VijayPande wrote:Are donors still seeing lots of Core15 on Kepler? We made some AS changes some time ago and thought that issue was resolved.
In the last weeks I´ve got mostly P8018.
EDIT: client-type=advanced
The disadvantage is that my computer has a GTX 780 and a GTX 680. As soon as a core 15 and a core 17 work simultaneously the GPU with core 15 changes to idle mode clock rate and the core 15 project runs extremly slow.
Heiko
PS: Some seconds ago the GTX 680 finished a P13001 and picked up a P8018. => GTX 780 with P8018 increased clock rate from idle clock rate back to normal. Weird!
Last edited by heikosch on Wed Apr 09, 2014 8:00 pm, edited 1 time in total.
Hardware configuration: 1) 1090t @3.6ghz, corsair 850w psu, 16gb gskill, asus m4a89td pro mb, asus 780 ti 2) fx-8350 @4ghz, cooler master 1000w psu, 8gb gskill, asus m5a99fx pro mb, two asus 780 ti's 3) am3 x4 @3.4ghz, rosewill 650w psu, asus M5A78L-M/USB3 mb, 8gb gskill, asus 780 ti
VijayPande wrote:Are donors still seeing lots of Core15 on Kepler? We made some AS changes some time ago and thought that issue was resolved.
I am 9 days straight on core 15 wu's, that's ok with me if you need me to run them, but I upgraded to 3 780 ti's for the future of FAH, I think you can use my cards better than that.
I am certain that the PG will do that once the Core_17 work is available. But as has been explained here and elsewhere, there has been some problems. One server that had Core_17 WU's has not been active for several weeks. I don't know why, but possibly its projects are finished or being analyzed to see where they need to go next. A second server had a disk failure. The project manager posted that they hoped to get a replacement project and server to continue its projects back up quickly, but it appears they have run into issues that need to be cleared up before the replacement project can go active for everyone. That leaves one server supplying Core_17 WU's at the moment.
As for the Core_15 work, I am certain the researchers who started projects with that core are happy to get as many units done as possible. Keep in mind that all of those projects started last year or even longer ago. Once finished they can go on to whatever the next phase of their research is.
iMac 2.8 i7 12 GB smp8, Mac Pro 2.8 quad 12 GB smp6
MacBook Pro 2.9 i7 8 GB smp3
VijayPande wrote:Are donors still seeing lots of Core15 on Kepler? We made some AS changes some time ago and thought that issue was resolved.
When I fire up my 2 GTX 670's all I ever get is Core_15's. I usually only ran the GPU's on the weekends/holidays (lower electricity cost) but I've stopped even doing that now. I did fire them up a couple of days ago to see what the status of Core_15 was and only got Core_15's still.
I know it's been said ad nauseam but if I have the option (and there is this option available) of getting 18k PPD for a Core_15 or 70K PPD for a Core_17 I want the 70K PPD. If I'm not getting the 70K PPD that is available then I'd rather not fold with these 2 GPU's. The relative "return on investment" just isn't there with Core_15 anymore.
Just a thought before the ad nauseam "it's for the science" crowed chimes in. If the points don't matter that much because we fold for the science of it then why not just artificially boost the remaining Core_15 WU's points? Then maybe people like me would want to fold them and we could get them (and the science behind them completed faster?
Bleeder wrote:
I know it's been said ad nauseam but if I have the option (and there is this option available) of getting 18k PPD for a Core_15 or 70K PPD for a Core_17 I want the 70K PPD.
Easy enough if you feel that strongly- run the GPUs under Linux.
Bleeder wrote:
I know it's been said ad nauseam but if I have the option (and there is this option available) of getting 18k PPD for a Core_15 or 70K PPD for a Core_17 I want the 70K PPD.
Easy enough if you feel that strongly- run the GPUs under Linux.
Can confirm, running Linux guarantees Core_17 (well, for nvidia at least)
Hardware configuration: Intel i7-4770K @ 4.5 GHz, 16 GB DDR3-2133 Corsair Vengence (black/red), EVGA GTX 760 @ 1200 MHz, on an Asus Maximus VI Hero MB (black/red), in a blacked out Antec P280 Tower, with a Xigmatek Night Hawk (black) HSF, Seasonic 760w Platinum (black case, sleeves, wires), 4 SilenX 120mm Case fans with silicon fan gaskets and silicon mounts (all black), a 512GB Samsung SSD (black), and a 2TB Black Western Digital HD (silver/black).
7im wrote:How does anyone tell if any GPU projects are running when new GPU projects are not announced on this forum, and the new servers do not show up in the Server Status Page?
Only indirectly, extrapolating via the volume of core_15-related posts on this and other fora
Really? You are missing the point of this rhetorical question.
Hardware configuration: 1) 1090t @3.6ghz, corsair 850w psu, 16gb gskill, asus m4a89td pro mb, asus 780 ti 2) fx-8350 @4ghz, cooler master 1000w psu, 8gb gskill, asus m5a99fx pro mb, two asus 780 ti's 3) am3 x4 @3.4ghz, rosewill 650w psu, asus M5A78L-M/USB3 mb, 8gb gskill, asus 780 ti
Hardware configuration: 1) 1090t @3.6ghz, corsair 850w psu, 16gb gskill, asus m4a89td pro mb, asus 780 ti 2) fx-8350 @4ghz, cooler master 1000w psu, 8gb gskill, asus m5a99fx pro mb, two asus 780 ti's 3) am3 x4 @3.4ghz, rosewill 650w psu, asus M5A78L-M/USB3 mb, 8gb gskill, asus 780 ti
When you post about 8018 you MUST specify which GPU you have and which flags are being used. (client-type=xxx). The core_15 WUs are the correct assignments for certain GPUs which currently cannot fold Core_17 assignments.
Windows 8
V7.4.4
client-type is advanced
GTX 660 Ti (Kepler GK104) -> Got 8018 (FahCore_15) @ 2014-04-09T12:20:19Z
GTX 650 (Kepler GK107) -> Got 13000 (FahCore_17) @ 2014-04-07T14:35:17Z
One observation that I have is over the last week or so is that my GTX 650 mostly gets FahCore_17 WUs while my GTX 660 Ti mostly gets FahCore_15 WUs. Currently, both WUs will finish in about two hours so will post what WUs will be assigned.
ETA:
Now ↞ Very Soon ↔ Soon ↔ Soon-ish ↔ Not Soon ↠ End Of Time