I'm going to buy one of these come payday.

http://www.newegg.com/Product/Product.aspx?Item=N82E16813128627

The fact that 1 card works just fine, but 2 cards go berserk make me think its either a driver/software issue or a failing crossfire chipset on the motherboard. I did buy the motherboard used off ebay


I mean I have a third HD 5970 sapphire edition on my MINISETI computer
http://setiathome.berkeley.edu/show_host_detail.php?hostid=7962112

And i'm running that card with a extreme overclock.

904 Core
-725 stock

4280mhz memory
-4000mhz stock

1.175 vcore
1.050 stock

70c full load fans at 100%

On its own dedicated 850w psu

d) could it be an interrupt problem?

Not sure how to I check?

Yeah, I was trying to find that info in the motherboard manual, but it doesn't say anything about it. Only that if you use both PCIe x16 slots, that they run at x8 a piece.

Do you have anything in the PCI slots?
I'll do a new search in a bit, just came back from a long dinner. :)

---

The fact that 1 card works just fine, but 2 cards go berserk make me think its either a driver/software issue or a failing crossfire chipset on the motherboard.

I agree, either that or it's an intermittent hardware failure on PCIe x16 slot 1, but those are so insanely hard to diagnose because they happen at random. And aside from hanging the whole motherboard onto a multimeter, I doubt it's easily figured out.

The crossfire thing is probably easily tested by not adding crossfire, as for BOINC/Seti that's not necessary. BOINC will see all 4 GPUs as independent entities.
Or you can temporarily run with one videocard, run in slot 1 first, then in slot 2. Just to see if it kicks in with just one card.

---

Is my GPU usage suppose to look like this?

I missed answering this one before, sorry for that.
Uhm yes, that's not a problem. The GPUs aren't under constant load, there's times that part of the calculations happen on the CPU. The rest of the times the CPU is handling the picking up of data from the hard drive, translating of that data to kernels the GPU can handle, move that into the GPU's memory, then waiting for it to finish, before moving it back to the computer's memory, translating it back into something the humans can read and storing that on disk.
Although this happens quite fast, it can happen that for a moment or two the GPU has to wait until the new data is in, which then shows as dips in your graph.

---

I'm running 4 separate gpu tasks though, wouldn't each task look different?

UGH, this thing is so confusing

Allright I did a clean driver install and installed 14.9

and this happened

It ran for about 30 seconds followed by a driver crash that took device1-3 but device (0) stayed working.

When the driver crashed it made a 5th gpu task?!?





I went to the bathroom and came back to find this



All 4 gpu's working, I know its a fluke.

Still its so weird.


________________________________

SO MUCH FOR IT WORKING



3/27/2016 6:38:25 PM | SETI@home | Computation for task 12jn10aa.29122.6210.13.40.209.vlar_2 finished
3/27/2016 6:38:25 PM | SETI@home | Starting task 12jn10aa.1477.23384.6.33.26_2
3/27/2016 6:38:28 PM | SETI@home | Started upload of 12jn10aa.29122.6210.13.40.209.vlar_2_0
3/27/2016 6:38:31 PM | SETI@home | Finished upload of 12jn10aa.29122.6210.13.40.209.vlar_2_0
3/27/2016 6:41:21 PM | SETI@home | Aborting task 27se10ab.2228.67.3.30.106_0: exceeded elapsed time limit 1005.91 (32657.53G/32.47G)
3/27/2016 6:41:23 PM | SETI@home | Computation for task 27se10ab.2228.67.3.30.106_0 finished
3/27/2016 6:41:23 PM | SETI@home | Starting task 27se10ab.2228.67.3.30.122_1
3/27/2016 6:41:25 PM | SETI@home | Started upload of 27se10ab.2228.67.3.30.106_0_0
3/27/2016 6:41:29 PM | SETI@home | Finished upload of 27se10ab.2228.67.3.30.106_0_0
3/27/2016 6:47:50 PM | SETI@home | Aborting task 27se10ab.2228.67.3.30.100_0: exceeded elapsed time limit 1005.91 (32657.53G/32.47G)
3/27/2016 6:47:52 PM | SETI@home | Computation for task 27se10ab.2228.67.3.30.100_0 finished
3/27/2016 6:47:52 PM | SETI@home | Starting task 27se10ab.2228.67.3.30.222_0
3/27/2016 6:47:54 PM | SETI@home | Started upload of 27se10ab.2228.67.3.30.100_0_0
3/27/2016 6:47:58 PM | SETI@home | Finished upload of 27se10ab.2228.67.3.30.100_0_0
3/27/2016 6:50:09 PM | SETI@home | Aborting task 27se10ab.2228.67.3.30.175_1: exceeded elapsed time limit 1005.91 (32657.53G/32.47G)
3/27/2016 6:50:10 PM | SETI@home | Computation for task 27se10ab.2228.67.3.30.175_1 finished
3/27/2016 6:50:10 PM | SETI@home | Starting task 27se10ab.2228.67.3.30.167_0
3/27/2016 6:50:12 PM | SETI@home | Started upload of 27se10ab.2228.67.3.30.175_1_0
3/27/2016 6:50:17 PM | SETI@home | Finished upload of 27se10ab.2228.67.3.30.175_1_0
3/27/2016 6:54:57 PM | SETI@home | Aborting task 27se10ab.2228.67.3.30.43_0: exceeded elapsed time limit 1005.91 (32657.53G/32.47G)
3/27/2016 6:54:58 PM | SETI@home | Computation for task 27se10ab.2228.67.3.30.43_0 finished
3/27/2016 6:54:58 PM | SETI@home | Starting task 24mr10ag.9401.3753.11.38.39_2
3/27/2016 6:55:00 PM | SETI@home | Started upload of 27se10ab.2228.67.3.30.43_0_0
3/27/2016 6:55:04 PM | SETI@home | Finished upload of 27se10ab.2228.67.3.30.43_0_0
3/27/2016 6:58:10 PM | SETI@home | Aborting task 27se10ab.2228.67.3.30.122_1: exceeded elapsed time limit 1005.91 (32657.53G/32.47G)
3/27/2016 6:58:12 PM | SETI@home | Computation for task 27se10ab.2228.67.3.30.122_1 finished
3/27/2016 6:58:12 PM | SETI@home | Starting task 27se10ab.2228.67.3.30.223_1
3/27/2016 6:58:14 PM | SETI@home | Started upload of 27se10ab.2228.67.3.30.122_1_0
3/27/2016 6:58:17 PM | SETI@home | Finished upload of 27se10ab.2228.67.3.30.122_1_0
3/27/2016 7:04:39 PM | SETI@home | Aborting task 27se10ab.2228.67.3.30.222_0: exceeded elapsed time limit 1005.91 (32657.53G/32.47G)
3/27/2016 7:04:41 PM | SETI@home | Computation for task 27se10ab.2228.67.3.30.222_0 finished
3/27/2016 7:04:41 PM | SETI@home | Starting task 27se10ab.2228.67.3.30.237_0
3/27/2016 7:04:43 PM | SETI@home | Started upload of 27se10ab.2228.67.3.30.222_0_0
3/27/2016 7:04:47 PM | SETI@home | Finished upload of 27se10ab.2228.67.3.30.222_0_0
3/27/2016 7:06:57 PM | SETI@home | Aborting task 27se10ab.2228.67.3.30.167_0: exceeded elapsed time limit 1005.91 (32657.53G/32.47G)
3/27/2016 7:06:58 PM | SETI@home | Computation for task 27se10ab.2228.67.3.30.167_0 finished
3/27/2016 7:06:58 PM | SETI@home | Starting task 14oc15ac.23489.7020.8.35.148_3
3/27/2016 7:07:01 PM | SETI@home | Started upload of 27se10ab.2228.67.3.30.167_0_0
3/27/2016 7:07:03 PM | SETI@home | Finished upload of 27se10ab.2228.67.3.30.167_0_0
3/27/2016 7:12:12 PM | SETI@home | update requested by user
3/27/2016 7:12:12 PM | SETI@home | Aborting task 24mr10ag.9401.3753.11.38.39_2: exceeded elapsed time limit 1032.73 (33528.25G/32.47G)
3/27/2016 7:12:14 PM | SETI@home | Computation for task 24mr10ag.9401.3753.11.38.39_2 finished
3/27/2016 7:12:14 PM | SETI@home | Starting task 27se10ab.2228.67.3.30.231_0
3/27/2016 7:12:16 PM | SETI@home | Started upload of 24mr10ag.9401.3753.11.38.39_2_0

Quick reminder, are you overclocking these HD5970s? or only that one in the MiniSeti machine?

Edit: yeah, better reboot that system, looks like you've got a bit crossed wrong.

---

No overclock on the quadfire system, I tried overclocking the fx8350 but the power supply trips and clicks on and off from to much power draw.

Rebooted and waiting to see if its still going to keep rejecting the tasks

EDIT
WEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEE I just blue screened

Time to check the dump log.

I'm going to install a different driver

I'm starting to regret buying amd/ati products. Why do I always have problems with them?

I'm starting to regret buying amd/ati products. Why do I always have problems with them?

Oh no, that one is too easy. Not going to react to that one. :)

By the way, what did the BSOD say?

---

Not sure, I have to download a program to check it and I'm just to exhausted. I need to go to sleep soon I have to get up in a few hours for work.

I'll pick back up tomorrow.

http://www.nirsoft.net/utils/blue_screen_view.html

---

Could my PSU be the problem?

I tried enabling OC genie in the bois but the computer black screens under load, so Ive been running the FX-8350 at stock clocks

Its a 900w off brand psu I bought off ebay.

If this is no-brand PSU that really may be the problem.
They lie about the rated W by big margin

According to this graph just one ATI Radeon HD 5970 card makes full system to use 476 W
http://www.tomshardware.com/reviews/radeon-hd-5970,2474-15.html


But lets try some testing:
cc_config.xml

<ignore_ati_dev>N</ignore_ati_dev>
Ignore (don't use) a specific ATI GPU. You can ignore more than one.

N=0 will ignore (for BOINC use) GPU 0 (half of the first card)
Change 0 to 1 2 or 3 to ignore other GPUs

If any combination of 3 GPUs work it should be a weak PSU

If 3 GPUs also fail you may try to ignore any 2 GPUs:

<cc_config>

   <log_flags>
   </log_flags>

   <options>
      <ignore_ati_dev>0</ignore_ati_dev>
      <ignore_ati_dev>2</ignore_ati_dev>
   </options>

</cc_config>

---

 


- ALF - "Find out what you don't do well ..... then don't do it!" :)