takes the Q6600 to do a WU. Which is what has me confused. Very odd, but yes another quad and I got it on sale. I just ordered 4 gbs of ram to max out the memory in the Phenom. Running Vista 64bit, it will see all 8gbs. I'll see if that does anything. |
||
|
|
|