mersenneforum.org

mersenneforum.org (https://www.mersenneforum.org/index.php)
-   GPU to 72 (https://www.mersenneforum.org/forumdisplay.php?f=95)
-   -   GPU to 72 status... (https://www.mersenneforum.org/showthread.php?t=16263)

Mark Rose 2015-10-16 23:39

More DCTF exponents are needed... Anonymous might grab nearly all of the available ones this weekend if history repeats.

dragonbud20 2015-11-03 08:35

[QUOTE=Mark Rose;412890]More DCTF exponents are needed... Anonymous might grab nearly all of the available ones this weekend if history repeats.[/QUOTE]

Do we still need more DC? I can move around 1000gHzd/d from TF to DC if needed.

LaurV 2015-11-03 09:16

[QUOTE=Mark Rose;412890]More DCTF exponents are needed... Anonymous might grab nearly all of the available ones this weekend if history repeats.[/QUOTE]
I don't think so, as we get to 73 bits, the time gets double now, so he would only get half of the assignments :razz:

Mark Rose 2015-11-03 12:04

When I wrote that two weeks ago the situation was different.

Mark Rose 2015-11-03 12:12

[QUOTE=dragonbud20;414783]Do we still need more DC? I can move around 1000gHzd/d from TF to DC if needed.[/QUOTE]

No, please don't, unless you want to. There is zero need to move resources to DCTF.

I was talking about GPU72 needing more available exponents for DCTF assignment.

Some of us are working on finishing all the DCTF years ahead of time, for fun. It would be better for the project if we were doing LLTF, so I don't encourage anyone to switch from LLTF to DCTF.

LaurV 2015-11-03 12:46

[QUOTE=Mark Rose;414793]When I wrote that two weeks ago the situation was different.[/QUOTE]
:redface:
(didn't see the date, the post that followed confused me, sorry; you were unlucky to be at the beginning of the page this time, hehe)

dragonbud20 2015-11-04 23:59

Honestly I always have this page open in a tab so I quickly forget when everything was posted lol

petrw1 2015-11-06 22:05

[QUOTE=Mark Rose;412890]More DCTF exponents are needed... Anonymous might grab nearly all of the available ones this weekend if history repeats.[/QUOTE]

And again...hey but it's a good thing ANONYMOUS is helping out so much.
Spidey just needs to put in a double shift.

LaurV 2015-11-07 04:46

Chris, please check those assignment limits. That is, 47M and 48M are supposed to be released at 72.

I switched my DCTF preference to "factor to 73". Before it was "to 70". I did not modify this for long time, following your update to feed us with 71 and later with 72, which was working ok for me, so I was lazy to "move my butt" to actualize local work preference. But then I did it yesterday, and now I am getting 48M to 73 (!?!), from which I did already a bucket, before seeing what's going on. But you won't expect me to continue to do so. I modified all local cached work by hand from "71,73" to "71,72". The last bit will "hang" on my assignment, I will try to go through them as they complete to 72 and unreserve, but this is a some work to do, and I can't promise I will really do it.

I think that if I select "73", you should either feed me with 48M to 72 (only! because that is the "release" point), or if you give me 73, then it should be 50M+ exponents. That is how I see it.

Walt 2015-11-07 15:39

Alas. There are people like me who when requesting assignments to a factoring limit, would like to get assignments that go to that factoring limit. I go to 76 on LLTF because it fits what I think I can do best with my machine to help the overall prime95 effort. I wouldn't want the system to derate my requests automatically to 75.

-Walt

Mark Rose 2015-11-07 18:43

[QUOTE=LaurV;415221]Chris, please check those assignment limits.[/QUOTE]

I've noticed this behaviour if I select "Lowest Exponent". If I pick "What makes sense" it works as expected, giving out assignments at 50M+.


All times are UTC. The time now is 23:15.

Powered by vBulletin® Version 3.8.11
Copyright ©2000 - 2021, Jelsoft Enterprises Ltd.