kudo motqalden has put on his turbo!! But it wasn't enough for phoenicis high end CC

Created 12 Jul 2019, 15:01:05 UTC
Coprocessors NVIDIA Tesla V100-SXM2-16GB (4095MB) driver: 418.67 OpenCL: 1.2

https://milkyway.cs.rpi.edu/milkyway/show_host_detail.php?hostid=809514
https://milkyway.cs.rpi.edu/milkyway/show_host_detail.php?hostid=809515
https://milkyway.cs.rpi.edu/milkyway/show_host_detail.php?hostid=809516
https://milkyway.cs.rpi.edu/milkyway/show_host_detail.php?hostid=809517

just a few
XS probaby get stomp by him in a few hours :LOL:

pututu, you forgot the Tesla V100 lmao


So my "turbo" was more "NOS" than turbo and my tank is now empty. just guessed lucky on the sprint and had a nice 8 day plus bunker built up
Nothings special to see here.
 
So about twice as fast as a VII - not bad, but considering the cost...
 
So about twice as fast as a VII - not bad, but considering the cost...

Absolutely, I'd never suggest buying them but renting at $0.80 per hour has it's place.

Assuming space in existing rigs, the purchase of 16 x R7s at UK prices = £9600 + electricity to run + dripping sweat without A/C = serious grief from wife

Renting 8 x v100s for a few hours to maybe stir things up during a sprint = priceless
 
phoenicis - would you do me a favor? If you still have access to the Tesla, can you run a few Einstein at 2x to see what's the run time on it? 2 concurrent tasks
 
phoenicis - would you do me a favor? If you still have access to the Tesla, can you run a few Einstein at 2x to see what's the run time on it? 2 concurrent tasks

Totally not worth it. They're complete pants at Einstein with 2 concurrent taking a little over 13 minutes each which is less than half the performance of a R7. I gave 4 per a try but it just doubled the time taken.

nvidia-smi says the V100 is at 100% utilisation but also that the card is only drawing 135W against a max of 300W. In contrast, when I was running MW yesterday the stated power draw was well over 200W.

So I took your words, tried to order one.
They said "you'd need 2 fuken warehouse of chips for 1 card"
:LOL:

Damned inflation :eek: A chip used to be worth something!
 
Last edited:
Thank you, sir! Looks like Einstein is really exclusive for AMD :LOL:

My grandma said "you better eat them chips, son. You looked skinny"
She doesn't know I've been starving for the last 6 months. Damn AMD took all my chips :D
 
Another crappy project with limited hardware support for only Nvidia GPUs (and not any new ones either lol).

FB needs to be re-organized and learn how to properly choose Projects with better hardware support for Sprints. Sorta takes the fun out of FB.

Have fun guys!
 
Another crappy project with limited hardware support for only Nvidia GPUs (and not any new ones either lol).

FB needs to be re-organized and learn how to properly choose Projects with better hardware support for Sprints. Sorta takes the fun out of FB.

Have fun guys!

No kidding. 90% of my GPUs are AMD. So it seems I'm sitting this one out.
 
Yes in Formula-BOINC's defense this is true as Skillz stated. We had a couple good years of projects popping up out of the woodwork that had good support and tons of work. Now, we are stuck with a few good legacy projects and a ton of home-brew projects that seem to be ran off of a home PC with a hobbyist for an admin rather than dedicated paid staff. GPUGrid has outlived itself as it at one time was a pretty solid project. Now they have limited support to work with. And worse... they know they have an abundance of hardware just thrown at them that they can pretty well just do what they want as they have.

FB has always had major flaws. But unfortunately, they now are heavily limited with what they have to work with.
 
Yeah, I'm not even talking workload support. I'm talking more hardware support, as in what you can use to run the project.
Either way, I see all sides have good points. And FB helps some of these little projects who normally wouldn't get this kind of user involvement either.
Just frustrating at times for everyone it seems.
 
Yeah... IIRC YAFU is or was ran off of a laptop. I could be completely wrong or it may have changed but it goes to show that some of this stuff is certainly not meant for the masses to focus a challenge on. It would be great if admins could learn the limitations of their environments before signing off on some of this stuff. FB doesn't ask the projects from what I have seen, so that is on FB IMO. The DC-Vault didn't do it until I came along and made that just part of the practice.
 
Yeah... IIRC YAFU is or was ran off of a laptop. I could be completely wrong or it may have changed but it goes to show that some of this stuff is certainly not meant for the masses to focus a challenge on.
but at least YAFU is transparent on that; rating itself as Alpha-stage to test new BOINC software. The only math project I kind of like a bit ... because of its name ? Or because of its intend to test software ? Don't know ...
 
We are now at 5th in SETI@home sprint. Need all the help we can get in the next 30 hours.

Got some spare GPUs or CPUs? The ETs are hiding somewhere.
 
Yeah, keep those rigs crunching till the last second.
 
Yeah, time for FB to go away if they don't want to fix or get it straightened out. FFS...at least try and show you give a shite about your own website that announces your competitions.

Good luck to all who decide to run this "MOD_PYTHON ERROR" project. I've ran it too many times it seems over the past couple years to be excited about it anymore LOL
 
This should help you folks out:


<app_config>
<app>
<name>MOD_PYTHON_ERROR</name>
<gpu_versions>
<gpu_usage>.05</gpu_usage>
<cpu_usage>.1</cpu_usage>
</gpu_versions>
</app>
<project_max_concurrent>128</project_max_concurrent>
</app_config>
 
This is a much better project than Universe for bunkering (y)

I don't know, seems to me Universe bunkers quite nicely :p Had Uni been picked for the Sprint things would have gotten interesting really quickly.

Of course that's just a guess on my part.
 
Last edited:
Had Uni been picked for the Sprint things would have gotten interesting really quickly
Yeah, we were so busy Wowing and I totally forgot about Uni. Last check all the short tasks are gone :(
How come no wowing from you?;) We miss ya'll.
 
I had 78M ready to drop when the sprint should have started. :) I don't know why I bother, I've never bunkered the right thing yet. We were just trying to get into the points on the Uni marathon and it dawned on me that if I put the lid on it I could do a fairly nice bunker. Then after I loaded up with the shorties they came out with the new ones. He said today he is considering raising the credit 1.5 or 2X. He should go 3X so they are equivalent.
 
would be funny if it starts working just after the sprint is over and whoever happened to guess it wins i guess. here is hoping its either WCG or MOO
i think we did WCG already did we do MOO?
 
Back
Top