CUDA has been crashing my GPU (I assume due to overheating) so I de-selected it as an option. Does the use x% of processor setting affect only the CPUs or the GPU too? I don't have any intentions of adding fans etc to my system to otherwise rectify this.

---

CUDA has been crashing my GPU (I assume due to overheating) so I de-selected it as an option. Does the use x% of processor setting affect only the CPUs or the GPU too? I don't have any intentions of adding fans etc to my system to otherwise rectify this.

Another good Cuda query......can Cuda usage be throttled?

---

"Time is simply the mechanism that keeps everything from happening all at once."

Can CUDA be throttled ? ..........
Hmmm, brings `something` to mind.........
.
Curiously Undeveloped Dubious Application :)

---

Can CUDA be throttled ? ..........
Hmmm, brings `something` to mind.........
.
Curiously Undeveloped Dubious Application :)

I was talking about the app, not the devs...........although...never mind.

---

"Time is simply the mechanism that keeps everything from happening all at once."

I too started processing CUDA units and began getting BSOD. I've never had those before, even in hard gaming. In analyzing the dump files and checking with nVidia, it does indeed look like my card too was overheating. Looks like CUDA was driving my card harder than any game ever has. But, my card is also factory 'superclocked', so it probably ran hotter anyway. Oh well, another excuse to get me one of the new GTX 295 cards. For those having the same problem, I've seen in other forums that you can tweak down the speed of the GPU to reduce heat, and of course, make sure your case is well ventilated as well. Also, if you're running any of the newer GTX 200 series, you need to be sure to have not only enough wattage on your power supply, but also enough amps available on the +12v rails. Many power supplies have lots of watts, but can't supply the ampage on the rails, and that is very often the source of stability issues when driving the nVidia processors hard. For example, the GTX 295 requires that the power supply be able to provide at least 46amps +12v, and few power supplies can do that. If you run high end cards with wimpy power supplies, you'll have problems.

---

I too started processing CUDA units and began getting BSOD. I've never had those before, even in hard gaming. In analyzing the dump files and checking with nVidia, it does indeed look like my card too was overheating. Looks like CUDA was driving my card harder than any game ever has. But, my card is also factory 'superclocked', so it probably ran hotter anyway. Oh well, another excuse to get me one of the new GTX 295 cards. For those having the same problem, I've seen in other forums that you can tweak down the speed of the GPU to reduce heat, and of course, make sure your case is well ventilated as well. Also, if you're running any of the newer GTX 200 series, you need to be sure to have not only enough wattage on your power supply, but also enough amps available on the +12v rails. Many power supplies have lots of watts, but can't supply the ampage on the rails, and that is very often the source of stability issues when driving the nVidia processors hard. For example, the GTX 295 requires that the power supply be able to provide at least 46amps +12v, and few power supplies can do that. If you run high end cards with wimpy power supplies, you'll have problems.

Stock setups do not consider 100% load 100% of the time as 'normal'.....
They are not engineered for that..........

So continues the sad saga of the lonely Cuda......

---

"Time is simply the mechanism that keeps everything from happening all at once."

CUDA has been crashing my GPU (I assume due to overheating) so I de-selected it as an option. Does the use x% of processor setting affect only the CPUs or the GPU too? I don't have any intentions of adding fans etc to my system to otherwise rectify this.

Another good Cuda query......can Cuda usage be throttled?

The short answer is No. Not in the current version of BOINC. In fact it won't even task switch them, they run until they finish (one way or the other) and then it will release the gpu for another task.

It probably won't be in the next version of BOINC either, but if you don't ask you don't get. Its a good idea so i'll put the suggestion in and we will see.

---

BOINC blog

but if you don't ask you don't get. Its a good idea so i'll put the suggestion in and we will see.

Added trak ticket 822 with the suggestion.

---

BOINC blog

CUDA has been crashing my GPU (I assume due to overheating) so I de-selected it as an option. Does the use x% of processor setting affect only the CPUs or the GPU too? I don't have any intentions of adding fans etc to my system to otherwise rectify this.

Just try to underclock GPU if you think overheating is the reason (I doubt in that )

For now common reason for driver restart/BSoD on Vista and BSoD on XP is VLAR task. And probably it caused BSOD by kernel that occupies GPU more than few seconds - known CUDA/videodriver issue - check nVidia CUDA page for this limitation.

CUDA has been crashing my GPU (I assume due to overheating) so I de-selected it as an option. Does the use x% of processor setting affect only the CPUs or the GPU too? I don't have any intentions of adding fans etc to my system to otherwise rectify this.

Just try to underclock GPU if you think overheating is the reason (I doubt in that )

For now common reason for driver restart/BSoD on Vista and BSoD on XP is VLAR task. And probably it caused BSOD by kernel that occupies GPU more than few seconds - known CUDA/videodriver issue - check nVidia CUDA page for this limitation.

Rather than make the statement thet 'Cuda is Death'......I suppose I should wish you all well........tune your Cuda, LOL...

---

"Time is simply the mechanism that keeps everything from happening all at once."

CUDA has been crashing my GPU (I assume due to overheating) so I de-selected it as an option. Does the use x% of processor setting affect only the CPUs or the GPU too? I don't have any intentions of adding fans etc to my system to otherwise rectify this.

Just try to underclock GPU if you think overheating is the reason (I doubt in that )

For now common reason for driver restart/BSoD on Vista and BSoD on XP is VLAR task. And probably it caused BSOD by kernel that occupies GPU more than few seconds - known CUDA/videodriver issue - check nVidia CUDA page for this limitation.

Rather than make the statement thet 'Cuda is Death'......I suppose I should wish you all well........tune your Cuda, LOL...

WTF, Mark? CUDA is just new technology and sooner or later it will be well used in SETI and not only in SETI as any another computational technology. Why so negative ???
Maybe it's time to stop to blame and to try to use it ?
Debugging in progress indeed. One of overflow bugs already fixed.

CUDA has been crashing my GPU (I assume due to overheating) so I de-selected it as an option. Does the use x% of processor setting affect only the CPUs or the GPU too? I don't have any intentions of adding fans etc to my system to otherwise rectify this.

Just try to underclock GPU if you think overheating is the reason (I doubt in that )

For now common reason for driver restart/BSoD on Vista and BSoD on XP is VLAR task. And probably it caused BSOD by kernel that occupies GPU more than few seconds - known CUDA/videodriver issue - check nVidia CUDA page for this limitation.

Rather than make the statement thet 'Cuda is Death'......I suppose I should wish you all well........tune your Cuda, LOL...

WTF, Mark? CUDA is just new technology and sooner or later it will be well used in SETI and not only in SETI as any another computational technology. Why so negative ???
Maybe it's time to stop to blame and to try to use it ?
Debugging in progress indeed. One of overflow bugs already fixed.

Oh, you detected my negativity? No problem with you, my friend. The biggest bitch I have is that this crap still belongs in BETA.....NOT HERE IN MAIN!!!!!!!!! AM I CLEAR ON MY POINT?????? Rant off........ It never should have seen the f-in light of day on main.........period.....exclamation point....end of line. Have fun with it......but don't waste my time or the project's time until it is producing valid science.........until then, it is a toy.......not a tool. Advances in technology are all well and good.......if the f'in wheels don't fall off of it when ya roll it out..........seems like Toy Story Bloopers......

---

"Time is simply the mechanism that keeps everything from happening all at once."

"For now common reason for driver restart/BSoD on Vista and BSoD on XP is VLAR task. And probably it caused BSOD by kernel that occupies GPU more than few seconds - known CUDA/videodriver issue - check nVidia CUDA page for this limitation."

I think you're right on this one, because it hasn't been running that hot (that I can tell). I noticed the newest drivers don't address this- does anybody have a fix?

---

Since your taking requests dealing with cuda here, I would like to request a setting/checkbox or something that I can click on so I wont get paired up with a wingman running cuda. I think people runnig cuda should get paired up together, and those of us running cpu's be paired up with no intermingeling between the 2.

Since your taking requests dealing with cuda here, I would like to request a setting/checkbox or something that I can click on so I wont get paired up with a wingman running cuda. I think people runnig cuda should get paired up together, and those of us running cpu's be paired up with no intermingeling between the 2.

Why do you think so? Do you have any idea why pairing is needed at all?
To ensure result validity. If you propose to pair buggy app with buggy app - you want to get database corruption? Think twice again.

"For now common reason for driver restart/BSoD on Vista and BSoD on XP is VLAR task. And probably it caused BSOD by kernel that occupies GPU more than few seconds - known CUDA/videodriver issue - check nVidia CUDA page for this limitation."

I think you're right on this one, because it hasn't been running that hot (that I can tell). I noticed the newest drivers don't address this- does anybody have a fix?

Fixing in progress.

V4c Raistmer? :) V4b is running good for me now. I'm going to give it a couple of days before I do anything else.

---

PROUD MEMBER OF Team Starfire World BOINC

Since your taking requests dealing with cuda here, I would like to request a setting/checkbox or something that I can click on so I wont get paired up with a wingman running cuda. I think people runnig cuda should get paired up together, and those of us running cpu's be paired up with no intermingeling between the 2.

For now, the preference setting to only accept Astropulse work would suit you.

I distinctly disagree with any suggestion that CUDA should be paired with CUDA now, that would put too many doubtful results in the master science database. Once the bugs are fixed I doubt many will care what pairings happen, but more flexibility would at least make sense.

                                                                  Joe

I distinctly disagree with any suggestion that CUDA should be paired with CUDA now, that would put too many doubtful results in the master science database. Once the bugs are fixed I doubt many will care what pairings happen, but more flexibility would at least make sense.

                                                                  Joe

[/quote]

Another point, if cuda is giving so many "doubtful" results as you had said, why is it being used in the mainstream crunching and not back in beta testing where it should remain until it works? Why cant I also crunch seti on my TI-84 calculator as well; seeing as the quality of results obviously dont matter to seti.

...
Why cant I also crunch seti on my TI-84 calculator as well; seeing as the quality of results obviously dont matter to seti.

Hey, if somebody wants to re-write the MB code into BASIC so that it can be loaded as an app, go for it. Though I don't know what kind of instruction set you would be stuck with. That Zilog Z80 is a relic. The 83+SE is clocked to 15MHz (I think the 84 is the same). The 83 and 83+ are only 6MHz.

---

Linux laptop:
record uptime: 1511d 20h 19m (ended due to the power brick giving-up)