Discussion Thread for the Continuous GW Search known as O2MD1 (now O2MDF - GPUs only)
Language
Copyright © 2024 Einstein@Home. All rights reserved.
Copyright © 2024 Einstein@Home. All rights reserved.
1. I wasn’t the first to
)
1. I wasn’t the first to bring it up.
2. it’s not a serious comment, just some tongue in cheek lol.
3. it’s not that deep, fam.
4. as someone who works in Earth Sciences at NASA, I can say definitively that the earth is not flat.
_________________________________________________________________________
Ian&Steve C. wrote:2. it’s
)
Well just saying "Jesus....." isn't religious. I might say "for crying out loud", but I don't want anyone to cry.
My faith in you is restored ;-)
What?
Ah that's why you have that logo. So you're a real astronaut then :-)
If this page takes an hour to load, reduce posts per page to 20 in your settings, then the tinpot 486 Einstein uses can handle it.
I haven’t been. I work in
)
I haven’t been. I work in remote sensing (satellites) .
_________________________________________________________________________
Server status: GW tasks
)
Server status:
GW tasks valid 1,152,567
GW tasks failed 298,298
Presumably this is the GPU RAM problem. Can the server not decide who to hand out which tasks to? The larger tasks should only go to those with bigger cards.
If this page takes an hour to load, reduce posts per page to 20 in your settings, then the tinpot 486 Einstein uses can handle it.
How come my signature has
)
How come my signature has "Idle projects" listed and nobody else's does? How can I remove that from the list? I can only filter by credits or BSrac. The credits is higher than some of my current projects, and the BSrac has a / instead of a number so doesn't filter.
If this page takes an hour to load, reduce posts per page to 20 in your settings, then the tinpot 486 Einstein uses can handle it.
Peter Hucker wrote: Server
)
I've been checking out my older ATI cards just to see what worked, and to refresh my memory. Using the Older Ubuntu 15.04 and the FLGRX Driver I was able to get the 10 year old HD6870 GPU to run the FGRPB1G task Very slowly...too slowly. I decided to try the HD7750 which also has 1 GB of vram, it's actually not that bad considering it's an 8 year old low-end GPU, it took 42 minutes and validated. Then I switched the preferences to GW to see if it would download a task.
This is what I got, https://einsteinathome.org/host/12822598/log
So, apparently as long as you have 2 Gb of vram, it will send you a GW task even though you need more than 3 on most current tasks. It didn't send any to the HD7750 although it seems to be still trying as the numbers in the link have changed a few times.
FYI this is what it looks
)
FYI this is what it looks like for a 8GB nvidia card.
_________________________________________________________________________
and the next
)
and the next request:
cross referencing with actual GPU mem use in nvidia-smi, it looks like that "GPU RAM calculated: min:" value is about how much the memory the WU will use on the GPU (i see about 1GB and 1.8GB used on the GPUs running GW tasks on that system). it doesn't look like it's referencing what the card actually has available at all though. at least not for the nvidia app.
https://einsteinathome.org/host/12803486/log
_________________________________________________________________________
TBar wrote: I've been
)
Nice to know they can and do check the GPU RAM. It just looks like somebody forgot to update the value!
If this page takes an hour to load, reduce posts per page to 20 in your settings, then the tinpot 486 Einstein uses can handle it.
Ian&Steve C. wrote: and the
)
So it's just some kind of internal calculation to do a request to the GPU? And not going to prevent you getting them when it's too big? How come it didn't send any GW to Tbar's HD7750? Something made it realise that card couldn't run it.
If this page takes an hour to load, reduce posts per page to 20 in your settings, then the tinpot 486 Einstein uses can handle it.