Just curious: VALIDATE error for a specific WU

DF1DX
DF1DX
Joined: 14 Aug 10
Posts: 108
Credit: 3909226167
RAC: 40211

I changed my host from a

I changed my host from a AMD7950 (too loud) to a 7970 last Friday, since then i have >95% Validate errors on opencl-ati 1.4.

ID 3781127: GHz, Win7-65, GPU 7970 @ 1 GHz, Catalyst 13.1, P55 with PCI-E 2.0 .
2-3 GPU tasks and 1-2 CPU tasks, about 275-290 VA from a 550W powersupply.

Yesterday i reinstalled the whole system (Win7+Boinc) , new ID is 6742381.
The first validate errors are coming again...

Strange.

Holmis
Joined: 4 Jan 05
Posts: 1118
Credit: 1055935564
RAC: 0

I can only see one "Validate

I can only see one "Validate error" on host #6742381. All the other errors are tasks that are deemed invalid when compared to the result from another volunteer.

Is the card overheating? Is it overclocked?

DF1DX
DF1DX
Joined: 14 Aug 10
Posts: 108
Credit: 3909226167
RAC: 40211

Hi Holmis, the GPU-temp is

Hi Holmis,

the GPU-temp is between 68 and 75° C, the same as with my old GTX460 for the last two years. Fan speed @ 47%, GPU utilisation abt. 88% with two tasks.

The 7970 runs at 1GHz, not overclocked.

I will try some Milkyway-WUs later.

Darren
Darren
Joined: 13 Nov 09
Posts: 38
Credit: 171972935
RAC: 0

I start to see validate

I start to see validate errors when the gpu HS get dirty and card starts to run hotter than normal. I also got alot when I added too much gpu for the PSU to deal with.

DF1DX
DF1DX
Joined: 14 Aug 10
Posts: 108
Credit: 3909226167
RAC: 40211

Minor update: My Tasks on

Minor update:

My Tasks on Milkyway:
Status: Alle (261) | In Bearbeitung (0) | Ausstehend (17) | Gültig (236) | Ungültig (8) | Fehler (0)

No dust, gpu-temp is below 70°C, also downclocked the gpu to 925 Mhz(from 1000), tested with 0 to 2 cpu-tasks, but no joy with opencl-ati.

Linux Memtest is ok, stresstest with Prime95 shows no errors, even flashed a new BIOS...

I'm puzzled.

Alex
Alex
Joined: 1 Mar 05
Posts: 451
Credit: 530410084
RAC: 318296

RE: Minor update: Linux

Quote:

Minor update:

Linux Memtest is ok, stresstest with Prime95 shows no errors, even flashed a new BIOS...

I'm puzzled.

There is one think that points to a bad power-supply: after upgrading to HD7970 >95% of the wu's fail.
The new card needs more power. One thing you could do is to exchange the PS between 2 pc's; that's not too much work to do.
If that fails, you could exchange ram-chips.
Has your PS 2 6-pin power-connectors for the graphic card or is it done with these adapters from the older 4-pin disc-drive power cables?
You could install the TRIXX-utility from AMD, this gives information about the card's input voltage, this should not be less than 11.7 volts.

Good luck!

DF1DX
DF1DX
Joined: 14 Aug 10
Posts: 108
Credit: 3909226167
RAC: 40211

Hi Alex, my Enermax

Hi Alex,

my Enermax "EPR525 AW" PS has 3 "6+2" connectors for PCI-E.

The Host draws about 275VA (Conrad Energy Monitor) with Einstein (2GPU+2CPU tasks). Milkyway uses ca. 330 VA with almost no errors, so i think there is enough power+ headroom available.

GPU Voltage is 1.175V.

Thanks for all suggestions. And excuse my english.

Jürgen.

Comment viewing options

Select your preferred way to display the comments and click "Save settings" to activate your changes.