Since my HD5770 has run out of AP, and i want to cut down on the lecky use, i've removed it,

Claggy

Since my HD5770 has run out of AP, and i want to cut down on the lecky use, i've removed it,

Claggy

I will not give up. I'll keep my 4850 running. Sooner or later, AP will flow freeely, and I'll be picking them up when they come. Until then, it will have to crunch MB, even if that pays much less/hour than AP on the GPU.

The kitties are awaiting developments on AP for NV...
Lunatics is working on it.
Meow!

---

"Time is simply the mechanism that keeps everything from happening all at once."

I'm not getting anything...no tasks available... :(
But it'll pick up

---

"Two things are infinite: The universe and human stupidity; and I'm not sure about the universe." - Albert Einstein

The kitties are awaiting developments on AP for NV...
Lunatics is working on it.
Meow!

Wonder if they will keep us updated here on Seti forum on development or keep it in house.

JohnDK, it's gone into Lunatics beta but there are still a few things to work out. And then of course is the fact SETI isn't giving us any new APs right now. :-(

---

PROUD MEMBER OF Team Starfire World BOINC

JohnDK, it's gone into Lunatics beta but there are still a few things to work out. And then of course is the fact SETI isn't giving us any new APs right now. :-(

Yes a bit difficult without any APs to work on, lets hope it will be better soon :)

I have 20 left than i´m dry also.
But got 1300 MBs today, so enough to reshedule back and forth.

---

With each crime and every kindness we birth our future.

I have 20 left than i´m dry also.
But got 1300 MBs today, so enough to reshedule back and forth.

How does one Reschedule from MB to AP (or vice versa)?
I'm very curious...

---

I have 20 left than i´m dry also.
But got 1300 MBs today, so enough to reshedule back and forth.

How does one Reschedule from MB to AP (or vice versa)?
I'm very curious...

You don't. They mean reschedule MB work to keep their GPU busy.



I saw the upload server was offline for a bit a few hours ago. It is back up now. So maybe when they guys leave for the day they will kick the AP splitters back on. I was going to do some playing with my HD3850 & it is just sitting there spinning its little fan waiting for work. :(

---

SETI@home classic workunits: 93,865 CPU time: 863,447 hours
Join the [url=http://tinyurl.com/8y46zvu]BP6/VP6 User Group[

I have 20 left than i´m dry also.
But got 1300 MBs today, so enough to reshedule back and forth.

How does one Reschedule from MB to AP (or vice versa)?
I'm very curious...

You don't. They mean reschedule MB work to keep their GPU busy.

You can reschedule AP too,

Claggy

I have 20 left than i´m dry also.
But got 1300 MBs today, so enough to reshedule back and forth.

How does one Reschedule from MB to AP (or vice versa)?
I'm very curious...

You don't. They mean reschedule MB work to keep their GPU busy.

You can reschedule AP too,

Claggy

Yes, but not from MB's, but between AP CPU, and AP ATI GPU. I did that often in the beginning, but rescheduling F's up the DCF, and it makes the scheduler think you're faster or slower depending on which way you reschedule, and ultimately you will get much lower credit for many WU's than you would otherwise have had.

So, I've totally stopped rescheduling, and gone to unticking/ticking CPU/GPU on the SETI@home preferences page. Since ATI GPU have no options there, it's harder to fix that, but one can always choose to get MB and AP ATI, by unticking CPU on the SETI@home preferences.

I do the same as you, i don't rescheduler ever any more, haven't for well over 6 months, But i do have a method of changing the device a Wu is going to run on,
I manually edit my client_state.xml and remove the two file_info sections, the result section and the workunit section for a Wu, then get the Server to resend the Wu to the other device,
(Disable CPU and Nvidia GPU work fetch, and only ATI GPU work fetch is active) and since you already have the Wu file, the download is skipped.

Claggy

I have 20 left than i´m dry also.
But got 1300 MBs today, so enough to reshedule back and forth.

How does one Reschedule from MB to AP (or vice versa)?
I'm very curious...

You don't. They mean reschedule MB work to keep their GPU busy.

You can reschedule AP too,

Claggy

Yes, but not from MB's, but between AP CPU, and AP ATI GPU. I did that often in the beginning, but rescheduling F's up the DCF, and it makes the scheduler think you're faster or slower depending on which way you reschedule, and ultimately you will get much lower credit for many WU's than you would otherwise have had.

So, I've totally stopped rescheduling, and gone to unticking/ticking CPU/GPU on the SETI@home preferences page. Since ATI GPU have no options there, it's harder to fix that, but one can always choose to get MB and AP ATI, by unticking CPU on the SETI@home preferences.

I do the same as you, i don't rescheduler ever any more, haven't for well over 6 months, But i do have a method of changing the device a Wu is going to run on,
I manually edit my client_state.xml and remove the two file_info sections, the result section and the workunit section for a Wu, then get the Server to resend the Wu to the other device,
(Disable CPU and Nvidia GPU work fetch, and only ATI GPU work fetch is active) and since you already have the Wu file, the download is skipped.

Claggy

Wow Claggy that was a creative solution to say the least. Heh, I would never had thought about that. That's a keeper, and I will blatantly steal the idea from you, if I ever need to do a reschedule in that way.

Other little snippits to know, if you do it to a Wu that's close to deadline, when it gets resent, it also gets a new deadline,
that was very handy on my Atom (when i completed my 10 validations of AP, and got more work, i got far too much to do within the existing deadlines)

Claggy

My I7-2600+2x HD5870s, now has only MB tasks, had quite a lot of
AstroPulse work.
MB, are doing OK also, doesn't even matter if a vlar slipped through,
cause they just take some more time, just like a CPU would.
ASTROPULSE does pay OK, depends on what hardware, GPU or CPU

---

Yes, but not from MB's, but between AP CPU, and AP ATI GPU. I did that often in the beginning, but rescheduling F's up the DCF, and it makes the scheduler think you're faster or slower depending on which way you reschedule, and ultimately you will get much lower credit for many WU's than you would otherwise have had.

Well maybe that's why the RAC on the Frozen 920 has been dropping since I did the reschedule and it's working the CPU again.

Drat....confounded newcredit system.

DCF will sort itself in a bit, I suppose.

---

"Time is simply the mechanism that keeps everything from happening all at once."

Running Boinc Version 6.10.60 and I had one box that was requesting gpu and cpu work but only receiving work for the gpu. Went on for days like this. I resceduled exactly 4 tasks to the quad cpu and immediately started receiving work for the cpu's when requesting.

Don't know but this does not seem as proper response on Boinc's part.

---

Boinc....Boinc....Boinc....Boinc....

The kitties are awaiting developments on AP for NV...
Lunatics is working on it.
Meow!

Wonder if they will keep us updated here on Seti forum on development or keep it in house.

Okay, I finished my first AP on NVIDIA GPU a few minutes ago. It took 2 hours 41 minutes and 46 seconds for 671.35 credits.(My wingman had already finished as it was a resend to me.)It was 2.39% blanked. According to EVGA Precision my memory usage stayed around 372MB and GPU usage bounced from 15 to 65%. Temp stayed at about 61C.

If anyone would like to look at it, this is it.. http://setiathome.berkeley.edu/workunit.php?wuid=756661080

---

PROUD MEMBER OF Team Starfire World BOINC

Okay, I finished my first AP on NVIDIA GPU a few minutes ago. It took 2 hours 41 minutes and 46 seconds for 671.35 credits.(My wingman had already finished as it was a resend to me.)It was 2.39% blanked. According to EVGA Precision my memory usage stayed around 372MB and GPU usage bounced from 15 to 65%. Temp stayed at about 61C.

If anyone would like to look at it, this is it.. http://setiathome.berkeley.edu/workunit.php?wuid=756661080

Thanks for the update :)

I wonder about how much credits you would get if you spend 2h41m doing MBs? My calculations sucks, but 671 credits and 2h42ms for one AP doesn't seem very much?

Running three WUs at a time, mid range work takes me about 45 minutes so about 15 minutes each at between 90 and 150 credits each. Right now, I have the app set to take the whole GPU to run but I'm thinking of trying to knock that down to .5 and try running one AP and one MB CUDA at the same time. The memory usage shouldn't be a problem but it might be a bit much for the GPU usage. Should be fun to try though.

---

PROUD MEMBER OF Team Starfire World BOINC

Okay, I finished my first AP on NVIDIA GPU a few minutes ago. It took 2 hours 41 minutes and 46 seconds for 671.35 credits.(My wingman had already finished as it was a resend to me.)It was 2.39% blanked. According to EVGA Precision my memory usage stayed around 372MB and GPU usage bounced from 15 to 65%. Temp stayed at about 61C.

If anyone would like to look at it, this is it.. http://setiathome.berkeley.edu/workunit.php?wuid=756661080

Thanks for the update :)

I wonder about how much credits you would get if you spend 2h41m doing MBs? My calculations sucks, but 671 credits and 2h42ms for one AP doesn't seem very much?

IIRC, the machine (and wingmen) need at least 10 WUs granted, before the Cr. are +/- correct/stable. OTOH, with the currently new Cr.system this is only theoretically.. ;-)

OTOH, if you compare the Cr./hour with MB vs. AP on a nVIDIA GPU, I guess AP is the winner.. ;-)

-----------------------------

BTW.
Some news about when the nVIDIA OpenCL AP app will be released for all for wide beta-test? Or maybe directly a release of a finished build?


- Best regards! - Sutaru Tsureku, team seti.international founder. - Optimize your PC for higher RAC. - SETI@home needs your help. -

---

I don't have a time frame. It's up to Raistmer when he thinks it's ready for primetime. Since it is in beta over on Lunatics it shouldn't take too much longer before he is happy with it but that all depends on how the beta testing goes. If there are any other Lunatics BETA testers over here that would like to try it, I'm sure he would like the help to get it out as fast as possible. I can only mess things up so much, I need a bit more help to really screw it up! From his reply to my report over there, I might have been the first to report on a Fermi card but I don't know. I know Jason was talking about maybe doing a barebones installer to upgrade to the x39e he is working on. Maybe Raistmer could come up with a way to add this to something like that when it is ready.

---

PROUD MEMBER OF Team Starfire World BOINC