Linux SMP -bigadv woes

[404 - File not found]

yeah 1.25v is great for 3.8 C0.
83C... you might have some trouble later this summer.. Not sure
I haven't played around much, but I'm also at 3.8ghz, at 1.35ish v. max 70C with scythe mugen.

I have some problems too now, at 3.8 with the third ram stick installed, memtest fails, but at stock all sticks were stable for a day, weird stuff. It's like one stick is weaker.

 

[T4rd]

Well it BSOD'd on me again after 45 mins into a -bigadv WU, so I bumped it up another notch to 1.2625v (my mobo goes in .00625v increments) and I'll see how that does. FAH doesn't seem to stress the CPU as much as Prime95 even though it's maxing all my threads out. I'm only getting up to ~76°C while folding. So that's good.

Is there any way to set the process priority inside the Linux VM so I don't have to use WinAFC or manually do it in the task manager? Because both those methods seem unreliable because the next time I sit down and look at it in the task manager, it's back up to "normal" priority and I have to reload the profile in WinAFC to get it back down. I just want to make sure this VM isn't going to steal any clocks away from what little my GPU clients need and hurting their performance.

That is odd about your third RAM stick. Have you tried switching out one of your sticks with that "weaker" stick to see if it fails the test again? Because if it passes after that, it's gotta be your mobo or something in the BIOS messing with it. Nearly every time I have issues with mobos making RAM seem bad, I can fix it with a new BIOS flash, so check that out and maybe see if there's a new BIOS version for you mobo. Just throwing that out there.. .

Thanks a lot for the input, 404.

 

[The_Heretic]

I have some problems too now, at 3.8 with the third ram stick installed, memtest fails, but at stock all sticks were stable for a day, weird stuff. It's like one stick is weaker.

Could this be because the i7 has the memory controller on the o/c'd CPU itself vs the northbridge like "older" systems ?

 

[404 - File not found]

I'm using WinAFC and its working fine for me, idle for the VM, belowNormal for the gpus; I can use the computer at full speed, I just close the gpus and I can use photoshop.
Just to be sure, did you remove the test mode in WinAFC and change the startup delay?

*\vmplayer.exe := ALL [priority=idle]
*\FahCore_11.exe := PAIR1::CPU0 [priority=BelowNormal]

Could this be because the i7 has the memory controller on the o/c'd CPU itself vs the northbridge like "older" systems ?

That does make sense, I'll try adding some juice somewhere

 

[newxperiment]

Is there any way to set the process priority inside the Linux VM so I don't have to use WinAFC or manually do it in the task manager
Thanks a lot for the input, 404.

Sure is
Just throw this into your .vmx configuration file:

priority.ungrabbed = "idle"
priority.grabbed = "idle"

 

[T4rd]

Sure is
Just throw this into your .vmx configuration file:

priority.ungrabbed = "idle"
priority.grabbed = "idle"

Perfect, thanks a lot, newxp! Does it matter where I put it in the file?

Update: had to bump the voltage up to 1.26875v. It crashed again last night. It's gonna be fun monitoring this all day to see what will keep this stable @ 3.8 GHz. I hope I don't have to get up to 1.3 cuz I think that'd make it run way too hot. I'm amazed you keep yours at 1.35v and still keep it at 70°C, 404. What are your ambient temps? All my D keep it pretty warm in my basement, but I'm not sure what exactly my temps are in my case.

 

[newxperiment]

Perfect, thanks a lot, newxp!

Update: had to bump the voltage up to 1.26875v. It crashed again last night. It's gonna be fun monitoring this all day to see what will keep this stable @ 3.8 GHz. I hope I don't have to get up to 1.3 cuz I think that'd make it run way too hot. I'm amazed you keep yours at 1.35v and still keep it at 70°C, 404. What are your ambient temps? All my D keep it pretty warm in my basement, but I'm not sure what exactly my temps are in my case.

I just recently gave up on getting my D0 920 stable above 4Ghz...it just required too much voltage and started to throttle; If I decide to put together a liquid cooling system I'll try again...but for now I settled at 3.92Ghz for the processor and 3.5Ghz for the unCore...

Voltages are set to 1.288v for the vCore and 1.4v for the unCore(qpi)...my particular processor loves unCore voltage, runs warmer than most people's I've read about, and won't stay stable above 4Ghz, crappy luck of the draw I guess!...Currently have prime running and temps are topping out at 78 on the corsair h-50 with a push/pull config.

I guess the advice I am trying to give is try bumping up your uncore voltage, or lower the uncore clock speed- it may be what's causing your rig to crash, and not the vCore voltage...That's what I found was causing my rig to crash ..Hope this helps,

Alex

edit: no it shouldn't matter where you add the priority information in the vmx...I just throw it in at the end so I can find it easily if I want to change it.

 

[404 - File not found]

I'm running it naked, only two low speed fans, mugen is really a great heatsink, third best in the review I found.

Maybe it's your ram the problem? try loosening the ram timings from specs: my bios by default put me up at 7-7-7 when its supposed to run at 9-9-9 1.65v, it was all good at stock, but when overclocked it wasn't good anymore. If your folding is crashing it's either your cpu or ram, you can rule out the ram for now.

Very possible you do need 1.3v, this is C0 after all, just tie-wrap another fan in there

 

[APOLLO]

Sure is
Just throw this into your .vmx configuration file:

priority.ungrabbed = "idle"
priority.grabbed = "idle"

I had tried that but I still find my GPU clients especially with the 353-pt WUs producing several hundred PPD less if I don't isolate them to a core. I'm assuming you still use an app to increase priority to your GPU clients even if you add those lines for the VMs?

 

[404 - File not found]

ok, then use the code I posted, it's working perfect. I'll help you set it up.

 

[Zero82z]

I'm assuming you still use an app to increase priority to your GPU clients even if you add those lines for the VMs?

I do.

 

[T4rd]

Cool, thanks a lot guys. And thanks again Newxp, that priority config works perfect . I'm pretty sure it's stable now @ 3.8 with a 1.26875v vCore. It's been crunching on a -bigadv for almost 6 hours now at 10.6k PPD and temps maxing around 78°C . I hope it doesn't need any more adjustment. If it does, I'll try playing with the uncore as you suggested (everything besides vCore is on Auto right now). Now if only I can find a way to keep my GPUs from downclocking . All will be right if I can fix this last issue.

Here's a little snip of times between steps

[18:56:34] Completed 53304 out of 250000 steps  (21%)
[19:20:45] Completed 55000 out of 250000 steps  (22%)
[19:55:00] Completed 57500 out of 250000 steps  (23%)
[20:29:46] Completed 60000 out of 250000 steps  (24%)
[21:05:35] Completed 62500 out of 250000 steps  (25%)
[21:40:13] Completed 65000 out of 250000 steps  (26%)
[22:14:35] Completed 67500 out of 250000 steps  (27%)
[22:49:10] Completed 70000 out of 250000 steps  (28%)

About 35 mins per step. Not too bad, right?

 

[nitrobass24]

Glad to see you got it working

 

[Zero82z]

About 35 mins per step. Not too bad, right?

That'll get you about 23k PPD.

 

[T4rd]

That'll get you about 23k PPD.

Saweetah. I'll be so glad if this thing stays stable from now on.

For the [H]orde!