• We’re currently investigating an issue related to the forum theme and styling that is impacting page layout and visual formatting. The problem has been identified, and we are actively working on a resolution. There is no impact to user data or functionality, this is strictly a front-end display issue. We’ll post an update once the fix has been deployed. Thanks for your patience while we get this sorted.

Some seti times for my new P4 3ghz @ 3.4ghz w/ 2 megs cache

W

Wiz

Diamond Member
This is finally running under 60c so I can turn in some preliminary times for seti1.
Remember, this is running two processes simultaneously so:
2 WU's in 2:26 = 1:13 each
and
2 HAR WU's in 1:49 = 54.5 minutes each.

I got 22 WU's from this thing in the past 24hours. That was with a lot of down time while figuring out the throttling issue.

2005 Mar 15 05:15am 62.235,18.08 0.426 BLACKY 2:26
2005 Mar 15 04:16am 62.213,18.08 0.426 BLACKY 2:26

2005 Mar 15 06:08am 41.934,16 7.574 BLACKY 1:49
2005 Mar 14 07:42pm 41.934,16 7.574 BLACKY 1:49
2005 Mar 14 07:28pm 41.739,26 7.591 BLACKY 1:50
2005 Mar 14 04:59pm 41.934,16 7.574 BLACKY 1:49

I bumped it a few more mhz on the fsb today & am watching temp & Throttling - so far 52c and zero throttling so I may try for 3.5ghz. 😉 This is all with a 1:1 fsb / mem ratio.

I discovered the cpu will go way up there but tne mem can not follow - besides heat gets to be too much of an issue & I don't have any Arctic Silver so I just put the HSF on with what they (intel) supplied. There's more room in this system with better memory & cooling it might get to 4ghz+
 
Nice cruncher!

I've decided to pass on the 925/915 generation, heat being the biggest issue. I read today that the new dual core CPU's coming 2H '05 will not be compatible with the 925 chipset. That means all these new S775 mobo's will be outdated soon, just like they did to the P4 S423. I still think S478 is the best bang/buck you can get. 🙂
 
I didn't go for a 915/925 mobo pretty much for that reason - and the fact that I could use my memory & video card with a P5P800, which seems to be a nice board for the price.

Oh, and finally I got an answer to the age old question of whether it was my memory or my AMD system that would not go over 170/340 fsb without problems.
Now running 231/462 fsb rock solid with the same ram.
 
A7N8X Deluxe running 170 / 340 fsb with a 1700+ clocked @ 2ghz

Also have a Epox 8RDA+ with a 1700+ running 180 / 360fsb @ 2.1ghz

They put out WU's in about the 2:20 - 2:30 range.
Just singles though, so those two systems put out less WU's per day than this one P4.
 
Originally posted by: Wiz

...I bumped it a few more mhz on the fsb today & am watching temp & Throttling...
Click to expand...

Hey Wiz, the Throttling you speak or interests me. I have a P4 2.8 with HT in a Toshiba notebook. A few months ago the WU time went up by about 50% and have stayed there. I use to get 12 - 15 WU/Day, now I am lucky to get 8. CPU-ID says it is running at 2.8 GHz. TaskManager show each log. CPU getting 49% of the CPU. I see nothing that would indicate a slowdown. How do I tell if she is Throttling?

-----Update

A web search for CPU Throttling brought me to a site where I download Throttle Watch by Panopsys. It shows my speed at 2792.85 MHz and 0% Throttling. So, what else could do this?

Originally I was running Client ver. 3.08, then when the problem started I deleted that and went with 3.03. No change in problem.

 
Even though you're doing 2 wu in 2:26, just diving by 2 doesn't mean that's how long 1 takes. I have a P4 3.06Ghz and I have HyperThreading turned on and I am processing 2 at a time. Try running 1 wu and you'll get a faster computing time that 2:26, but more that 1:13. With HT on, you are able to crunch more WU's in a 24 hr period than running 1 at a time. If you run SetiSpy, you'll notice each wu taking close to that 2:26. That's wall clock time.
 
Originally posted by: vrmuds
Even though you're doing 2 wu in 2:26, just diving by 2 doesn't mean that's how long 1 takes. I have a P4 3.06Ghz and I have HyperThreading turned on and I am processing 2 at a time. Try running 1 wu and you'll get a faster computing time that 2:26, but more that 1:13. With HT on, you are able to crunch more WU's in a 24 hr period than running 1 at a time. If you run SetiSpy, you'll notice each wu taking close to that 2:26. That's wall clock time.
Click to expand...


You are correct vrmuds, we all tend to cut those times in half. But, we are really looking at throughput and the effective time for each WU. I have a Dell XPS notebook that runs two WUs in 2h16m. Regardless of how one braggs about it, that is nice output for one computer!
 
Originally posted by: vrmuds
Even though you're doing 2 wu in 2:26, just diving by 2 doesn't mean that's how long 1 takes. I have a P4 3.06Ghz and I have HyperThreading turned on and I am processing 2 at a time. Try running 1 wu and you'll get a faster computing time that 2:26, but more that 1:13. With HT on, you are able to crunch more WU's in a 24 hr period than running 1 at a time. If you run SetiSpy, you'll notice each wu taking close to that 2:26. That's wall clock time.
Click to expand...


Hey your right its actually 1.09090909 according to the calculator.

try it yourself divide 24 by 22 and see what you get.

Whether or not it actually does this in 2.26 or ?.... 2 instances are done in that time, and 22 are done in a 24 hour period.

Bottom line is he is correct,.

Please correct me if i am wrong.

Mike
 
You can do that, but really, that's assuming you're doing it one at a time... I used to run this on a few quad P4 servers and each server was pushing about 64 WU a day. Does that mean each only took 22.5 minutes to crunch??

I really don't want to get into a big discussion about all of this, which proably has happened over and over around here.

Anyway, every wu counts, and that's what really matters around here.
 
Originally posted by: vrmuds
You can do that, but really, that's assuming you're doing it one at a time... I used to run this on a few quad P4 servers and each server was pushing about 64 WU a day. Does that mean each only took 22.5 minutes to crunch??

I really don't want to get into a big discussion about all of this, which proably has happened over and over around here.

Anyway, every wu counts, and that's what really matters around here.
Click to expand...

I'll drink to that, the :beer: is on me :beer:😛:beer:

I do agree with you btw, on all counts 😛

:beer:😛:beer:

Mike
 
Personally I don't care what time per wu is. Total production in a 24 hour time period is what counts and hyperthreaded p4s or higher processers are hard to beat as far as seti goes.
 
I think we should be bragging on how much we can push/day...

Hehehe... I am in the middle of a new build, AMD FX-55 ,OCZ VX memory 2-2-2-8 at PC4000, and dfi's sli-dr board..... I'll be cranking this thing up.. Beers all around!!!!
 
Originally posted by: vrmuds
I think we should be bragging on how much we can push/day...

Hehehe... I am in the middle of a new build, AMD FX-55 ,OCZ VX memory 2-2-2-8 at PC4000, and dfi's sli-dr board..... I'll be cranking this thing up.. Beers all around!!!!
Click to expand...

How many a day do yiou figure that beast will produce?

MMMM :beer:

Mike
 
Originally posted by: vrmuds
Even though you're doing 2 wu in 2:26, just diving by 2 doesn't mean that's how long 1 takes. I have a P4 3.06Ghz and I have HyperThreading turned on and I am processing 2 at a time. Try running 1 wu and you'll get a faster computing time that 2:26, but more that 1:13. With HT on, you are able to crunch more WU's in a 24 hr period than running 1 at a time. If you run SetiSpy, you'll notice each wu taking close to that 2:26. That's wall clock time.
Click to expand...

Indeed. Actually, time per wu [/b]increases[/b] using HT because each process is sharing resources.

That said, HT is by FAR the best thing going for Dist Computing today. The FX55 system below will be a mighty cruncher, but it's daily output is still not going to best a high MHz P4 system with HT.

 
Originally posted by: Wiz
This is finally running under 60c so I can turn in some preliminary times for seti1.
Remember, this is running two processes simultaneously so:
2 WU's in 2:26 = 1:13 each
and
2 HAR WU's in 1:49 = 54.5 minutes each.

I got 22 WU's from this thing in the past 24hours. That was with a lot of down time while figuring out the throttling issue.

2005 Mar 15 05:15am 62.235,18.08 0.426 BLACKY 2:26
2005 Mar 15 04:16am 62.213,18.08 0.426 BLACKY 2:26

2005 Mar 15 06:08am 41.934,16 7.574 BLACKY 1:49
2005 Mar 14 07:42pm 41.934,16 7.574 BLACKY 1:49
2005 Mar 14 07:28pm 41.739,26 7.591 BLACKY 1:50
2005 Mar 14 04:59pm 41.934,16 7.574 BLACKY 1:49

I bumped it a few more mhz on the fsb today & am watching temp & Throttling - so far 52c and zero throttling so I may try for 3.5ghz. 😉 This is all with a 1:1 fsb / mem ratio.

I discovered the cpu will go way up there but tne mem can not follow - besides heat gets to be too much of an issue & I don't have any Arctic Silver so I just put the HSF on with what they (intel) supplied. There's more room in this system with better memory & cooling it might get to 4ghz+
Click to expand...

Let me know how many wu's you get done at 3.5GHz Wiz. Last night I cranked up my 3.06 system to ~3.5GHz (150 fsb, 600QDR)......we can get a nice comparison of Prescot versus Northwood and fsb speeds.

Last time I did an evaluation of ram speed vs seti times, I found that raw CPU clock had the biggest impact on seti times. Increasing the ram speed did almost nothing! It was quite interesting, I posted it here almost a year ago.

 
Originally posted by: Wolfsraider
Originally posted by: vrmuds
I think we should be bragging on how much we can push/day...

Hehehe... I am in the middle of a new build, AMD FX-55 ,OCZ VX memory 2-2-2-8 at PC4000, and dfi's sli-dr board..... I'll be cranking this thing up.. Beers all around!!!!
Click to expand...

How many a day do yiou figure that beast will produce?

MMMM :beer:

Mike
Click to expand...

I'm not too sure.... I started running seti last night to do a base test. Everything is at base settings. I started at 10pm and when I check things at 6:30, at the rate it's going, it's going to best my P4 3Ghz with HT on and only 1 wu is being processed at a time.

I still need to install the UV sleeving and clean up the cables. My second video card is coming in on Friday so I can run SLI (Asus 6800GT 256MB cards). I got to admit, the asus card looks cool. There's a blue diode that lights up which creats some nice ground effects and the fan system lights up!!... My next step is to overclock the system, then run some more tests. After I'm satisfied, I'm going to reinstall the OS.
 
what times are you getting per wu with that beast (the fx-55)? im assuming your 3.0 P4 HT is cranking out anywhere from 14-17 wu per day?!?
 
Originally posted by: Unforgiven
what times are you getting per wu with that beast (the fx-55)? im assuming your 3.0 P4 HT is cranking out anywhere from 14-17 wu per day?!?
Click to expand...

I'm not sure yet... I'm whatchin smoke's seti page and saw my score bump up pretty fast.. I'm already at 20 for today and I'm usally do about 15-16/day depending on if I game or not, which sometimes drop to 14 (On the weekends 😛).
 
Originally posted by: Wiz
This is finally running under 60c so I can turn in some preliminary times for seti1.
Remember, this is running two processes simultaneously so:
2 WU's in 2:26 = 1:13 each
and
2 HAR WU's in 1:49 = 54.5 minutes each.

I got 22 WU's from this thing in the past 24hours. That was with a lot of down time while figuring out the throttling issue.

2005 Mar 15 05:15am 62.235,18.08 0.426 BLACKY 2:26
2005 Mar 15 04:16am 62.213,18.08 0.426 BLACKY 2:26

2005 Mar 15 06:08am 41.934,16 7.574 BLACKY 1:49
2005 Mar 14 07:42pm 41.934,16 7.574 BLACKY 1:49
2005 Mar 14 07:28pm 41.739,26 7.591 BLACKY 1:50
2005 Mar 14 04:59pm 41.934,16 7.574 BLACKY 1:49

I bumped it a few more mhz on the fsb today & am watching temp & Throttling - so far 52c and zero throttling so I may try for 3.5ghz. 😉 This is all with a 1:1 fsb / mem ratio.

I discovered the cpu will go way up there but tne mem can not follow - besides heat gets to be too much of an issue & I don't have any Arctic Silver so I just put the HSF on with what they (intel) supplied. There's more room in this system with better memory & cooling it might get to 4ghz+
Click to expand...

How does that compare to the 1MB Presc & the Northwoods?

Btw the A7N8X v1.0x were not actually rated for 200MHz FSB (nor were many early Nfrc2 mbrds) (despite what some parts of Asus's website make out!:roll: ),they usually topped out between 175-190 though some did 200.This can be got around by upping the chipset voltage which can only be done via mbrd mod!
I had got mine upto 180MHz ,though I dropped it down to 179 yesterday as I was getting occasional BSOD, though that might be down to spyware!:| ,haven't determined which yet but it was ok for about 3mths at 180MHz.
At 180FSB my XPM 2500 is at 2.53GHz @1.66v 😀

 
It had a string of HAR units last night - that being said it put out 15 wu's in the last 12 hours @ 231 / 462 fsb 3.5ghz cpu 😉

BTW, it may seem simplistic but getting an -average- time for each unit by dividing the total time for two simultaneous units by 2 works quite well for me.

You can also take the total number of units done in a number of hours and divide by that number of hours. In which case for the past twelve hours my average WU time is 48 minutes each.

I've been doing seti for almost 6 years, I know that the actual processing time per individual unit is actually greater, (who doesn't know this?) however it's obvious that more work gets done this way than doing them one at a time.
To me it's just like having more than one machine running doing the same project. My average number of WU's per hour is greater than if I relied on only one cpu.

Badthad - it won't stay stable @ 3.5ghz - I lowered it to 3.46 after only a few minutes of trying it @ 3.5 - my ram is PC3200. It's been totally stable @ 3.46

JWMiddleton, run Throttlewatch & keep it open, hit F5 to start logging & ler her run a long time then check the log file. It might only be throttling at times you don't notice.
Otherwise, maybe it's doing some power saving?
 
JW, also make sure to clean the HS\F from time to time, they clog up fast on notebooks. How about running sandra cpu benchmark just to see whats going on with the floating point score's.
 
You must log in or register to reply here.

TRENDING THREADS

Back
Top