Message boards :
Number crunching :
Windows port of Alex v8 code
Message board moderation
Previous · 1 · 2 · 3 · 4 · 5 . . . 50 · Next
Author | Message |
---|---|
JDWhale Send message Joined: 6 Apr 99 Posts: 921 Credit: 21,935,817 RAC: 3 |
John, well done and congratulations on putting your money where you mouth is. Thanks, They ran, but won't validate, I'll try again and not send the WU's in till the numbers match Next time I won't promote "real time". I've got a bunch of detatched WUs that are being validated by wingmen and will end up "too late" to be accepted... Lots to practice on. LOL I wonder if the runtimes are pretty close, though? Later, John |
zoom3+1=4 Send message Joined: 30 Nov 03 Posts: 65738 Credit: 55,293,173 RAC: 49 |
:) I've had over 4,100 RAC on just one PC alone with SSSE3 v2.4V and 6.10. I've got a cooling problem with the cpu that forced that PC down from 3.3GHz to 2.85GHz, So I'd rather stay with what I have. But then thanks anyway. The T1 Trust, PRR T1 Class 4-4-4-4 #5550, 1 of America's First HST's |
Logan Send message Joined: 26 Jan 07 Posts: 743 Credit: 918,353 RAC: 0 |
:) I said 2.4 from lunatics, Joker. Not 2.4v SSSE3 from Crunch3r. Regards. Logan. BOINC FAQ Service (Ahora, también disponible en Español/Now available in Spanish) |
zoom3+1=4 Send message Joined: 30 Nov 03 Posts: 65738 Credit: 55,293,173 RAC: 49 |
:) I did understand what You typed, I just prefer what works better for Me and XP x64 here. The T1 Trust, PRR T1 Class 4-4-4-4 #5550, 1 of America's First HST's |
Logan Send message Joined: 26 Jan 07 Posts: 743 Credit: 918,353 RAC: 0 |
:) I'm sorry my friend. I see after my last message what you are running 64 bits windows in your computers. (and sorry for my bad english...) Logan. BOINC FAQ Service (Ahora, también disponible en Español/Now available in Spanish) |
zoom3+1=4 Send message Joined: 30 Nov 03 Posts: 65738 Credit: 55,293,173 RAC: 49 |
:) That's ok, My 2.4v apps are native x64 apps and last I looked the 2.4 are not. The T1 Trust, PRR T1 Class 4-4-4-4 #5550, 1 of America's First HST's |
Logan Send message Joined: 26 Jan 07 Posts: 743 Credit: 918,353 RAC: 0 |
:) I think I found it... When you see SSE3, forget it, really is SSSE3. Take the C2D version.... for 64 bits... Logan. BOINC FAQ Service (Ahora, también disponible en Español/Now available in Spanish) |
JDWhale Send message Joined: 6 Apr 99 Posts: 921 Credit: 21,935,817 RAC: 3 |
Here's an interresting WU where I was teamed with a Q6600 running optimized 2.4V... So if he isn't OC'd we are running about the same clock as my E4500 at 2418MHz though I've only got 2MB shared L2 vs. quads 2x4 shared L2 cache. I replaced the mismatched memory with a 2x1GB DDR2-800 Mhz set from Fry's ($44) yesterday. So I've been crunching both processors for the past 24 hours or so, CPU times are stable as I've only been crunching the 63.98 credit WUs. To bad the MOBO tops out at DDR2-667MHz... maybe a little more with the 10% OC. I'm moving that CPU to another MOBO so will probably get a new host ID later tonight. Still working to resolve the mid-angle issues though I'm happy with the low and high angles for now. I've run out acceptable WU's to crunch for now so am switching back to 2.4V since I'm heading to Las Vegas for some R&R. I'll Check back in about a week. Ciao, JDWhale |
Fred W Send message Joined: 13 Jun 99 Posts: 2524 Credit: 11,954,210 RAC: 0 |
Here's an interresting WU where I was teamed with a Q6600 running optimized 2.4V... So if he isn't OC'd we are running about the same clock as my E4500 at 2418MHz My Q6600 (G0) clocked at 3336MHz (XP Home) reports 3273 FP and 7476 Integer speeds; your wingman (Vista) reports more than twice the FP and 25% more Integer. I know the benchmarking is flakey, but those figures are closer to Mark Sattler's than mine so my guess is he's overclocked (by a lot!!). F. [Edit]Enjoy the R&R[/Edit] |
Jason A. Countryman Send message Joined: 29 Aug 03 Posts: 139 Credit: 50,172,873 RAC: 2 |
Nope. he is running stock. This is right out of his stderr out: Speed: 4 x 2401 MHz |
SATAN Send message Joined: 27 Aug 06 Posts: 835 Credit: 2,129,006 RAC: 0 |
When I upgrade this mac to 6.1.0 my benchmark score almost double. |
kittyman Send message Joined: 9 Jul 00 Posts: 51468 Credit: 1,018,363,574 RAC: 1,004 |
When I upgrade this mac to 6.1.0 my benchmark score almost double. While that my look pretty, benchmarks mean nothing in Seti.....and it has been documented in this forum that benchmarks vary wildly depending on OS and other factors...... Have your crunch times changed, on equal AR WUs? "Freedom is just Chaos, with better lighting." Alan Dean Foster |
Fred W Send message Joined: 13 Jun 99 Posts: 2524 Credit: 11,954,210 RAC: 0 |
And stderr_txt on one of my WU's says: Speed: 4 x 3752 MHz and I running the same Op App but at 4 x 3336. So how much credence can we give to the stderr_txt?? I'll get his crunch-times to compare against mine; that will be definitive... F. |
SATAN Send message Joined: 27 Aug 06 Posts: 835 Credit: 2,129,006 RAC: 0 |
Mark, that's my point, the benchmarks mean diddlysquat. Can't answer about the other as Alex's doesn't give proc speed in the output files. Although it is stock 2.8 |
Andre Howard Send message Joined: 16 May 99 Posts: 124 Credit: 217,463,217 RAC: 0 |
Was brought up the other day I believe, someone mentioned if you run q6600 at 8x multipier it inflates the speed shown in stderr_txt |
kittyman Send message Joined: 9 Jul 00 Posts: 51468 Credit: 1,018,363,574 RAC: 1,004 |
Mark, that's my point, the benchmarks mean diddlysquat. Yes, I always found the CPU speed in result file to be very helpful in comparing the performance of various computers.....I think all apps, optimized or stock, should include that feature..... "Freedom is just Chaos, with better lighting." Alan Dean Foster |
Fred W Send message Joined: 13 Jun 99 Posts: 2524 Credit: 11,954,210 RAC: 0 |
Brownie-point to you sir. Data-Vac says that the stderr is correct in this case. Apologies for doubting your assurance and for muddying the waters, here. Benchmarks are sigmificantly more meaningless than I thought!! [Edit] Here's what I mean: Direct Link [/Edit] F. |
kittyman Send message Joined: 9 Jul 00 Posts: 51468 Credit: 1,018,363,574 RAC: 1,004 |
Here's an interresting WU where I was teamed with a Q6600 running optimized 2.4V... So if he isn't OC'd we are running about the same clock as my E4500 at 2418MHz Have a nice vacation, but when you get back...... jason gee has posted a few beta compiles over at Lunatics, which I ran a knabench test on..... If you wanna post some test code over there, I would luv to do a comparo..... "Freedom is just Chaos, with better lighting." Alan Dean Foster |
JDWhale Send message Joined: 6 Apr 99 Posts: 921 Credit: 21,935,817 RAC: 3 |
Here's an interresting WU <snip> @Fred - The dice don't stand a chance ;-) I thought my quad teammate was/is running stock, that is why I brought up the comparison... I've got some compatable WU's, both VHAR & VLAR, coming up on my quads, but don't want to suspend a bunch of tasks. So I don't know if I'll get them crunched before I leave. But if I finish getting this code sorted... I might be able to turn it loose on one before I leave. Okay, so my 10% OC'd E4500 @ 2418 might not be able to keep up with an OC'd quad, the point that I failed to make is the performance we're seeing from a chip with only 1MB vs. 2MB L2 cache/core on the Q6600. Running virtually clock for clock, the Conroe didn't run away from the Allendale. Rather the preliminary findings are that the "Alex-v8" port to WinXP is ~17% faster per core than v2.4 for this AR on a lesser machine. But then it just might be the new memory sticks I put in that explains the performance we're seeing ;-) -JDWhale @Mark - We'll see, I kind of took this quick port on in response to a couple challenges by folks here on the MB that made it sound hard... So I thought I'd see what I could do in a day. I really don't enjoy coding all that much any more. This was the first code I've messed with in nearly 2 years... other than the BlackJack "basic strategy" tester I wrote in Java a couple months back to test my play (the "Simpsons" yell at me when I make a mistake). Personal record is 500+ hands without making a mistake, but as time drags on the mistakes become much more frequent. Cheers, |
David Send message Joined: 19 May 99 Posts: 411 Credit: 1,426,457 RAC: 0 |
Apologies for doubting your assurance and for muddying the waters, here. Benchmarks are sigmificantly more meaningless than I thought!! I have 2 PC's here that have massively different benchmarks, but one runs at 3033 & the other 3100, same chip, same ram, same board (Almost), almost identical video cards & same OS. PC1: Measured floating point speed 3477.79 million ops/sec Measured integer speed 11049.44 million ops/sec PC2: Measured floating point speed 2986.16 million ops/sec Measured integer speed 6593.91 million ops/sec The RAC's are close enough to the same, so I cant see there being such a massive difference, but there is in the benchmarks, just not in reality |
©2024 University of California
SETI@home and Astropulse are funded by grants from the National Science Foundation, NASA, and donations from SETI@home volunteers. AstroPulse is funded in part by the NSF through grant AST-0307956.