• Welcome to TechPowerUp Forums, Guest! Please check out our forum guidelines for info related to our community.

WCG-TPU Cruncher's Hardware/Tech Support Discussion Thread

Attachments

  • boinc 3.png
    boinc 3.png
    41.1 KB · Views: 379
Hey guys! Been crunching the past few days but all of a sudden today I'm getting BSODs (first MEMORY_MANAGEMENT MANAGEMENT and then now PNF_LIST CORRUPTED or something similar. Rig now reboots every ten minutes roughly. Could this a problem with my settings (I picked 100% cpu time and usage) or something worse? D:
 
Having a guess here, while it indicates it may be memory, it might also be hard drive related, check Page File size and reduce it, also make sure allowable disc usage isn't too big or WCG will max it out.
Hopefully others will have suggestions or a solution.
 
Also you could shut down boinc and run something else that will load the system to rule out Boinc.
 
ThE_MaD_ShOt said:
Also you could shut down boinc and run something else that will load the system to rule out Boinc.
Click to expand...
Will try Prime95 as one of the tasks was running said "avx" in the name and I believe prime's similar instruction set (something like that?

Caring1 said:
Having a guess here, while it indicates it may be memory, it might also be hard drive related, check Page File size and reduce it, also make sure allowable disc usage isn't too big or WCG will max it out.
Hopefully others will have suggestions or a solution.
Click to expand...

I'll try adjusting page file also, the disk settings I've left on the default. WCG shouldn't be maxing it out by default, no?

edit: running prime95 blend i got ATTEMPTED_TO_WRITE_TO_READONLY_MEMORY this time!

edit 2: got a DRIVER_IRQL_NOT_LESS_OR_EQUAL when i click run on Prime95 the second time, then a MEMORY_MANAGEMENT BSOD just browsing the internet straight after that, so at least it's not just BOINC!

edit 3: DRIVER_OVERRAN_STACK_BUFFER? haha I have no idea what's going. I got that one running Prime when I tried to run CPU Z (coincidence?) running Windows Memory Diagnostic now, and probably gonna run Memtest86 after that too
 
Last edited:
Sounds to me like you have some bad memory in that system @El_Mayo .

Download and boot using Memtest and see if it throws errors.
 
stinger608 said:
Sounds to me like you have some bad memory in that system @El_Mayo .

Download and boot using Memtest and see if it throws errors.
Click to expand...
around 5600+ so far 1 hour in... I don't know how my memory's gone bad all of a sudden. Will test the sticks individually without xmp speeds after a cmos reset now. Is it safe to turn off memtest prematurely?


:(
 
thebluebumblebee said:
Yes.
Click to expand...
I think I've identified the stick that was causing problems, one of them is showing (around 1100+) errors within the first 4 tests while the other isn't. I will run a full pass on he "good" stick just to be safe. Hopefully all good, I can live with 8GB

edit: GeiL supposedly have a lifetime warranty so i've filled out a request. They're based in Taiwan so part of me is expecting this to be a logistical nightmare! (if they respond)
back on topic if this stick passes memtest in single channel I'll be back crunching tonight! :D
 
Last edited:
Haha, that is just what I expected you to find @El_Mayo ! I kind of suspected one or both ram sticks to throw up errors in memtest. Glad that is all it was.

And I have heard that GeiL has a US based RMA center and from past research have heard they are pretty easy to deal with when RMA's are concerned. Of course your experience may be different. :p
 
stinger608 said:
Haha, that is just what I expected you to find @El_Mayo ! I kind of suspected one or both ram sticks to throw up errors in memtest. Glad that is all it was.

And I have heard that GeiL has a US based RMA center and from past research have heard they are pretty easy to deal with when RMA's are concerned. Of course your experience may be different. :p
Click to expand...

Haha well I bought it second hand from eBay from a guy who only had it for testing so it was practically new. It definitely passed memtest when I bought it cos it's the first thing I did. I'm hoping that lifetime warranty transfers between owners, but I've left out the second hand part in my RMA request anyway. Shipping to Taiwan can't be that expensive tho? this ram stick could literally go in a letter envelope with some bubble wrap. either way I'm back up and running, glad it's just the one stick!
 
I've done an RMA of Ram to Taiwan directly, from Australia, it wasn't expensive and they dealt with it professionally and fairly fast, even though I wasn't the original owner as I had bought from eBay also. I can't remember if it was Geil though, but I do use that brand also.
 
Caring1 said:
I've done an RMA of Ram to Taiwan directly, from Australia, it wasn't expensive and they dealt with it professionally and fairly fast, even though I wasn't the original owner as I had bought from eBay also. I can't remember if it was Geil though, but I do use that brand also.
Click to expand...
Finally heard back from them! :D
On an unrelated note though I haven't had any tasks on BOINC for about 2-3 days now. This ever happened to anyone? I tried closing and reopening BOINC and reseting a few times but nothing
 
In advanced view, make sure you haven't accidently hit "no new tasks" under the projects tab.
 
Caring1 said:
In advanced view, make sure you haven't accidently hit "no new tasks" under the projects tab.
Click to expand...

I don't think so, but everything (Update down to Properties) is greyed out so maybe I have? D:

edit: never mind, I hadn't clicked on WCG first. I didn't click no new tasks, and I've clicked Update now but I'm getting communication deferred
 
El_Mayo said:
I don't think so, but everything (Update down to Properties) is greyed out so maybe I have? D:

edit: never mind, I hadn't clicked on WCG first. I didn't click no new tasks, and I've clicked Update now but I'm getting communication deferred
Click to expand...

Are you getting new tasks now?
 
manofthem said:
Are you getting new tasks now?
Click to expand...

Not as of this morning, but I just read the thread you mentioned me in. I will cancel the .png transfers that are stalling in the transfers tab and try again
 
Didn't even have to cancel those... it just started working today all of a sudden... now I missed out on 3 days for nothing!
 
Been having bsod crashes running mapping cancer markers. One time at around 1% while the other I'm almost done completing the task. Error code has something to do with whea something. System was stable on Prime 95 (blend and small ffts) 12 hours each test. Specs on my system specs profile. What you guys think?
 
Whea uncorrectable error is usually a CPU issue, and as they are CPU intensive it could be that, or a PSU issue, are the VRM's running hot?
Run HWInfo and have a look at the board temps.
 
Caring1 said:
Whea uncorrectable error is usually a CPU issue, and as they are CPU intensive it could be that, or a PSU issue, are the VRM's running hot?
Run HWInfo and have a look at the board temps.
Click to expand...

The 960t is unlocked btw. It's running on 5 cores with an overclocked freq. of 3.8ghz. Dunno which part here is the vrm but the highest temp that I've noticed is the T2 sensor at 62c.

hwinfo.png


I just find it odd that it can withstand prime95 fine but when it comes to crunching it goes nuts. Isn't Prime95 supposedly more demanding than boinc? My cpu temps are even significantly higher in Prime95.
 
Your temps look fine, it might be related to the unlocked core not being as stable as the others.
 
Caring1 said:
Your temps look fine, it might be related to the unlocked core not being as stable as the others.
Click to expand...

currently crunching with just the default cores, no unlock. Running good so far.
 
@brandonwh64 you'll love this bud. Girlfriends hdd stop responding and I can't figure out why.
20160309-212043-large.jpg
 
You must log in or register to reply here.
Back
Top