• Some users have recently had their accounts hijacked. It seems that the now defunct EVGA forums might have compromised your password there and seems many are using the same PW here. We would suggest you UPDATE YOUR PASSWORD and TURN ON 2FA for your account here to further secure it. None of the compromised accounts had 2FA turned on.
    Once you have enabled 2FA, your account will be updated soon to show a badge, letting other members know that you use 2FA to protect your account. This should be beneficial for everyone that uses FSFT.

Help a n00b

Xilikon

Xilikon

[H]ard|DCer of the Year 2008
2FA
Joined
Oct 12, 2004
Messages
15,022
Hi [H]orde,

I have heard a lot about your kindness to help each other when it come to folding so let me tap this for my own good ;)

I have a i7 920 with a GTX 260 (both under water), a Celeron E1200 WHS and a E5200 HTPC. I wanted to get the most juice out of them with the new advances in F@H warfare so suggest me how I should setup all this to get the most PPD out of them.

Regards,
Stef
 
n00b? Crap, you are oldie than Oldie. :p

Welcome back brother!
 
Nice to see you around. I heard you were folding somewhere else?

What is going on with you?
 
I may be older than Oldie but still a n00b when it come to the newest stuff like BigWU, GPU3 and SMP2 so my request is genuine... I phrased it as a poke to the [H]orde :p

Pocatello, I never folded for any other team. I just stopped folding so I can run WGC to help a new team at another forum get on the track. Now that I find WCG a bit lacking in terms of crunching power usage (no real GPU client beside some buggy and obscure ones), I'm getting back to F@H and helping team 33 extend the deadline a bit before getting dethroned out of the #1. The urge came when MtM poked me to make some comments about my official installation guides so thank him for that :)

I'm glad to see familiar faces around.
 
Xilikon:

Crack open a cold beer
Overclock that 920 as fast as you can
Install the new SMP client per Zero's sticky at the top of the forum being sure to add the -bigadv flag
Stick with the GPU2 client on the GTX 260
Have another cold beer (plenty around for n00bs)
Install the new SMP client on the E5200 Dual Core
Install the classic client (or SMP client, just don't use the -smp flag) on the Celery
Install HFM.net to monitor it all
Have another cold beer and some tamales
Sit back and watch the PPD roll in

:cool:
 
And when do I eat some tamales between those steps ?

Thanks, that's what I wanted to know. I'm unsure whenever to use GPU3 or not since it's unclear for which card it benefit the most. I won't upgrade for now because I just moved to a new house since 2 weeks with a few thousands lined up for a few additions like a fence and a deck so my dog could get outside in the rear instead of getting out from the front. However, I might slide a upgrade or two if the price is right later.
 
on the gtx 260 do not use the gpu3 client.. use the gpu2 client.. theres a HUGE difference in PPD with the g200 cards on GPU2 and GPU3..you lose between 3-4k PPD on the GPU3 client..
 
Xilikon said:
And when do I eat some tamales between those steps ?
Click to expand...
Tamales are best savored after all the work is done and I think Apollo cooked the last batch so you'll need a few beers in you first. :eek:

GPU3 is generally stable. However, in my testing, some of the Core 15 WUs for non-Fermi cards run a bit slower on G200 cards compared to GPU2 WUs. Interestingly enough, when you compare G92 Core15 to G92 Core11 WUs you don't notice any point difference at all. I definitely notice it on my G275's though. So, to maximize PPD, the point of your post, stick with GPU2 for now on G200 hardware.
 
i'm having the burrito before i get the system going ;) i guess i'm breaking the order


(substituting Burrito for Tamale, si nce well i have a burrito and not a tamale)
 
I would look into a Corsair H50/70 to get that bad boy to 4ghz+ @ reasonable temps and you should be able to crush the -bigadv flag that way. My understanding from other 920 folders is 4ghz gets the WU completed with time to spare to give you a nice PPD avg.

with the new 6.30 windows client -bigadv is even easier to setup than ever IMO. I'm a Linux guy and it was easy for me so that should say something :)
 
Mtnduey said:
I would look into a Corsair H50/70 to get that bad boy to 4ghz+ @ reasonable temps and you should be able to crush the -bigadv flag that way. My understanding from other 920 folders is 4ghz gets the WU completed with time to spare to give you a nice PPD avg.

with the new 6.30 windows client -bigadv is even easier to setup than ever IMO. I'm a Linux guy and it was easy for me so that should say something :)
Click to expand...

Well, if you consider a watercooling setup with a Heatkiller 3.0 CU nickel-plated, BlackIce SR-1 120.3 + Swiftect MCR120, Dual MCP355 with a XSPC Dual-Bay Reservoir, D-Tek Fuzion GPU2 with UNISink and MIPS X58 SLI LE motherboard block set good enough, then I'm set (hint, check my sig) ;)

I know -bigadv is easy, I'm just unsure which client to use as I was off the folding scene for over 1 1/2 years :eek:
 
Xilikon, welcome back!! :cool: :cool:

Was wondering for a long time when you'd be coming back. :)
 
APOLLO said:
Was wondering for a long time when you'd be coming back. :)
Click to expand...
The folding Gods were listening to me last night. I was just writing about him on my belated HDCOTM interview questions that I finally gave Poc'y last night.
 
I switched to the GPU3 client on my GTX 295s, simply because I moved them to a machine with GPU3 already configured. However, I'm using the "-forcegpu nvidia_g80" command line option to receive (mostly) core 11 WUs. I haven't noticed any PPD difference, so it seems like there's no reason to avoid GPU3.
 
SazanEyes said:
However, I'm using the "-forcegpu nvidia_g80" command line option to receive (mostly) core 11 WUs. I haven't noticed any PPD difference, so it seems like there's no reason to avoid GPU3.
Click to expand...
If you are receiving mostly Core 11 WUs it is likely because you are not running with -advmethods turned on. Most of the non-Fermi Core 15 WUs are still in -advmethods but many should start being released to the public (non -advmethods) shortly. If you enable -advmethods, I bet you'd see a greater assignment of Core 15 WUs.
 
Mtnduey said:
I'm a Linux guy and it was easy for me so that should say something :)
Click to expand...
I thought you were an "AIXPERT"? :p Man, that brought back memories.. I did a fair amount of work a long time ago on PowerPC (some Power1's) based RS/6000's.. *shudder* I was so glad when I moved to a Sun shop. :p
 
Tobit said:
I thought you were an "AIXPERT"? :p Man, that brought back memories.. I did a fair amount of work a long time ago on PowerPC (some Power1's) based RS/6000's.. *shudder* I was so glad when I moved to a Sun shop. :p
Click to expand...

I am the AIXPERT :D But like 1 out of 100 geeks even know wtf AIX is so I just tell people I'm a Linux guy. If I can get some time I wanna give the PowerVM Lx86 a try and see if I can get those Power 7 CPU's to fold. 3.7ghz, 4 threads per core, 8 cores of joy per cpu, oh yeah
 
Mtnduey said:
If I can get some time I wanna give the PowerVM Lx86 a try and see if I can get those Power 7 CPU's to fold. 3.7ghz, 4 threads per core, 8 cores of joy per cpu, oh yeah
Click to expand...
Funny, we were just talking about this on irc this morning. :D

[11:00] <&rmdashrrootsplat> I wish there was a client for IBM Power... Got a p570 with 12 free Power7 cpu's here at work. :p
[11:01] <&rmdashrrootsplat> I could erm... break it in... :D
[11:01] <&Tobit> hmm, install linux and wine on that bad boy.
[11:04] <&rmdashrrootsplat> Don't think I could get permission tho... there is a production lpar running on it. :p

:cool:
 
Tobit said:
If you are receiving mostly Core 11 WUs it is likely because you are not running with -advmethods turned on. Most of the non-Fermi Core 15 WUs are still in -advmethods but many should start being released to the public (non -advmethods) shortly. If you enable -advmethods, I bet you'd see a greater assignment of Core 15 WUs.
Click to expand...

So what is recommended? I'm not running -advmethods (and in general never do) but I'm willing to turn it on.
 
SazanEyes said:
So what is recommended? I'm not running -advmethods (and in general never do) but I'm willing to turn it on.
Click to expand...
Well, since we are in maximize PPD mode right now team-wide.. don't fix what isn't broken. If you are getting mostly Core 11's awesome.. don't change a thing. :cool:
 
Thanks for the warm welcome and for the advices :D

Right now, I'm working to finish the networking backbone in my new house (ie, installing a 24 ports patch panel and configure the server and switches). After that, I will start unpacking and connecting the computers to run them. That's because I'm moved to a new house since less than 3 weeks and it's a real mess.
 
A last n00b question...

Are the new 6.30 SMP client insensitive to network hiccups ? I assume it's the case since the MPICH is the one causing issues due to networking.
 
Xilikon said:
A last n00b question...

Are the new 6.30 SMP client insensitive to network hiccups ? I assume it's the case since the MPICH is the one causing issues due to networking.
Click to expand...

I haven't had any issues with network problems for a long time.
 
Thanks Kendrak !

I have a laptop at work which is put aside for a few weeks and it's brand new. I thought about folding on it using a DSL connection and wifi (we use it for testing purpose and isn't monitored). In the past, it's impossible to fold since wifi would often drop causing the client to lose the WU.
 
Xilikon said:
Thanks Kendrak !

I have a laptop at work which is put aside for a few weeks and it's brand new. I thought about folding on it using a DSL connection and wifi (we use it for testing purpose and isn't monitored). In the past, it's impossible to fold since wifi would often drop causing the client to lose the WU.
Click to expand...

Not sure on that......

All my boxen that fold are hard wired.

I have 4 different spots in the house you can plug in and get 1000/100 based. Never had to use wifi but for netbook and my phone.
 
I've run the GPU client on a wifi box with no issues, even when the connection drops and then reconnects. Never tried it with SMP.
 
I wouldnt run wireless on a bigadv client but others should be fine.
 
Not bigadv for sure, the laptop doesn't have a Core i5/i7 in it, just standard SMP.

So far, it went to 17% and chugging along on a 6060. I now think it should work fine :)
 
You must log in or register to reply here.
Back
Top