I just got:

28/02/2012 01:27:34 | SETI@home | Reporting 2 completed tasks, requesting new tasks for NVIDIA GPU
28/02/2012 01:27:46 | SETI@home | Scheduler request completed: got 0 new tasks
28/02/2012 01:27:46 | SETI@home | No tasks sent
28/02/2012 01:27:46 | SETI@home | No tasks are available for Astropulse v5
28/02/2012 01:27:46 | SETI@home | No tasks are available for Astropulse v505
28/02/2012 01:27:46 | SETI@home | This computer has reached a limit on tasks in progress

But only have 1,437 WUs. Given I have 4 GPUs I would be allowed 1,600 for the GPUs alone. Is there yet another limit?

Probably ghost tasks but that can't be checked right now with the results disabled.

Ghosts would have been resent unless they were VLAR in which case they'd have expired on that request.

We don't know how and when the counting is done.
Those tasks reported might have been CPU.

Or the tasks were CPU and would have been resent, but for the limits...

Thank you for the ideas, but the request was only for GPU WUs so I would not expect the CPUs being at the limit should matter.

Thank you for the ideas, but the request was only for GPU WUs so I would not expect the CPUs being at the limit should matter.

Complete a CPU Wu (and have CPU work fetch disabled), then try and get GPU work,
i believe if you're reached eithier CPU or GPU limit,

Claggy

How do I disable CPU work fetch but still keep processing on my CPUs please?

Setting "Use CPU" to "no" in your SETI@home preferences only prevents new CPU tasks from being downloaded, not already downloaded ones from running.

Gruß,
Gundolf

How do I disable CPU work fetch but still keep processing on my CPUs please?

Via the SETI@home preferences page.

The second item

Use CPU: no
Enforced by version 6.10+

does exactly what you want.

How do I disable CPU work fetch but still keep processing on my CPUs please?

By changing your SETI@home preferences, select 'Use CPU' to 'No', the next time Boinc connects to Seti the Seti preferences will be updated, now Boinc won't asks for CPU work again, when you feel you have enough GPU work, reenable Use CPU.

Note, if you want to do this to only one host, move that host to a different venue, and make separate SETI@home preferences for it,

ie, My PIII & C2D E8500 are on the default venue, my C2D T8100 & Atom N450 on the home venue, the XP3200+ on the school venue, and the i7-2600K/GTX460/HD5770 on the work venue,
this enables me to have the Laptop & Netbook only ask for Astropulse work & have Nvidia GPU work fetch disabled, the PIII & E8500 not ask for Astropulse work, and i can adjust the preferences for the i7-2600K as i like.

Claggy

I may not be following the flow of this thread quite correctly but ...

if the 400 limit per gpu is a pressing problem for you and running 2 WU's per GPU doesn't affect the 400 limit ... then ...

if you are an undauntable BOINC adventurer try running 2 BOINC's and split each GPU between the 2 BOINCS ... each will get the 400 limit ...

I have one of my computers split into 2 with 4 CPU's running lunatics SSE3 (cpu's 1,3,5,7) and 4 CPU's running lunatics ssse3 code (cpu's 0,2,4,6) the 1 GPU I have running lunatics x41g_win32_cuda32.exe*32 from each BOINC instance (one running on cpu's 0,2,4,6 and the other running on cpu's 1,3,5,7). (I'm doing the even/odd cpu stuff just for fun ;-) ).

on the even CPU BOINC instance I have 200 cpu WU's and 400 GPU WU's
on the odd CPU BOINC instance i have 200 cpu WU's and 400 GPU WU's.

if you have a monster rig, this can be expanded to 4 or 8 boincs (it just takes MORE work on your part to keep it all together ;-).

If this idea helps you other crazy's out there ... have at it

Ed F

I may not be following the flow of this thread quite correctly but ...

if the 400 limit per gpu is a pressing problem for you and running 2 WU's per GPU doesn't affect the 400 limit ... then ...

if you are an undauntable BOINC adventurer try running 2 BOINC's and split each GPU between the 2 BOINCS ... each will get the 400 limit ...

I have one of my computers split into 2 with 4 CPU's running lunatics SSE3 (cpu's 1,3,5,7) and 4 CPU's running lunatics ssse3 code (cpu's 0,2,4,6) the 1 GPU I have running lunatics x41g_win32_cuda32.exe*32 from each BOINC instance (one running on cpu's 0,2,4,6 and the other running on cpu's 1,3,5,7). (I'm doing the even/odd cpu stuff just for fun ;-) ).

on the even CPU BOINC instance I have 200 cpu WU's and 400 GPU WU's
on the odd CPU BOINC instance i have 200 cpu WU's and 400 GPU WU's.

if you have a monster rig, this can be expanded to 4 or 8 boincs (it just takes MORE work on your part to keep it all together ;-).

If this idea helps you other crazy's out there ... have at it

Ed F

A cheeky monkey could also have one instance for CPU. Then 2 or 3 more for each GPU to effectively circumvent the limits as well. For that purpose I would think it is more work than it is worth, but some people are bored enough to do just that.

---

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

red-ray wrote:

I just got:

28/02/2012 01:27:34 | SETI@home | Reporting 2 completed tasks, requesting new tasks for NVIDIA GPU
28/02/2012 01:27:46 | SETI@home | Scheduler request completed: got 0 new tasks
28/02/2012 01:27:46 | SETI@home | No tasks sent
28/02/2012 01:27:46 | SETI@home | No tasks are available for Astropulse v5
28/02/2012 01:27:46 | SETI@home | No tasks are available for Astropulse v505
28/02/2012 01:27:46 | SETI@home | This computer has reached a limit on tasks in progress

But only have 1,437 WUs. Given I have 4 GPUs I would be allowed 1,600 for the GPUs alone. Is there yet another limit?

red-ray wrote:

Thank you for the ideas, but the request was only for GPU WUs so I would not expect the CPUs being at the limit should matter.

The BOINC logic to try to give reasons when no work is sent doesn't always make a lot of sense. You're right that being at the CPU limit would not prohibit getting GPU work, but it will cause the "reached a limit" message if work is not sent for some other reason. Given the shown messages, I think there were some SETI@home Enhanced tasks available but they were .vlar so couldn't be sent. That special case doesn't generate a message of its own.

The idea that it might be ghosts doesn't work; the check for lost work is done so early the messages about no available AP tasks would not have happened.

                                                                  Joe

Seems so strange to me that people who are donating processor time to this project are being limited to how much they can actually perform. I thought the point of the project was to identify signals from space that may be of a non earth origin. There is a whole lot of sky to search and enough work to keep everyone busy for years. This bottleneck of work distribution is kind of silly in todays computing environment. So get with the program or somebody else will.

Seems so strange to me that people who are donating processor time to this project are being limited to how much they can actually perform. I thought the point of the project was to identify signals from space that may be of a non earth origin. There is a whole lot of sky to search and enough work to keep everyone busy for years. This bottleneck of work distribution is kind of silly in todays computing environment. So get with the program or somebody else will.

Jack.....first of all, welcome to the forums.

Seti is steadily advancing in the distribution of work, the hardware infrastructure to distribute that work, and the science behind that work.

The current limitations in the amount of work that can be cached on each computer are due to some problems with Boinc server code that were introduced some months back, and are still being worked on.
They only become a problem when the servers go down or slow down, which we hope will become far less often in the near future when 2 brand new donated servers come online.

Nobody likes the limits less than me and the kitties....we never like to see our crunchers without kibble to munch. But I understand the reasons that they cannot be lifted right now, and am biding my time whilst crunching everything that comes my way.

---

"Freedom is just Chaos, with better lighting." Alan Dean Foster

Seems so strange to me that people who are donating processor time to this project are being limited to how much they can actually perform. I thought the point of the project was to identify signals from space that may be of a non earth origin. There is a whole lot of sky to search and enough work to keep everyone busy for years. This bottleneck of work distribution is kind of silly in todays computing environment. So get with the program or somebody else will.

It isn't really strange at all to limit how work people are allowed to have at once. Even without the current limits that are in place there is still a limit of x days of work that a machine is allowed to build up.

In generally with the volume of people donating their processor time we can crunch through the data faster then they can receive and rip it out to us.

Another thing to remember is that in today's environment this project has limited funding to pay for everything. Including their network & server maintenance.

One thing they really need to add on the project requirements page is patience. If you don't have that then you will quickly get frustrated.

---

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