just landed a biggie!!!

2605 171.64.65.56 p2605_Protein in POPC 77373 3.00 4.00 1760.00

oh my god!!!!......1760 points

currently chewing at 20 minutes a frame on my SMP box at home......hope it dont EUE!!!

i wish it was on my mac pro

lee

stop it and copy it as is quick !!!!

what are the settings on that machine? big packets? advmethods?

how much ram does the machine have?

give us some screenies of EMI crunching away on it !!!

Nice.

Nice i came into work to find all of my smp boxs siting there doing nothing after euing over the weekend

leelegend said:

2605 171.64.65.56 p2605_Protein in POPC 77373 3.00 4.00 1760.00

oh my god!!!!......1760 points

currently chewing at 20 minutes a frame on my SMP box at home......hope it dont EUE!!!

i wish it was on my mac pro

lee

Click to expand...

machine spec??

It's an SMP only WU that has a low weight, so you don't see them very often, although they seem to have given more out as they transition in some new WUs.

My E6600@3400Mhz does about 10m 20s per frame.

LandShark said:

machine spec??

Click to expand...

Lee, which machine is that? I've got one on my P5B/[email protected] doing about 19 minutes, which I thought was slow; however, the P5B is back on kubuntu "live" which I thought was the reason to the TPF length. I have another 2605 on the K8N-DL(2x2)Opty rig which ironically is also running live. TPF for the x2 Opty rig is 17:11.

Tim, running live sounds like it is killing your perf.

I have all 3 of the smp rigs running the 2605 atm with frame times of 11:30 - 12:02 ... they are using 90% of both cores vs 80% on the 3000's and 375-405 MB.

Rigs are 6600's running at 2.9-3 ghz.

I wonder if stanford benchmarking is off cuz my TPF translates to 2100-2200 ppd per rig??!!??

/edit one of them is going to finish today so it should be a noticable bump

my X2 4400 with 2gb of ram.....it doesnt like it much

[04:02:47] Completed 185000 out of 500000 steps (37 percent)
[04:05:47] Timered checkpoint triggered.
[04:08:47] Timered checkpoint triggered.
[04:11:47] Timered checkpoint triggered.
[04:14:48] Timered checkpoint triggered.
[04:17:47] Timered checkpoint triggered.
[04:20:48] Timered checkpoint triggered.
[04:22:26] Writing local files
[04:22:26] Completed 190000 out of 500000 steps (38 percent)

give me a normal SMP unit any day of the week....9mins a frame usually

lee

Wow, you guys folded those slow! My E6300 @3.15ghz folded it at 11.5 minutes a frame!

I've got 'em on the farm as well... my 2.6GHz X2 keeps carping out on this one @ ~41%.

Project: 2605 (Run 0, Clone 45, Gen 0)

Anyone else get this particular R/C/G WU? This rig was great on p3007-p3009. So I'm apt to blame the WU & not the OC that was also 100% stable in Windows. I tried to totally delete the WU and download a new one, but guess what... same p2605 (Run 0, Clone 45, Gen 0)... again. So I'll let it Fold until this evening and see if I can get past 41%. If not, this rig is going back to Windows until more fixes arrive for SMP.

btw- grats to lee on his overtake on me that I assume will happen today.

After 1 6600 rig turned one in last night, I thought they were over after a quick eue on a 3007 followed by a 3008.

So I decided to play with the the oc a bit on this latest 6600 and managed to get it folding at 3.5 with 1.375V still on air. The p5w64 seems to overvolt quite a bit so it is really running at ~1.4+ ? although i don't know how to monitor this stuff in linux.

Then checked it this monrning to find the 3008 finished and another 2605 folding ... at 10:00 per frame

leelegend said:

my X2 4400 with 2gb of ram.....it doesnt like it much

[04:02:47] Completed 185000 out of 500000 steps (37 percent)
[04:05:47] Timered checkpoint triggered.
[04:08:47] Timered checkpoint triggered.
[04:11:47] Timered checkpoint triggered.
[04:14:48] Timered checkpoint triggered.
[04:17:47] Timered checkpoint triggered.
[04:20:48] Timered checkpoint triggered.
[04:22:26] Writing local files
[04:22:26] Completed 190000 out of 500000 steps (38 percent)

give me a normal SMP unit any day of the week....9mins a frame usually

lee

Click to expand...

You seem to have set your ckeckpoint to a low value ... I configure all of mine to the max 30 min. Don't know how much effect this has on folding, but i am guessing the overhead of checkpointing is probably higher in smp than regular folding since it needs to sync all the cores?

I got in the habit of long checkpoints to minimize the risk of corruption if checkpointing during a power hit etc, but it is a tradeoff if you start and stop cleints frequenetly.

harlam357 said:

I've got 'em on the farm as well... my 2.6GHz X2 keeps carping out on this one @ ~41%.

Project: 2605 (Run 0, Clone 45, Gen 0)

Anyone else get this particular R/C/G WU? This rig was great on p3007-p3009. So I'm apt to blame the WU & not the OC that was also 100% stable in Windows. I tried to totally delete the WU and download a new one, but guess what... same p2605 (Run 0, Clone 45, Gen 0)... again. So I'll let it Fold until this evening and see if I can get past 41%. If not, this rig is going back to Windows until more fixes arrive for SMP.

Click to expand...

Harlam, if you post the EUE over at FC in the smp forum one of the mod's can check to see if someone else has turned in the wu sucessfully. It could take a while if it was reassigned for someone else to complete it but if it eue's fast on other rigs it should show up pretty soon.


I have noticed especially on one of my smp rigs that i had to reduce the oc quite a lot from what i had it folding at under wxp - about 25 fsb to stop getting freezes.

Also, on the c2d's at least, this wu seems to drive avg cpu to 90%+ which the 3000's don't - the 3000's seem to hit about 80%.
I read about others at FC who said this 2605 was worse than the 3000's in that on their rigs it would only drive cpu to 70%? Hardware/cpu/platfrom archticture seems to have a lot to do with smp performance, or at least wiith how the smp cores are implemented.

The 2605's also seem to use about 180 MB more memory than the 3000's.

Thanks pete!

I'm gonna post it up on FCF when I get home... funny thing is, the sucker finally quit on its own and downloaded a p3008. Hopefully that one goes better. I do have another p2605 on my other SMP rig (completely stock clocks ATM) and it's cranking right along. So hopefully :cross fingers: the WU that was giving me trouble was indeed a dud.

nightmare!!!

takes 20minutes a frame, i go to portugal for 3 days (turning my rigs off)....come home and boot them up, and guess what

missed the deadline!!!

man, these SMP deadlines are harsh!

lee

leelegend said:

2605 171.64.65.56 p2605_Protein in POPC 77373 3.00 4.00 1760.00

oh my god!!!!......1760 points

currently chewing at 20 minutes a frame on my SMP box at home......hope it dont EUE!!!

i wish it was on my mac pro

lee

Click to expand...

What a relief. I'm getting 24:57 on my Athlon 64 X2 4200+ and 27:57 on my lightly OC'ed PD 805. (3.32 GHz) On another post someone said they were getting ~6 mins fps on a E6300 ...? (p2605) I was getting ready to retire my two slow boxes and add them to the small collection of outdated equipment I have in the closet. (Celey 346, Athlon 64 3000+, P4 2.60, Celey 900 MHz, etc) Yeah, I like that 1760 points thing, now if everything goes alright with the two machines.

Edit: Now, after reading this post again, I'm coming to the conclusion my machines are slow. Does anyone have any suggestions as to how I could speed up either one of these slowpokes?

i have a couple of these going again ... on my best 6600 @3.5 it is 10 min per frame, and on the next best @3.2 it running 11 min per frame.
I think Shell quoted a 6 Min time but i am wondering if he may have confused it with a 30xx wu? He would need to be folding around 4Ghz or more to get that kind of frame time, assuming frame times are consistant across 2605's. I do recall Shell had a very nice OC on his 6300.

JWS ... i am not sure if your tpf's are out of line with others on similar hardware. So far I only have smp running on 6600's. I recall harlam reporting around 19 min? on his amd x2? Clocks should be the the difference between yours and his.

Also, the 805 has a smallish cache and mem bandwidth which may be limiting things on it since smp is running 4 cores.

I don't have any amd's on the farm so intels i am more familiar with and i don't have an 805 ... just 1 830 left on the farm.

How does the 805 compare ppd vs running 2 windows clients?

I hope to get some time soon to try smp on one of my 930's so i can get a feel for the differnce but the fsb and cache delta vs a smithfiled/805 is quite signiificant to 'normal' folding performance even after normalizing for clocks. The type of wu's to compare is also important ... patty melts aren't going to compare well to anything ... while the smp wu's all have nice bonus, especially the 2605. Ribo's or 1495/97 would be a better comparison ... although bottom line for me is which config gets better ppd at any point in time.

Optie 165 at 2.8ghz is getting 17.5 mins a frame. Averaging about 89%. Both the c2d's seem to be about the same on processor usage w/frame times on the e6300's of about 12 mins.

I'm doing 2 at a time at 13min per frame

pscout said:

I recall harlam reporting around 19 min? on his amd x2? Clocks should be the the difference between yours and his.

Click to expand...

Currently running a p2605 on a stock 4400 @ 21:45 tpf & ram @ 2-2-2. I imagine the larger cache and the tight lats are the difference as the clocks on a 4200 stock are the same.