The 3rd Folding@Home Holiday Season Race in December 2008

[Pokey]

Thanks for putting the stats together Peter. :thumbsup:

It doesn't look good for the Folders team this year.

But we have a ways to go yet; maybe my hampsters will get their second wind..........................

 

[Drsignguy]

Originally posted by: Pokey
Thanks for putting the stats together Peter. :thumbsup:

It doesn't look good for the Folders team this year.

But we have a ways to go yet; maybe my hampsters will get their second wind..........................

Thats right, we have 3 weeks to go......and lets see if we can make it more interesting? Hmmm? :evil:

 

[petrusbroder]

I certainly hope so ...

and I have a noob question:

I have a 8600GT GPU from ASUS running on a computer with a 4 core Phenom @ 2.6GHz and with 3 GByte RAM.
According to a thread in this forum the card should deliver some 1 200 - 1 700 ppd (the latter when OC'ed)
I have the latest CUDA-driver (180.60)
I have the FAH GPU systray client 6.23.
The Project run is 2582 (R3, C60, G28, using the Amber-core).
FahMon reports just a measly 156.66 PPD. This project has taken morer than 1 d 11h ...

I tried to follow the guide @ Folding@home and installed the CUDA-driver first, then rebooted. Installed FAH client, rebooted. Then I configured the the client with my name and the TeAm#. No other configuration ...
I am running only Win XP SP3 -all updates, and Firefox.

Is there anything wrong, if yes, how should I fix it?

Should I have (e.g.) uninstalled the original NVIDIA-driver first? Cleaned out the registry? Anything?

Edit: all the spelling- and grammer errors.

 

[VirtualLarry]

When I built my quad-GPU rig with four 9600GSOs, I installed XP, chipset drivers, network drivers, then I installed ForceWare 180.70, from Guru3d. I didn't install any seperate CUDA drivers.

I downloaded the GPU systray client from the Folding@Home beta/advanced client download page, and I followed the instructions for multi-GPU (copying the F@H directory in appdata, and renaming the copies, then creating the shortcuts, renaming the shortcuts, and adding -gpu X to them). Then when I double-clicked on the shortcuts for the first time, I configured them. I added my name, team 198, checked the box for big WUs (10MB or something), configured it to use "slightly higher priority", and "do not LOCK CPU core" or something like that, then I set the MachineID to 2 (and higher for the other shortcuts).

Did you remember to check the box enabling the bigger WUs? Because I think that my WUs are using the Gromacs core, not the Amber core. I could be wrong.

 

[theAnimal]

Originally posted by: petrusbroder
The Project run is 2582 (R3, C60, G28, using the Amber-core).
FahMon reports just a measly 156.66 PPD. This project has taken morer than 1 d 11h ...

Project 2582 is not a GPU2 project.

 

[Insidious]

TheAnimal is right. Peter, I think you accidentally installed the basic CPU client instead of the GPU client.

-Sid

 

[petrusbroder]

Thanks mates, got it. I had accidently downloaded and installed the basic CPU client ... LoL
I have now installed the GPU client ... I'll see what the PPD will be!

Running now GPUv2 Gromacs ...

 

[petrusbroder]

After 4% done: 1070 PPD. Nice!

 

[Insidious]

You can run both Peter!

:thumbsup:

 

[petrusbroder]

I know ... but I'll take one client at a time ... the next turn will come to the SMP-client because I have a 4-core Phenom.

 

[Assimilator1]

rabrittain
Not pi$$ed me off at all, or AFAIK anyone else .
Re RAM speeds ,you're dead right, I forgot that the Ath64 family CPUs has RAM dividers that give odd speeds on some CPUs.
But it'd still be worth checking that you've got the right voltage set for your RAM even at default speed :thumbsup:.

When you're ready give me & Drsignguy , & anyone else who offers a shout .

 

[rabrittain]

I certainly will give you a shout - maybe starting now.

Since I changed the RAM speed back to stock, that machine now processes 1% of a FAH WU in 48 minutes instead of 35 minutes with DDR 400. I like the stability that I have now, but I would like to have that same stability with the faster memory speed.

You seem to indicate that raising the voltage a little bit could give me that.

Hmmmm.

 

[Insidious]

rabrittain, I remember using This thread for help when I was building mine last year.

I don't go for the MEGA overclocks, but it is still good to have a plan. First, I decide just how hot I am willing to let my CPU get. (for a Q6600, I think 60C is a good worst case number)

At stock clocks, using CoreTemp to monitor temperatures, I start by bumping Vcpu in very small ncrements until I get to 45C or so idling. I also like to set my RAM voltage to it's specified value immediately (this value is normally above the default of the Motherboard so it needs to be set correctly to let the RAM perform at it's rated speed and latency settings.)

Next, I do the isolate and maximize method to find the max speed of CPU and RAM independently.

Next, I find the best balance of the max speeds I found. I still haven't loaded my CPU, so my temperatures better be low enough to allow for that. (I like 45C or less to start with.)

Finally, I load my CPU with F@H while closely monitoring core temperatures. If they start to go too high, I back off the CPU speed a little. If still too high, I back off the Vcpu a little.

Eventually, you get a system that loads to ~60C or less and now you watch for stability. Memtest86 is a good tool for a quick indicator of gross instabilities, but nothing beats a few days of folding to really prove it out.

Good Luck!

:beer: :music: :music: :beer:

-Sid

EDIT: IMPORTANT, the first temps I entered were too high (those were AMD numbers) fixed now.

 

[VirtualLarry]

Intel C2D chips can handle up to 80C or more. It's the AMD chips that have to be kept cooler, like 60C, or depending on your chip, 70C.

 

[petrusbroder]

Due to time limits I'll do the stats every second day ... they will be done later tonight.

 

[Insidious]

Originally posted by: VirtualLarry
Intel C2D chips can handle up to 80C or more. It's the AMD chips that have to be kept cooler, like 60C, or depending on your chip, 70C.

This might explain some of the stability problems you've been working with. That's too hot.

-Sid

 

[biodoc]

Originally posted by: petrusbroder
Due to time limits I'll do the stats every second day ... they will be done later tonight.

Would you like me to take over the stats Peter? I know you've been very busy lately!

Folders!!! Recruit more crunchers!! Hard to believe just a few TAS people can beat the folders!!

 

[Assimilator1]

Originally posted by: VirtualLarry
Intel C2D chips can handle up to 80C or more. It's the AMD chips that have to be kept cooler, like 60C, or depending on your chip, 70C.

Hmm, maybe for some but I know for example that Intel quote a max of 71C for the Q6600 G0.

Personally speaking I would stay below 70C when loaded with OCCT or P95 (many DC projects won't get the CPU as hot by about 5-7C on my rig).

 

[VirtualLarry]

Originally posted by: Assimilator1

Originally posted by: VirtualLarry
Intel C2D chips can handle up to 80C or more. It's the AMD chips that have to be kept cooler, like 60C, or depending on your chip, 70C.

Hmm, maybe for some but I know for example that Intel quote a max of 71C for the Q6600 G0.

Personally speaking I would stay below 70C when loaded with OCCT or P95 (many DC projects won't get the CPU as hot by about 5-7C on my rig).

Intel specs quote Tcase, CoreTemp/RealTemp measure Tjunction, there is a gradient of about 10C or more between the two. So running at 80C Tjunction is still within 71C Tcase.

 

[VirtualLarry]

My production is going to be down a bit. I turned off my quad-GPU rig, it seems to have a problem with the video card(s). I was unable to wake it from monitor standby, I shut it down, and powered it back on again, and I got no display. I could tell that Windows was loading, because of the HD LED. I shut it down, and tested it with both DVI ports, and an analog VGA adaptor on both ports. Nothing.

Looks like I cooked my top "Top" video card, in less than two weeks, at 91-96C. Oh well, the thing had a damn inefficient cooler anyways.

 

[Markfw]

Sorry to hear about it. My take is to always use dual-slot cooler solutions and a well ventilated case, and 65% fan speed always on the cards.

 

[Insidious]

I don't have a slice on my drives. This book says.... right here.... that I'm doing it right.

Yeah, I know it looked like my ball curved and it is laying over there in the next fairway, but the book said I am swinginig the right way.

The ball didn't curve dernit

Moral of the story: READ A DIFFERENT BOOK

-Sid

 

[petrusbroder]

The teams' stats as of December 09, 21:30 UTC:

The Folders
ppd __________ points ________ ppd __________ points _________ points _______ points _________ Name
at start _______ at start ______ Dec.09 ________ Dec. 09 ________ in Dec. ______ past 2 days ____________
_____ 0 _____ _ 134 846 ______ 3 560 _______ _ 159 767 _______ 24 921 _________ 5 817 _______ [TA]Assimilator1
__4 271 _____ 1 902 475 ______ 6 785 _______ 1 957 519 _______ 55 044 ________ 13 345 _______ Cutthroat
__5 870 ____ __ 836 014 _____ 12 245 _______ _ 936 460 ______ 100 446 ________ 25 545 _______ Drsignguy
_65 769 ____ 15 196 034 _____ 55 375 ______ 15 712 281 ______ 516 247 _______ 110 834 _______ Gravity
_14 646 _____ 2 137 843 _____ 13 135 _______ 2 262 425 ______ 124 582 ________ 30 147 _______ jonesthewine
__3 855 ____ __ 701 617 ______ 4 555 _______ _ 740 890 _______ 39 273 _________ 8 947 _______ m1ldslide1
_____ 0 ____ __ 629 388 _____ __ 784 _______ _ 634 870 ________ 5 482 ________ ____ 0 _______ paulnt04 = PCTC2
__5 382 _____ 1 730 509 _____ 12 523 _______ 1 832 683 ______ 102 174 ________ 31 171 _______ Pokey_TA
___ 957 _____ __ 12 657 ______ 1 458 _______ __ 21 572 ________ 8 915 _________ 4 339 _______ theAnimal
___ 239 _____ __ 13 862 _____ __ 328 _______ __ 16 161 ________ 2 299 ________ __ 643 _______ TA_andy
__5 510 _____ __ 34 764 _____ 12 662 _______ _ 137 788 ______ 103 024 ________ 27 100 _______ VirtualLarry
105 303 ____ 23 303 490 ____ 123 409 ______ 24 412 416 __ __ 1 082 407 _______ 257 888 _______ Sum for the Folders


TaAm AnandTech Sentinels (TAS)
ppd __________ points ________ ppd __________ points _________ points _______ points _________ Name
at start _______ at start ______ Dec.09 ________ Dec. 09 ________ in Dec. ______ past 2 days _____________
___ 274 _____ __ 33 556 _____ __ 714 _______ __ 43 316 ________ 9 760 _________ 1 920 _______ Baldurs_Gate
_13 522 _____ 1 531 760 _____ 25 097 _______ 1 760 566 ______ 228 806 ________ 51 413 _______ biodoc
_____ 0 _____ ___ ___ 0 ______ 4 248 _______ __ 34 806 _______ 34 806 ________ 10 899 _______ Chocobo_Navy
__2 856 _____ 2 562 004 ______ 3 616 _______ 2 593 015 _______ 31 011 _________ 9 432 _______ GLeeM
__3 874 _____ 2 480 959 _____ 28 366 _______ 6 359 180 ______ 259 249 ________ 61 323 _______ Insidious
_10 971 _____ 2 018 284 _____ 11 715 _______ 2 130 646 ______ 112 362 ________ 25 558 _______ Lightstar
_52 968 ____ 12 408 074 _____ 54 962 ______ 12 894 029 ______ 485 955 _______ 122 137 _______ Markfw900
__3 566 ____ __ 943 946 ______ 4 242 _______ _ 982 181 _______ 38 235 _________ 8 076 _______ Neurodog
_____ 0 ____ ___ 25 758 ______ 1 349 _______ __ 38 878 _______ 13 120 _________ 3 840 _______ rabrittain
___ 949 ____ __ 446 828 ______ 1 322 _______ _ 457 812 _______ 10 984 _________ 4 484 _______ TAS-petrusbroder
__6 773 _____ 2 082 441 ______ 7 564 _______ 2 147 651 _______ 65 210 ________ 15 964 _______ TurtleBlue
__1 728 ____ __ 482 339 ______ 1 820 _______ _ 497 414 _______ 15 075 _________ 2 966 _______ Tom_Philippart
126 138 ____ 31 115 880 ____ 148 367 ______ 32 448 918 __ __ 1 333 038 _______ 323 265 _______ Sum for TAS


Difference between the Racing teams:
Day 1 __________ 62 029 points ________ TAS wins this day!
Day 2 __________ 10 423 points ________ TAS wins this day!
Day 3 __________ 46 068 points ________ TAS wins this day!
Day 4 __________ 23 849 points ________ TAS wins this day!
Day 5+6 ________ 42 885 points ________ TAS wins this day!
Day 7+8 ________ 65 377 points ________ TAS wins this day!
Difference: _____ 250 631 points ________ TAS in the lead.

In the past 2 days the two teams produced 581 153 points! Well done!
In December, the two teams produced 2 415 445 points!

 

[Insidious]

Thanks for the stats Peter!

I think it's awsome that our merry little bunch has processed two and a half million points in this venture.

KUDOS

:beer:

-Sid

PS: TAS rules!

 

[Markfw]

Thanks for the stats Peter !

My new cruncher is down again, 12000 PPD down until I can reformat. The bad PSU (new one coming thursday) keeps locking up the system, and the repair install locked up, and now it won't repair, have to start from scratch.