Socket 939/940 folders, question

[=ACID RAIN=]

What is your typical PPD? I'm getting below 200 on each of my 939 systems and I'm wondering if this is normal or not.

3500+: 140 PPD
Opteron 146 @ 2.5GHz: 165 PPD

Does this seem on par?

[WarriorII]

Harlam can answer this one.
He's a big fan of AMD Folding.

-was. :-/

[=ACID RAIN=]

LOL.

If I can get my hands on a better heatsink for the opty, I can really abuse it in terms of OCing (or a radiator/heatercore, I have 939 watercooling equipment + pump).

The 3500+ system won't OC at ALL. I'm not sure if swapping out the caps mattered on the mobo, or if it's the CPU, but it won't do squat haha.

[ChasR]

60 ppd/Ghz is "normal" for small gromacs WUs on 939/940 socket . Bonus (large) WUs provide 65% to 180% more, the upper range depending on cache size on "tet" WUs (292 pointers).

I've got one left, the FX-53 I use at home daily. I put a HD 3850 AGP 512 MB in it and now make 1200+ ppd on it.

[VinnyTAMU]

What is your typical PPD? I'm getting below 200 on each of my 939 systems and I'm wondering if this is normal or not.

3500+: 140 PPD
Opteron 146 @ 2.5GHz: 165 PPD

Does this seem on par?

That is about what I was getting for a stock 146. I decided it wasn't worth the PPD/energy usage so I stopped folding with it.

[harlam357]

Yep... ChasR has the numbers down. :)

I have a 3500+ as my development box at work, soon to be one of my only single core Folders left, and I'm seeing 138ppd out of FahMon on p3734.

...and WII is right, I was (really still am), a big AMD fan. I hadn't had an Intel rig since I built my first TBird 1.2GHz back in 2001. But AMD just rested on their success and now they can't compete. I waited for Phenom but it just can't compete with Intel for Folding, so I have been migrating. I'll be keeping a few AMD rigs, but the majority of my production will now come from s-I-n-tel. :)

[15:35:49] *------------------------------*
[15:35:49] Folding@Home Gromacs Simulated Tempering Core
[15:35:49] Version 1.10 (Oct 4, 2007)
[15:35:49]
[15:35:49] Preparing to commence simulation
[15:35:49] - Assembly optimizations manually forced on.
[15:35:49] - Not checking prior termination.
[15:35:49] - Expanded 365641 -> 1800796 (decompressed 492.5 percent)
[15:35:49]
[15:35:49] Project: 3734 (Run 87, Clone 0, Gen 16)
[15:35:49]
[15:35:49] Assembly optimizations on if available.
[15:35:49] Entering M.D.
[15:36:10] (Starting from checkpoint)
[15:36:10] Protein: p3734_Seq18_Amber03_Extended
[15:36:10]
[15:36:10] Writing local files
[15:36:52] Completed 285000 out of 1500000 steps (19)
[15:36:52] Extra SSE boost OK.
[16:06:52] Timered checkpoint triggered.
[16:13:55] Writing local files
[16:13:55] Completed 300000 out of 1500000 steps (20)
[16:43:55] Timered checkpoint triggered.
[16:50:30] Writing local files
[16:50:30] Completed 315000 out of 1500000 steps (21)
[17:20:30] Timered checkpoint triggered.
[17:28:54] Writing local files

[=ACID RAIN=]

I've got one left, the FX-53 I use at home daily. I put a HD 3850 AGP 512 MB in it and now make 1200+ ppd on it.
THAT is what I'm talking about. Nice. I've been considering the same thing for the 939 folders here.

Thanks folks, this has definitely been helpful. If I can get my laptop sold for ~150, I can get me a 3870 for the farm rig (it's PCI-E).

['Cuda340]

I'll be keeping a few AMD rigs, but the majority of my production will now come from s-I-n-tel.

You can lead a horse to water, but you can't make it drink. ;) j/k H :beer:

[=ACID RAIN=]

LMAO.

Oh lord. Let's not get that started haha. AMD vs Intel can turn the best of friends into a knock-em-out bloodfeud!

['Cuda340]

H knows i'm just having fun. :beer:

For that matter i could care less who folds on what as long as they are FOLDING.

[drshivas]

My wife's 939 X2 @2.6 gets anywhere between ~150-350ppd per core, depending on the WU spread, with the classic FAH client.

My Opty 170 @2.7 gets ~850-1k PPD with WinSMP.

[harlam357]

Oh lord. Let's not get that started haha. AMD vs Intel can turn the best of friends into a knock-em-out bloodfeud!

Yeah... no worries. I can take it. ;)

My faith was misplaced in AMD... I've known that for a while now. Not that I really like Intel as a company... but I have to go with what "works" for me. Again, I'm still an AMD fan and wish they could get their stuffing together. I would gladly move back if it made sense.

['Cuda340]

and wish they could get their stuffing together.

As do i & i believe they will ! :beer:

[harlam357]

btw cuda- I owe you a PM, just haven't had time to sit down and put all my thoughts down.

It will be forthcoming. ;)

[=ACID RAIN=]

Oh I hear ya H. I bought AMD exclusively until C2D came out. That goes all the way back to the K6-2 350MHz haha. In the end though, I am a performance fan more than a brand fan. This time the Q6600 won hands down.

Next time, Gadget! NEXT TIIIIME!!
http://stoneywageslave.files.wordpress.com/2007/05/drclaw.jpg

[Wega!]

I've been running the system in my sig for a couple of weeks now, and it gave me around 750PPD, running in Win XP with the SMP console.

At the moment I'm testing out a VWware/Ubuntu/SMP console setup with a single core client still running in win XP. But it seams it can't give me more than the 750PPD I made running only win XP... Don't know why..

[=ACID RAIN=]

Might be because those chips have much less performance in a virtual setting than current intel chips. No optimizations for it.