It's not so much the model of GPU that is supported or not, but the minimum amount of memory it needs to have. See http://einsteinathome.org/node/195578 for the BRP CUDA requirements.
I think the projects Collatz and PrimeGrid still support the older less memory laden cards, there could be others. Most newer cards have at least 1gb of onboard memory, any non onboard memory does NOT count! So when laptops 'add' memory, it does not count towards Boinc and gpu crunching.
It's not so much the model of GPU that is supported or not, but the minimum amount of memory it needs to have. See http://einsteinathome.org/node/195578 for the BRP CUDA requirements.
Ageless - your link reads - ca 300 MB of free RAM required on graphics card.
The GPU has 512 MB. That would seem to indicate that the card is acceptable? http://www.geforce.com/hardware/desktop-gpus/geforce-210/specifications
I tried to use this card for a WCG task and it cause all the GPU tasks to run for the hills (Abort). I'm asking here so the active tasks don't all abort when I insert the card.
Secondary question: I didn't get an email notification of your post. Why not? What didn't I do? I get the feeling you need to subscript to the Thread, even when you created the threat. Yes?
I've used some time ago a GT9500 with 512MB (that is older and slower than the 210) and with the less optimized apps for BRP that was in use in that time it took around 7 or 8 hours to crunch them... So your 210 should not have issues with Einstein, but it depends on your cache sizes, the overal performance of the host and what other projects you run...
It helps to know in advanced what is expected, but every host is a different world, and the only way to be absolutely sure of something is just trying it...
AFAIK, there is no automatic subscription to threads not even for your own ones, but Im not sure if there is some optional setting out there for achieve something like that...
I've used some time ago a GT9500 with 512MB (that is older and slower than the 210) and with the less optimized apps for BRP that was in use in that time it took around 7 or 8 hours to crunch them... So your 210 should not have issues with Einstein, but it depends on your cache sizes, the overal performance of the host and what other projects you run...
It helps to know in advanced what is expected, but every host is a different world, and the only way to be absolutely sure of something is just trying it...
AFAIK, there is no automatic subscription to threads not even for your own ones, but Im not sure if there is some optional setting out there for achieve something like that...
After reading your responses, I decided to make my GPUs ATI. Installing a NVIDIA just to fill the GPU slots just didn't make any sense. But, thanks for the responses. You guys are the best.
[Resolved] GeForce 210
)
It's not so much the model of GPU that is supported or not, but the minimum amount of memory it needs to have. See http://einsteinathome.org/node/195578 for the BRP CUDA requirements.
RE: Will E@H support a
)
I think the projects Collatz and PrimeGrid still support the older less memory laden cards, there could be others. Most newer cards have at least 1gb of onboard memory, any non onboard memory does NOT count! So when laptops 'add' memory, it does not count towards Boinc and gpu crunching.
RE: It's not so much the
)
Ageless - your link reads - ca 300 MB of free RAM required on graphics card.
The GPU has 512 MB. That would seem to indicate that the card is acceptable?
http://www.geforce.com/hardware/desktop-gpus/geforce-210/specifications
I tried to use this card for a WCG task and it cause all the GPU tasks to run for the hills (Abort). I'm asking here so the active tasks don't all abort when I insert the card.
Secondary question: I didn't get an email notification of your post. Why not? What didn't I do? I get the feeling you need to subscript to the Thread, even when you created the threat. Yes?
Thanks
I've used some time ago a
)
I've used some time ago a GT9500 with 512MB (that is older and slower than the 210) and with the less optimized apps for BRP that was in use in that time it took around 7 or 8 hours to crunch them... So your 210 should not have issues with Einstein, but it depends on your cache sizes, the overal performance of the host and what other projects you run...
It helps to know in advanced what is expected, but every host is a different world, and the only way to be absolutely sure of something is just trying it...
AFAIK, there is no automatic subscription to threads not even for your own ones, but Im not sure if there is some optional setting out there for achieve something like that...
RE: I've used some time ago
)
After reading your responses, I decided to make my GPUs ATI. Installing a NVIDIA just to fill the GPU slots just didn't make any sense. But, thanks for the responses. You guys are the best.
Oops.
)
Oops.
Marking as [Resolved].
)
Marking as [Resolved].