Notice recently that my 460 is down clocking. A reboot usually will fix this, but not this time. Keep getting a Win 7 message stating the driver has stopped and then recovers. Also been getting computation errors on the GPU WU's..should I be concerned? http://setiathome.berkeley.edu/result.php?resultid=2147946714

Thanks
Win 7 64-bit
Intel i7-970

---

Notice recently that my 460 is down clocking. A reboot usually will fix this, but not this time. Keep getting a Win 7 message stating the driver has stopped and then recovers. Also been getting computation errors on the GPU WU's..should I be concerned? http://setiathome.berkeley.edu/result.php?resultid=2147946714

Thanks
Win 7 64-bit
Intel i7-970

What drivers the newer drivers cause this and I had to go to 258 driver some use the 260 driver.

---

Notice recently that my 460 is down clocking. A reboot usually will fix this, but not this time. Keep getting a Win 7 message stating the driver has stopped and then recovers. Also been getting computation errors on the GPU WU's..should I be concerned? http://setiathome.berkeley.edu/result.php?resultid=2147946714

Thanks
Win 7 64-bit
Intel i7-970

You're also producing inconclusive results, you might want to check the GPU temps, clean the heatsink if required, and adjust the GPU fan speed higher, then think about trying different drivers.

There is also a x39e Diagnostics build you could try instead of x38g

Claggy

You might want to grab a copy of a cuda memory tester and let it loose on the card. MemtestG80 and MemtestCL can be found here

I had 2 EVGA Superclocked GTX460's and both failed after 1 year. It seems the memory had died.

---

BOINC blog

Thanks guy's, I will give it a shot after work.

---

Hi all,

I am having similar trouble with my 1 month old 560TI.

I have downloaded & run the memtests. I ran memtestCL with no errors over 50 iteration,s however when I run memtest G80 - yikes - causes the driver to fall over and reset every time, and the result....

Final error count after 50 iterations over 128 MiB of GPU memory: 1809797992 err
ors

So does this mean I have managed to kill my new video card, and do you think I can claim it under warranty seeing as I have been using the card for BOINC rather than anything graphics related?

Cheers

So does this mean I have managed to kill my new video card, and do you think I can claim it under warranty seeing as I have been using the card for BOINC rather than anything graphics related?

Cheers

Absolutely. If it is a dud and still under warranty (minimum 1 year in Australia) then return it to the shop. Under the Trade Practices Act (or whatever its called now) the shop is responsible for replacement at no cost to you. If it has an extended warranty (the EVGA's usually have 3 years) and was purchased after 1/01/2011 then they are also responsible to sort out during its extended warranty period.

Both my GTX460's were used for cuda crunching. Unfortunately I got them before 1/01/2011 so they were in the extended warranty. I have to pay for shipping, which is still cheaper than 2 new cards. Its one reason why I buy EVGA cards, most brands only give you 1 year as required by law.

---

BOINC blog

Sweeeet!!

My concern is when I take my system back to the shop, if they ask me to replicate the fault, I will have to run BOINC, which I assume they can say is an inappropriate use of the video card i.e.:- not a video related app - no?

Don't get me wrong, I will give it a go, but I would rather go in to the battle fully prepared with highest chance of success rather than falling over at the first hurdle.

Cheers.

Can't you just run the memtest G80 again?

---

PROUD MEMBER OF Team Starfire World BOINC

Considering Nvidia made the original CUDA app for SETI, I would think that makes it a legitimate use of the video card.

---

Considering Nvidia made the original CUDA app for SETI, I would think that makes it a legitimate use of the video card.

                                 ---[Snipped]--- 

Final error count after 50 iterations over 128 MiB of GPU memory: 1809797992 errors.

So does this mean I have managed to kill my new video card, and do you think I can claim it under warranty seeing as I have been using the card for BOINC rather than anything graphics related?

Maybe someone can and will test a 560Ti and 55Ti cards and lower the core-
shader and memory clock-speed, if I up the clock of my 470 or 480 above
1.75 GHz, both started making errors, with SETI MB, Einstein or GPUgrid WUs.

All 560Ti GPUs, I've seen, in a canonical-result, were running @ 1.8GHz, >90%
errored out. Mostly with a Result Overflow or -9 error, so I'm curious
if this has something todo with the clock-speed, too low voltage and or bad
or memory running at it's edge, maximum speed or above.
Resulting in too much heat and or oscillation, due to instabillity.
(Also the Mobo, CPU type/speed; Chipset; DDR2/3 speed and PSU,
+ # of +12V rails and Amps.).

Tried to find an error in an 'old' mobo, it turned out to be a burned PCIe lane?!

I'm using the 275.33 driver on the 480 and 280.22(?) on the 470 and haven't
noticed any downclocking, except if nothing is computed besides graphics.
No General Purpose Computing, anyway.
But that is normal behavior, even in a 'game', it keeps max frequency.
WIN XP 32/64BIT; Q6600(x86); QX96509(x64).

---

As it turns out...my PC was Dead after getting back from work. Power supply only lasted 3 months, not sure if I just got a bad one or not. 850Watts should be big enough for a i7-970 with one GTX 460?

---

As it turns out...my PC was Dead after getting back from work. Power supply only lasted 3 months, not sure if I just got a bad one or not. 850Watts should be big enough for a i7-970 with one GTX 460?

Probably a bad one, in 10 years and over 25 computers, I've only seen two PSUs dead. Silent, a blown triac in most cases and sometimes also
capacitors and the rectifier, a diode-bridge.
In a worst case, the PSU of 500Watt, could produce 50Watts heat, efficiency 90% and fully loaded.

I use a 500Watt PSU on my Q6600 + GTX470 rig, doesn't get warm, ~32C.
It uses, according to a Kill-a-Watt, ~ 385Watt.
Doing 2 SETI MB WUs on 470, 4 on Q6600.

The GTX470 is the hottest part with 90C, well the Q6600, also is @ 90C and have been for 18 months.

The GTX480 & X9650(@3.5GHz) uses 410Watt and has a 650Watt PSU, which
doesn't get warm, 29C either.
(In both cases PSU is placed outside, the computer case.
But in some, bad 'cases', the hot air of CPU(s) and GPU(s), can also push
the PSU 'over the edge', i.e. too hot! If PSU is on top of a tower.............

---