i have a galaxy 560ti gc and a 550ti gc

to anyone with nvidia downclocking problems, i have a solution, not for the faint of heart, but a solution none the less.

*all the rest snipped*

Why not just do the easy thing and upgrade to the latest Lunatics build and run whatever driver you want to begin with?

---

Traveling through space at ~67,000mph!

It's a Tagan TG-580 (nominally a 580W PSU).
The card is a Gigabyte factory overclock. 900MHz core, 1GHz memory (being DDR5 that probably gets translated to 4GHz for advertising purposes), 1.8GHz shader. System is a C2D E6600 running at stock with 1x7200 & 1x10,000RPM HDD.
According to my UPS, the load is about 35%.
Which means about 525VA, so probably around 420W (assuming PF of .8). So depending on your system power, an honest 550W PSU would be up to the job.

Thanks! that gives some good clues.

Jason

---

"Living by the wisdom of computer science doesn't sound so bad after all. And unlike most advice, it's backed up by proofs." -- Algorithms to live by: The computer science of human decisions.

i have a galaxy 560ti gc and a 550ti gc

to anyone with nvidia downclocking problems, i have a solution, not for the faint of heart, but a solution none the less.

*all the rest snipped*

Why not just do the easy thing and upgrade to the latest Lunatics build and run whatever driver you want to begin with?

That looks-like what I will do when the new card arrives and just skip any issues.

at least for me, the card downclocks with the latest lunatics. could be wrong. i downgraded last nite to the 267.59 drivers, http://ru.download.nvidia.com/Windows/267.59/267.59_WinXP_Desktop_Release_Notes.pdf

last nite i flashed back to oem bios so i could mess around with the downclocking problem.

i tried the latest nvidia drivers, and i still got downclocking problems with them. so last nite i tried the earliest drivers available for the 560ti, the 267.59 drivers, and i still got downclocking problems.....

back to the modded bios for me.

so i am guessing it's probably a driver thing, but it could be a video card bios thing. video cards now days do have a 2d downclocked mode, and only really go to 3d when sensing a 3d environment, or at least i am guessing thats how it works, i dont really know for sure.

what i do know, is that flashing the bios how i describes DOES solve the problem, i probably flashed the bios 5 times last nite, and about 25 times total so far. anyone know how many times you can safely flash a so called "read only memory" chip? i'm getting scared....

at least for me, the card downclocks with the latest lunatics. could be wrong. i downgraded last nite to the 267.59 drivers, http://ru.download.nvidia.com/Windows/267.59/267.59_WinXP_Desktop_Release_Notes.pdf

last nite i flashed back to oem bios so i could mess around with the downclocking problem.

i tried the latest nvidia drivers, and i still got downclocking problems with them. so last nite i tried the earliest drivers available for the 560ti, the 267.59 drivers, and i still got downclocking problems.....

back to the modded bios for me.

so i am guessing it's probably a driver thing, but it could be a video card bios thing. video cards now days do have a 2d downclocked mode, and only really go to 3d when sensing a 3d environment, or at least i am guessing thats how it works, i dont really know for sure.

what i do know, is that flashing the bios how i describes DOES solve the problem, i probably flashed the bios 5 times last nite, and about 25 times total so far. anyone know how many times you can safely flash a so called "read only memory" chip? i'm getting scared....

You need to report that to Lunatics then because you are the only person it's happening to, or it could be just the 560Ti since not many people are running them it just hasn't become a known yet. It could be something has went amiss with your bios flashing back and forth, hope you are checking hashes.

---

Traveling through space at ~67,000mph!

...You need to report that to Lunatics then because you are the only person it's happening to, or it could be just the 560Ti since not many people are running them it just hasn't become a known yet.

Adding some information on what is known so far:
Some of the newer cards are under scrutiny for some odd behaviour with newer builds specifically made to give more useful diagnostic information.

In some cases it's looking like unexpectedly high power demands under certain conditions, other cases yet to be determined. In all cases fully resolved so far the downclock has shown to be a driver feature protecting against *something*, boincapi non-threadsafe behaviour in older builds, and establishing what that something is in each case is unbroken ground for some cards in particular.

The current wider diagnostic build for those that need it (x39e) is available at: http://lunatics.kwsn.net/12-gpu-crunching/x38g-reports.msg39472.html#msg39472 . It just 'hardens' some code & gives more useful stderr output in the case of a failure.

It's not an installer, but exe only, so anyone that needs this but isn't familiar with the process should sing out for help here.

Isolating, & if possible refining for, the major niggles experienced with some of the newer cards, would lead to an updated wider release (installer update)

Jason

---

"Living by the wisdom of computer science doesn't sound so bad after all. And unlike most advice, it's backed up by proofs." -- Algorithms to live by: The computer science of human decisions.

...You need to report that to Lunatics then because you are the only person it's happening to, or it could be just the 560Ti since not many people are running them it just hasn't become a known yet.

Adding some information on what is known so far:
Some of the newer cards are under scrutiny for some odd behaviour with newer builds specifically made to give more useful diagnostic information.

In some cases it's looking like unexpectedly high power demands under certain conditions, other cases yet to be determined. In all cases fully resolved so far the downclock has shown to be a driver feature protecting against *something*, boincapi non-threadsafe behaviour in older builds, and establishing what that something is in each case is unbroken ground for some cards in particular.

The current wider diagnostic build for those that need it (x39e) is available at: http://lunatics.kwsn.net/12-gpu-crunching/x38g-reports.msg39472.html#msg39472 . It just 'hardens' some code & gives more useful stderr output in the case of a failure.

It's not an installer, but exe only, so anyone that needs this but isn't familiar with the process should sing out for help here.

Isolating, & if possible refining for, the major niggles experienced with some of the newer cards, would lead to an updated wider release (installer update)

Jason

Yep. Jason hammered out that my 750W PSU wasn't enough to run two 560ti's without the occasional bork. Replacing that with a 1000W soon to see what happens.

---

Executive Director GPU Users Group Inc. -
brad@gpuug.org

Jason,
Just for your reference, I have switched to 39e on my 560TI. Also just for reference, it is the Galaxy 560TI GC edition which is factory overclocked to 835 at .975V. I have pushed the overclock to 850 at the same voltage and have been stable with the 275.33 drivers. When I pushed the overclock up without a corresponding increase in voltage is when I would see downclocking. That hasn't happened since 38g came out though.
Sounds like there shouldn't be a problem but I will keep an eye on it.
Thanks
Dan

---

Yep. Jason hammered out that my 750W PSU wasn't enough to run two 560ti's without the occasional bork. Replacing that with a 1000W soon to see what happens.

Same basic issue I believe I'm having with my server box. Had to take out the second gpu because it kept throwing bad results. And this after an RMA and replacement on the first card!

---

Traveling through space at ~67,000mph!

I just triggered the downclocking issue on one of my GTX 560 Ti's :(

It was getting a bit chilly in my room, so with a bit of extra thermal headroom I proceeded to overclock... I took the GPU Core Clock up from 900MHz to 1000MHz (using the same process I use for my GTX 460 in another machine) and then noticed in GPU-Z that the GPU Core Clock for that card was now only 405 MHz. The other card remains unaffected, and even if I set the one I meddled with back to "Factory shipped frequencies" it's still downclocked according to GPU-Z :(

I'm using 270.61 drivers under Windows 7 and the Lunatics thingy.

Help!?

---

Brian.

I just triggered the downclocking issue on one of my GTX 560 Ti's :(

It was getting a bit chilly in my room, so with a bit of extra thermal headroom I proceeded to overclock... I took the GPU Core Clock up from 900MHz to 1000MHz (using the same process I use for my GTX 460 in another machine) and then noticed in GPU-Z that the GPU Core Clock for that card was now only 405 MHz. The other card remains unaffected, and even if I set the one I meddled with back to "Factory shipped frequencies" it's still downclocked according to GPU-Z :(

I'm using 270.61 drivers under Windows 7 and the Lunatics thingy.

Help!?

Quick fix is a restart. Long term I have no clue.

---

Executive Director GPU Users Group Inc. -
brad@gpuug.org

Can I be honest and say I didn't think about restarting until after I'd posted that?! :$

I get so used to leaving my computer running 24/7, and rarely needing to restart I forget all about that option!

I suppose a permanent fix will arrive in a driver update at some point. For the time being I'll leave the OCing alone on those cards.

---

Brian.

Can I be honest and say I didn't think about restarting until after I'd posted that?! :$

I get so used to leaving my computer running 24/7, and rarely needing to restart I forget all about that option!

I suppose a permanent fix will arrive in a driver update at some point. For the time being I'll leave the OCing alone on those cards.

I'd recommend asking Jason about this issue. He's been dealing with it with my 560's for ages and is far smarter than I am.

---

Executive Director GPU Users Group Inc. -
brad@gpuug.org

How many WU's/card are you guys running?

---

Executive Director GPU Users Group Inc. -
brad@gpuug.org

I'm running 2 at a time.

---

Grant
Darwin NT

I'm beginning to think the 460 may still be the best bang-for-the-buck card and the 560 Ti looks better on paper than in practice... at twice the price.

But mine hasn't arrived, so I haven't had the opportunity to screw it up -- yet.

I was really thinking that with the new Lunatics I should be able to get three at a time going, but I haven't seen anyone report success with that.

I'm running 2 at a time with the x39e app from Lunatics. I have the 275.33 driver. I am currently running at 850 MHz core and 2000 MHz memory clock. This is at the factory set 0.975 V. This overclock is all based on what is stable clock frequency by voltage that f_n_t reported.
Units are running at just under 15 minutes each for the normal ARs. Shorties were running as fast as 3 minutes. The RAC for the box has been climbing but may be stabilizing around 17.5k. Now I want to get another card and really crank out the work.
Dan

---

I'm beginning to think the 460 may still be the best bang-for-the-buck card and the 560 Ti looks better on paper than in practice... at twice the price.

But mine hasn't arrived, so I haven't had the opportunity to screw it up -- yet.

I was really thinking that with the new Lunatics I should be able to get three at a time going, but I haven't seen anyone report success with that.

It works just fine.
A friend of mine is running it a couple days without any problems on a tiny 450.

---

With each crime and every kindness we birth our future.

I'm beginning to think the 460 may still be the best bang-for-the-buck card and the 560 Ti looks better on paper than in practice... at twice the price.

But mine hasn't arrived, so I haven't had the opportunity to screw it up -- yet.

I was really thinking that with the new Lunatics I should be able to get three at a time going, but I haven't seen anyone report success with that.

It works just fine.
A friend of mine is running it a couple days without any problems on a tiny 450.

Side semi-related query: Is everyone remembering to turn off web browser GPU acceleration ?

---

"Living by the wisdom of computer science doesn't sound so bad after all. And unlike most advice, it's backed up by proofs." -- Algorithms to live by: The computer science of human decisions.

Side semi-related query: Is everyone remembering to turn off web browser GPU acceleration ?

Thanks for that Jason as it just maybe what gives me the odd errored w/u so I'll see what happens now. (Y)

Cheers.

---