I have a couple of quad-core systems running BOINC, and have them configured to keep 3 days of work in queue (some other projects...errr...can't provide new work for days at a time). With MW, this results in a "Message from server: (reached limit of 24 tasks in progress)" every couple of minutes. Since there's no project-specific queue size setting, is there another way to tell MW how many tasks to request?

It is 6 tasks per cpu core.
It has been explained numerous times. This projects needs the results to make more tasks. So having a system have 3 days of work for gpus would be many thousands per system and by the time the results are in they would be "out of date" for their calculations. It would produce bad results for the project.

---

Doesn't expecting the unexpected make the unexpected the expected?
If it makes sense, DON'T do it.

I know it has been explained, but did you make it to the end of my post before you got annoyed? Because of other "lazy" projects, I want a 3 day queue (in fact, that isn't enough most of the time), but I can't set it for THEM without setting it for MW. So I was wondering if there was another way to suppress the error messages (a MW-specific setting that says, "you don't need 3 days for this project, thank you!")

It is 6 tasks per cpu core.
It has been explained numerous times. This projects needs the results to make more tasks. So having a system have 3 days of work for gpus would be many thousands per system and by the time the results are in they would be "out of date" for their calculations. It would produce bad results for the project.

correct me if i'm wrong, but CPU WUs takes literally hours to produce the result. does it mean that MW do not "need the results to make more tasks"? deadline for these WUs is EIGHT days, so it means that the project do not need the result before. right?

in my understanding it's really weird to give 6 WUs per core while i'm using GPU to crunch. hey, guys, you mixing apples and oranges. it make way more sense to use type and quantity of GPUs to estimate the number of WUs to be sent.

---

Its a shame they can't create an app/WU designated for GPU only. IF they did and knowing that most GPU's finish the WU's in under 5 minutes, then we could possibly have more than 1 hour of GPU work on our computers

---

Blah blah blah you know the rest

I'd be happy if they made this project a GPU only project, there'd be no excuse then for increasing the number of wu's cached.

Here's an oddity. I'm only getting 10 WU's per machine now. this really makes me worry about keeping work running

---

Blah blah blah you know the rest

I can't recive wu for cpu because gpu took all 6 tasks of limit.I can see this problem here for some time and no1 fixed it yet?
CTAPbIi,nice point...

I can't recive wu for cpu because gpu took all 6 tasks of limit.I can see this problem here for some time and no1 fixed it yet?
CTAPbIi,nice point...

You should get 24 wus (6 per core). Make sure you have 'Use CPU' checked in 'My Account' on the MW site. Also you seem to have some errors too, that can limit the wus you get.

"Task 284888777
Name de_separation_17_3s_fix_1_1727375_1294444842_1
Workunit 214649357
Client state Compute error
Exit status -1073741819 (0xffffffffc0000005)
Computer ID 251742
Report deadline 16 Jan 2011 2:45:08 UTC
Run time 571.360001
CPU time 66.58123
stderr out <core_client_version>6.10.58</core_client_version>
<![CDATA[
<message>
- exit code -1073741819 (0xc0000005)
</message>
<stderr_txt>
Running Milkyway@home ATI GPU application version 0.23 (Win32, CAL 1.4) by Gipsel
ignoring unknown input argument in app_info.xml: -np
ignoring unknown input argument in app_info.xml: 20
ignoring unknown input argument in app_info.xml: -p
ignoring unknown input argument in app_info.xml: 0.6287266820817938000000000
ignoring unknown input argument in app_info.xml: 7.2275680020594000000000000
ignoring unknown input argument in app_info.xml: -1.6796468342614490000000000
ignoring unknown input argument in app_info.xml: 164.7750503709946800000000000
ignoring unknown input argument in app_info.xml: 28.2823511614459300000000000
ignoring unknown input argument in app_info.xml: -2.0123427713877600000000000
ignoring unknown input argument in app_info.xml: -0.2223713225459644200000000
ignoring unknown input argument in app_info.xml: 2.8824382083202060000000000
ignoring unknown input argument in app_info.xml: -1.1614431260952500000000000
ignoring unknown input argument in app_info.xml: 179.7365264641748300000000000
ignoring unknown input argument in app_info.xml: 2.4068725982289365000000000
ignoring unknown input argument in app_info.xml: -6.2831853071795860000000000
ignoring unknown input argument in app_info.xml: 6.2831853071795860000000000
ignoring unknown input argument in app_info.xml: 5.0438907481288010000000000
ignoring unknown input argument in app_info.xml: -0.8270843487885879000000000
ignoring unknown input argument in app_info.xml: 190.0317938522658800000000000
ignoring unknown input argument in app_info.xml: 4.8516910488575510000000000
ignoring unknown input argument in app_info.xml: -6.2831853071795860000000000
ignoring unknown input argument in app_info.xml: -1.0097060205432127000000000
ignoring unknown input argument in app_info.xml: 12.7312019801904960000000000
instructed by BOINC client to use device 0
CPU: Intel(R) Core(TM)2 Quad CPU Q6600 @ 2.40GHz (4 cores/threads) 2.39396 GHz (344ms)

CAL Runtime: 1.4.880
Found 1 CAL device

Device 0: ATI Radeon HD3800 (RV670) 512 MB local RAM (remote 1470 MB cached + 1470 MB uncached)
GPU core clock: 777 MHz, memory clock: 1126 MHz
320 shader units organized in 4 SIMDs with 16 VLIW units (5-issue), wavefront size 64 threads
supporting double precision

Starting WU on GPU 0

main integral, 640 iterations (1600x1400), 3 streams
predicted runtime per iteration is 699 ms (33.3333 ms are allowed), dividing each iteration in 21 parts
borders of the domains at 0 80 152 232 304 384 464 536 616 688 768 840 920 992 1072 1144 1224 1296 1376 1448 1528 1600
Calculated about 3.28897e+013 floatingpoint ops on GPU, 2.47165e+008 on FPU. Approximate GPU time 502.263 seconds.

cut number 0, 640 iterations (400x1400), 3 streams
predicted runtime per iteration is 174 ms (33.3333 ms are allowed), dividing each iteration in 6 parts
borders of the domains at 0 72 136 200 272 336 400


Unhandled Exception Detected...

- Unhandled Exception Record -
Reason: Access Violation (0xc0000005) at address 0x6B777FFF write attempt to address 0x00000004

Engaging BOINC Windows Runtime Debugger...

"

---

Doesn't expecting the unexpected make the unexpected the expected?
If it makes sense, DON'T do it.

Only 6 tasks because i have allowed 1 of 4 cores for crunching and errors happens after some1 forgets turn off boinc before comp :/
I think its not going to solve anything(6 or 24). because the system gave me gpu replacement for finished cpu task.This can happen with 24 tasks too i guess.

Only 6 tasks because i have allowed 1 of 4 cores for crunching and errors happens after some1 forgets turn off boinc before comp :/
I think its not going to solve anything(6 or 24). because the system gave me gpu replacement for finished cpu task.This can happen with 24 tasks too i guess.

Never mind,now i got your post fully.U are right,i have unticked "use cpu" thx for solving problem.