*** F@H Tech Assistance Thread ***

[Feanor]

OK everything is up and running.

The linux bigadv client i'm running on 11 threads in a virtual box (8 gb hard disk, 2500 mb ram) is good for almost 44000ppd at 4360ghz. Around 3000 more than my gpu tracker bigadv client at the same speed. Not too bad for an hour or so of work. VirtualBox is the only freeware i found that let me dedicate more than 4 threads to the vm.

Is the -bigbeta flag enough for me to get bigbeta wu's? The client is currently running with -smp 12, which is needed for the client to be able to fully utilize 11 threads (for some weird reason with only -smp or any other number or nothing, it starts 10 threads), -verbosity 9 for the log, and -bigbeta flags. What are your dedicated rigs settings?

[bogmali]

Settings are the same as yours already; -smp -bigbeta -verbosity 9. Unfortunately, I found out that running -bigbeta on virtual has issues. It has to be on native Linux, read more here:

http://foldingforum.org/viewtopic.php?p=190905#p190905

[Feanor]

I'll check out this morning, but for the first 40% of my 6900 wu, the vm was using around 1,5 gig of ram, and that's for linux and folding in total. So i don't see any issues here. And the fact that this wu is folding faster in a vm than in windows tells me the same. I'll see when i get one of those 6903 or 6904 wu if anything goes wrong.

[BUCK NASTY]

If the WU is actually running in a VM and producing more PPD, what's the chance of running a VM with an i7-870. Would it produce more PPD than the 20K I'm getting now?

[theoneandonlymrk]

having some major issues on my main rig still wont start up at all today psu goes instantly off damn ting i just added a 5850 to get more done too, can a hdd not spinning up cause this?

[thebluebumblebee]

Is this the system in your system specs?

[Feanor]

Originally Posted by BUCK NASTY

If the WU is actually running in a VM and producing more PPD, what's the chance of running a VM with an i7-870. Would it produce more PPD than the 20K I'm getting now?

If you're already running native Linux, there's no point talking about vms. A dedicated rig should run Linux, as it's client produce more ppd than the windows client. However, gpu clients are almost out of question, anyone correct me on that one. Been years since i tried setting up linux gpu folding, i just remember i gave up at the time.

If you're running windows and still want to get linux client small bonus, then i tried virtual box 4.1.2. Setting up a vm is pretty easy, you just click the tabs one at a time and choose what you want to allocate to it. I always keep a thread for gpu client and my personnal use of the computer, so it will be 7 for you and 11 for me. My vm (ubuntu 10.10 64 bit and full tilt bigbeta) is running at 1,5-1,6 of ram out of the 2,5gb i gave it, so your guess for how tight you could fit. Set the hard disk to whatever you wish, i left mine at 8 gb and set it to fixed. Network type NAT works ok for me. After the vm is set up, just follow the musky guide at http://hardforum.com/showthread.php?t=1601608. Install linux, updates, copy and paste all the shit, then hfm, and your all set.

Just remember that the MAIN problem with vm's is that if it crashes YOU LOSE EVERYTHING. The only way i found to start and stop my vm is by saving the state when i hit close. This way you can restart your computer, but still it wont protect you from power lost or yourself.

[BUCK NASTY]

Originally Posted by Feanor

If you're already running native Linux, there's no point talking about vms. A dedicated rig should run Linux, as it's client produce more ppd than the windows client. However, gpu clients are almost out of question, anyone correct me on that one. Been years since i tried setting up linux gpu folding, i just remember i gave up at the time.

If you're running windows and still want to get linux client small bonus, then i tried virtual box 4.1.2. Setting up a vm is pretty easy, you just click the tabs one at a time and choose what you want to allocate to it. I always keep a thread for gpu client and my personnal use of the computer, so it will be 7 for you and 11 for me. My vm (ubuntu 10.10 64 bit and full tilt bigbeta) is running at 1,5-1,6 of ram out of the 2,5gb i gave it, so your guess for how tight you could fit. Set the hard disk to whatever you wish, i left mine at 8 gb and set it to fixed. Network type NAT works ok for me. After the vm is set up, just follow the musky guide at http://hardforum.com/showthread.php?t=1601608. Install linux, updates, copy and paste all the shit, then hfm, and your all set.

Just remember that the MAIN problem with vm's is that if it crashes YOU LOSE EVERYTHING. The only way i found to start and stop my vm is by saving the state when i hit close. This way you can restart your computer, but still it wont protect you from power lost or yourself.

I'm running Win7 Ultimate because of the 2x GPU's. I used to run VMWare for the CPU client several years ago. Got 4gb of ram, so I'll try the virtualbox this weekend. Thanks for the info Feanor!

[Feanor]

You're welcome Buck! Let me know how it turns out, i'll be out of game for a while, i just blew my trustworthy x58a-ud3r. ALWAYS check for spills before turning on your watercooled pc...

[bogmali]

Originally Posted by Feanor

I'll be out of game for a while, i just blew my trustworthy x58a-ud3r. ALWAYS check for spills before turning on your watercooled pc...

That sucks bro Any other parts damaged besides the mobo?

[Feanor]

No. I rushed to the store and bought the only 1366 mobo they had (307$ for a p6x58d-e. I almost had a heart attack when he said the price...) then tested my brand new 970. Everything is fine for the cpu.

Opened a cold beer and relaxed at that point...

I didn't test my two gtx 460 as, first i don't want to set up my second loop only to part it out the next day (just bought the mobo for testing purposes, gonna return it in a few days), and second i don't think a blown vreg can get that far as frying my gpus. So all in all i think all is good, but what a great way to transform 140$ into electric and plastic smelling smoke...

I already found another x58a-ud3r, hope the price will stay as low as he is now. Just one more day to wait for the auction to end!

[DRDNA]

lol where the heck did my folding at home button go?

[Feanor]

I just love ebay. One open box gigabyte x58a-ud3r, with all accessories, for 120$, that's what i call a deal! So my new mobo is ordered, should arrive by the end of the week. I'll finally be able to get back to folding full throttle!

My 970 doesn't like the asus p6x58d-e it's in now, as i cannot go as high as in my last x58a-ud3r (the one i blew last week). It just refuse to go beyond 4200.

Can anyone tell me what kind of settings and voltages they use for their max 24/7 overclock on a 970? I'm under water, so cooling is not a problem. It's gonna be the volts that will limit me! I know for sure this chip will do 4,2 @ around 1,375 vcore, but i really want it to be run at the max.

Also, anyone noticed a problem in folding with the huge memory bandwidth drop from bloomfield to gulftown?

[bogmali]

Originally Posted by Feanor

Can anyone tell me what kind of settings and voltages they use for their max 24/7 overclock on a 970?

Mine is @ 3.9 with 1.275v currently stable for 24/7 folding. I know Oily has a couple in his stable

[oily_17]

Originally Posted by bogmali

Mine is @ 3.9 with 1.275v currently stable for 24/7 folding. I know Oily has a couple in his stable

IIRC ... I have both running @ 4.2, but my voltage is higher, I run about 1.36V.Also use watercooling to keep the temps down.

I have also noticed that a native linux system requires slightly more volts than windows.

EDIT:Feanor, I have both my 970's in Gigabyte x58a-ud3r boards, I can check for sure what my settings are in the morning.

[hat]

Hey guys, I've been having this issue for a while now, but I haven't been too bothered by it until now. For some reason, my Server/Cruncher1 rig randomly dies on me. I'll be able to tell by not being able to log in with teamviewer and seeing the client inactive in hfm.net. It's an intermittent issue.

A few noteworthy things is that the E2140 is overclocked to hell, almost a 100% overclock. It's stable in OCCT Linpack for over 8 hours, so I don't think that's it. It does run quite hot though, around 75c at worst, so I wonder if maybe the CPU just gives up after a while.

The GTX260 is also overclocked, at 640 core, 1405 shaders, with the 1.18vgpu mod. I have it running at 70% min fan speed, so it runs cool enough. Doesn't hit 70c. I ran the GTX260 in my main machine for a long time until I got my Radeon 5870, and it never gave me issues at these settings.

[thebluebumblebee]

Originally Posted by hat

Hey guys, I've been having this issue for a while now, but I haven't been too bothered by it until now. For some reason, my Server/Cruncher1 rig randomly dies on me. I'll be able to tell by not being able to log in with teamviewer and seeing the client inactive in hfm.net. It's an intermittent issue.

A few noteworthy things is that the E2140 is overclocked to hell, almost a 100% overclock. It's stable in OCCT Linpack for over 8 hours, so I don't think that's it. It does run quite hot though, around 75c at worst, so I wonder if maybe the CPU just gives up after a while.

The GTX260 is also overclocked, at 640 core, 1405 shaders, with the 1.18vgpu mod. I have it running at 70% min fan speed, so it runs cool enough. Doesn't hit 70c. I ran the GTX260 in my main machine for a long time until I got my Radeon 5870, and it never gave me issues at these settings.

Okay, since no one stepped up to bat, I'll giver 'er a try. You know more about OC'ing than me, so I'm just giving educated guesses.
*Many around here have found that there is a difference between benchmark stable and 24/7 DC stable. Maybe back off on your CPU OC a little. .2-.4 GHz.
*I'd also give your PSU a good look. Check that the 24 pin connector is in good condition. How big is the PSU?
*Check your BIOS for temperature settings. If you're approaching a high temperature setting in BIOS, a spurious high temperature reading could cause an instant re-boot.
*Is there a time of the day that it tends to re-boot? Maybe put a Kill-a-Watt meter on your line to see if you have any poor electrical conditions. NE sells them all the time for $17-18. Also, some libraries have them to check out.
*I don't think the GPU could be causing this, but I will defer to those who Folded on the GTX260.

[hat]

If you look in my system specs, the machine in question (Server/Cruncher1) is listed in detail. The power supply is a Corsair CX600, which is more than enough for that rig. I think the CPU is stable, as I said 8.5 hours of OCCT linpack has been passed, more than once.

I don't think any safety features are kicking in. I usually turn those off anyway, but if I did leave it on by chance, it doesn't reboot, it just dies. Fans spin, lights are on, but the system is unresponsive and there's no video.

It doesn't do it at a certain time of day, it's intermittent.

I already moved the GPU clockspeeds back to stock, with the 70% fan speed modification still there. I wonder if the GPU really is causing it to crap out, it never was a valiant OC'r, in fact it was well below average. Maybe it degraded further.

We'll see what happens, thanks for the reply.

[Feanor]

I've finally received my 2nd x58a-ud3r, and i'm back folding full time. Next step is trying to send back the first one for rma... Hope the pictures i send them doesn't tell the vrm were blown by water and not dust build-up or something else....!

[hat]

Well, since I moved the GPU back to stock, it hasn't died yet. I wonder if maybe it really was the video card crapping out on me... it seems unlikely since it seemed fine in my other rig, but like I said maybe it just degraded even further. It never was a champion OC'r.

[F150_Raptor]

@ Hat, if I was to take a guess, I'd say your power supply is the culprit. It might have enough to power every thing, but when every thing peaks at once it locks up. I had a corsair hx1000 that did the same thing to me. It was the last thing I swapped out of the rig, put it in another and it did it in the other rig too. How old is the psu?

[hat]

Looks like the machine died again, so the video card wasn't it.

The PSU is fairly new, I've only had it for about two months.

[Chicken Patty]

Guys, as I was posting in the F@H Team Thread, can't get neither of my two cards to fold 100%. I already have my desktop extended and still nothing. When I first set it up it was working fine, but now it's not. Any input?

[[H]@RD5TUFF]

Stupid question how does one obtain an SMP passkey, and one of those nifty signatures that display your PPD and such.

[oily_17]

Originally Posted by [H]@RD5TUFF

Stupid question how does one obtain an SMP passkey, and one of those nifty signatures that display your PPD and such.

For your passkey go here -

http://folding.stanford.edu/English/FAQ-passkey


And you can get some info on the sig image here -

http://folding.extremeoverclocking.com/?nav=IMAGES