• Welcome to Overclockers Forums! Join us to reply in threads, receive reduced ads, and to customize your site experience!

new to seti/forum (sorta)

Overclockers is supported by our readers. When you click a link to make a purchase, we may earn a commission. Learn More.

capedcrewsader

Registered
Joined
Feb 27, 2012
I remember my dad running seti on all of our old computer years ago while idling and until recently had a crappy lil system I built 7 years ago (decent at the time ..dated obviously by now)

just stepped up into the realm of building a new rig (I5-2500ks are a steal for the performance!) and ventured back into overclocking, got everything up/running/stable and where I'd like it at 4.5ghz.

use to lurk here for information and did a bit more in my research before putting this system together...decided to start on seti@home also since this thing has power to spare and joined the team.

hope I will be a useful boost in crunching some numbers!
 
same on SETI capedcrewsader just joined this afternoon probably hasn't updated. I couldn't find myself on any lists yet either
 
http://boincstats.com/stats/user_graph.php?pr=sah&id=9674240 looks like team update hasn't put you in our team yet. Maybe (if we have a clean outage window) tomorrow.
http://setiathome.berkeley.edu/show_user.php?userid=9674240

Hmm Core i5-2500k, Radeon 6x00 series...that should be a nice bit of RAC once it gets maxed out. You've run the optimized installer, right?

the first day or so running it solo I had not ran any sorta optimizer this afternoon I ran the unified installer from lunatics and matched it up for our CPU/GPU been running it full steam ahead since.

*edited after figuring out the answer after reading up some more*

anyone have any idea what kinda RAC I should expect to average running pretty much 24/7 with about 15+ of those hours at 100% CPU/GPU and during the other hours max available -gpu just wanna see if things look right after a few days of running it optimized/make sure it is proper.

also what a "decent" time for completing an astropulse work unit start to finish would be? those look to be huge jobs in comparison to the seti@home enhanced work units?
 
Last edited:
Definitely look into getting BOINC rescheduler as well. Move those AP units to the GPU, ATI cards are great at doing AP workunuits. My 5850 finishes them in around an hour. Also, there's some tweaks you can manually make to if you're running a 68x0 or 69x0 to let it run two WUs at the same time.

My i7-930 @ 4GHz with just a single 5850 is pulling over 17k RAC right now. See other thread in our group about that. Right now, I'm only using 1 (2) cores as well. Not sure which ATI card you're using, but if its a 68 or 69, I would expect at least that much, and probably more RAC. I'm hoping that once I get the other 5850 back in my machine, and figure out my crazy Crossfire issue, that I'll see my main PC get closer to 30k.
 
looks like I picked the wrong day to clear out all my work and try to make adjustments since it will be down for 3 days.

has not been making proper use of my GPU at all and had trouble with the rescheduler so was going to start back over from scratch

whoops LOL
 
quick question. just reinstalled the project, reran the optimizer and now filling the WU's back up

I currently have no astropulse workorders at all, would this be due to the servers catching back up from downtime? I set my cache to have 10 days of work logged at a time so hoping it fills up

right now it is doing 5 tasks at a time with reasonable speed... my GPU is running at full speed but only showing around an average of 33% load on these tasks which are all using applications:
setiathome_ehanced 6.03 / 6.10

basically does it sound like I did something wrong to get it to process astropulse jobs or should they cycle in as the servers catch back up? also for running those types one of the 6.10s says running (0.05 CPUS + 1.00 ATI GPU's) I'm guessing that is the one using 33% load of the GPU for a non-astropulse WU is that just its limitations?

hope my questions make sense the way I'm asking more or less I'm just curious if when an astropulse WU comes up if it will make more than 33% use of my GPU / or if perhaps I did something wrong

edit: got my settings switched around in rescheduler looks like it is working hard now! GPU running at 95% 5 tasks running and it dropped the task using the GPU from "six hours estimated" to less than two i imagine it will be done in about 20 minutes from the rate its counting down right now
 
Last edited:
AP workunits are few and far between and have been for a while now.

As far as your GPU being at 33%, that can be fixed. You have to manually update the app_info.xml to allow it to run multiple instances.

Here's how I have mine set up for a Radeon 5850. If you're running a 68x0 or 69x0 these settings should work nicely.

<app_info>
<app>
<name>setiathome_enhanced</name>
</app>
<file_info>
<name>AK_v8b2_win_x64_SSE41.exe</name>
<executable/>
</file_info>
<app_version>
<app_name>setiathome_enhanced</app_name>
<version_num>603</version_num>
<platform>windows_intelx86</platform>
<file_ref>
<file_name>AK_v8b2_win_x64_SSE41.exe</file_name>
<main_program/>
</file_ref>
</app_version>
<app_version>
<app_name>setiathome_enhanced</app_name>
<version_num>603</version_num>
<platform>windows_x86_64</platform>
<file_ref>
<file_name>AK_v8b2_win_x64_SSE41.exe</file_name>
<main_program/>
</file_ref>
</app_version>
<app>
<name>astropulse_v505</name>
</app>
<file_info>
<name>ap_5.06_win_x86_SSE2_OpenCL_ATI_r521.exe</name>
<executable/>
</file_info>
<file_info>
<name>AstroPulse_Kernels_r521.cl</name>
<executable/>
</file_info>
<app_version>
<app_name>astropulse_v505</app_name>
<version_num>506</version_num>
<avg_ncpus>0.04</avg_ncpus>
<max_ncpus>0.2</max_ncpus>
<platform>windows_intelx86</platform>
<plan_class>ati13ati</plan_class>
<cmdline>-instances_per_device 2 -unroll 12 -ffa_block 8192 -ffa_block_fetch 4096</cmdline>
<coproc>
<type>ATI</type>
<count>0.5</count>
</coproc>
<file_ref>
<file_name>ap_5.06_win_x86_SSE2_OpenCL_ATI_r521.exe</file_name>
<main_program/>
</file_ref>
<file_ref>
<file_name>AstroPulse_Kernels_r521.cl</file_name>
<copy_file/>
</file_ref>
</app_version>
<app_version>
<app_name>astropulse_v505</app_name>
<version_num>505</version_num>
<avg_ncpus>0.04</avg_ncpus>
<max_ncpus>0.2</max_ncpus>
<platform>windows_intelx86</platform>
<plan_class>ati13ati</plan_class>
<cmdline>-instances_per_device 2 -unroll 12 -ffa_block 8192 -ffa_block_fetch 4096</cmdline>
<coproc>
<type>ATI</type>
<count>0.5</count>
</coproc>
<file_ref>
<file_name>ap_5.06_win_x86_SSE2_OpenCL_ATI_r521.exe</file_name>
<main_program/>
</file_ref>
<file_ref>
<file_name>AstroPulse_Kernels_r521.cl</file_name>
<copy_file/>
</file_ref>
</app_version>
<app_version>
<app_name>astropulse_v505</app_name>
<version_num>506</version_num>
<avg_ncpus>0.04</avg_ncpus>
<max_ncpus>0.2</max_ncpus>
<platform>windows_x86_64</platform>
<plan_class>ati13ati</plan_class>
<cmdline>-instances_per_device 2 -unroll 12 -ffa_block 8192 -ffa_block_fetch 4096</cmdline>
<coproc>
<type>ATI</type>
<count>0.5</count>
</coproc>
<file_ref>
<file_name>ap_5.06_win_x86_SSE2_OpenCL_ATI_r521.exe</file_name>
<main_program/>
</file_ref>
<file_ref>
<file_name>AstroPulse_Kernels_r521.cl</file_name>
<copy_file/>
</file_ref>
</app_version>
<app_version>
<app_name>astropulse_v505</app_name>
<version_num>505</version_num>
<avg_ncpus>0.04</avg_ncpus>
<max_ncpus>0.2</max_ncpus>
<platform>windows_x86_64</platform>
<plan_class>ati13ati</plan_class>
<cmdline>-instances_per_device 2 -unroll 12 -ffa_block 8192 -ffa_block_fetch 4096</cmdline>
<coproc>
<type>ATI</type>
<count>0.5</count>
</coproc>
<file_ref>
<file_name>ap_5.06_win_x86_SSE2_OpenCL_ATI_r521.exe</file_name>
<main_program/>
</file_ref>
<file_ref>
<file_name>AstroPulse_Kernels_r521.cl</file_name>
<copy_file/>
</file_ref>
</app_version>
<app>
<name>astropulse_v505</name>
</app>
<file_info>
<name>ap_5.05r409_SSE.exe</name>
<executable/>
</file_info>
<app_version>
<app_name>astropulse_v505</app_name>
<version_num>505</version_num>
<platform>windows_intelx86</platform>
<file_ref>
<file_name>ap_5.05r409_SSE.exe</file_name>
<main_program/>
</file_ref>
</app_version>
<app_version>
<app_name>astropulse_v505</app_name>
<version_num>505</version_num>
<platform>windows_x86_64</platform>
<file_ref>
<file_name>ap_5.05r409_SSE.exe</file_name>
<main_program/>
</file_ref>
</app_version>
<app>
<name>setiathome_enhanced</name>
</app>
<file_info>
<name>MB6_win_x86_SSE3_OpenCL_ATi_HD5_r390.exe</name>
<executable/>
</file_info>
<file_info>
<name>MultiBeam_Kernels_r390.cl</name>
<executable/>
</file_info>
<app_version>
<app_name>setiathome_enhanced</app_name>
<version_num>610</version_num>
<platform>windows_intelx86</platform>
<avg_ncpus>0.05</avg_ncpus>
<max_ncpus>0.05</max_ncpus>
<plan_class>ati13ati</plan_class>
<cmdline>-period_iterations_num 20 -instances_per_device 2</cmdline>
<coproc>
<type>ATI</type>
<count>0.5</count>
</coproc>
<file_ref>
<file_name>MB6_win_x86_SSE3_OpenCL_ATi_HD5_r390.exe</file_name>
<main_program/>
</file_ref>
<file_ref>
<file_name>MultiBeam_Kernels_r390.cl</file_name>
<copy_file/>
</file_ref>
</app_version>
<app_version>
<app_name>setiathome_enhanced</app_name>
<version_num>610</version_num>
<platform>windows_x86_64</platform>
<avg_ncpus>0.05</avg_ncpus>
<max_ncpus>0.05</max_ncpus>
<plan_class>ati13ati</plan_class>
<cmdline>-period_iterations_num 20 -instances_per_device 2</cmdline>
<coproc>
<type>ATI</type>
<count>0.5</count>
</coproc>
<file_ref>
<file_name>MB6_win_x86_SSE3_OpenCL_ATi_HD5_r390.exe</file_name>
<main_program/>
</file_ref>
<file_ref>
<file_name>MultiBeam_Kernels_r390.cl</file_name>
<copy_file/>
</file_ref>
</app_version>
</app_info>
 
AP workunits are few and far between and have been for a while now.

As far as your GPU being at 33%, that can be fixed. You have to manually update the app_info.xml to allow it to run multiple instances.

Here's how I have mine set up for a Radeon 5850. If you're running a 68x0 or 69x0 these settings should work nicely.

[[[snipped code]]]


went ahead and updated all the code in the app_info.xml file, reran optimizer, and it still is processing the same 5 WUs currently should it have adjusted to run more instances with that or should it begin working on more after these complete? still only using around 33-40% of the GPU but perhaps it is just the unit being worked on?
edit: after pausing those and starting everything over again looks like the changes took affect!
 
Last edited:
Excellent. You should be running 5 tasks now still, however, now it should be 2GPU and 3 CPU, right?

i changed the app_info.xml file and adjusted the last two to .50/.50

now it shows me having 3 CPU and 2GPU running with both GPU showing:
Running (0.50 CPUS + 0.50 ATI GPUS)

looking at the properties for one GPU it shows an estimated size/app speed that says it would take it 10 hours to complete (estimated task size 169761 GFLOPs/ estimated app speed 4.51GFLOPs/sec). looks like it will take around one hour currently

sound about right? slow for this build/decent?

some completed tasks (from GPU..CPU times are just *slightly* better than the "run times"):
Run time 4,326.76
CPU time 137.75

Run time 3,150.88
CPU time 107.97


last question for now I think, other than modifying the settings in app_info.xml to pull .50 CPU for both GPU applications is there somewhere else/another setting I should be swapping to change the preferences to only use 3 cores to leave the 4th open to feed the GPU as you recommended?
 
Last edited:
In BOINC Manager, go to advanced, preference, processor usage and on the "On multuiprocessor systems, use at most" set it to 75%.

I don't think you'd need .50/.50 usually GPU WUs need just enough CPU time to get fed data. On your setup, .04/.20 should be plenty.
 
gotcha! I was misunderstanding how it used that setting, swapped the final two back down to .05/.05 with the first applications set to .04/.2

change it over to using 75% of the processors at most also. thanks for the help, now to leave it crunching through the night and see what things look like in the morning!

GPU load is staying right at around 98% solid now
 
RAC is slowly climbing up, any ideas for how long it takes to stabliize/become more accurate? seems half my completed tasks are still pending so far in about 2-3 full days:

27k work done, RAC 2319

I see others talking about 20-30k+ RAC I seem to be pretty far off that mark
 
BoincStats shows your machine did 7k yesterday, that would be a good place to start watching. Generally, from what Ive seen RAC is usually 1-2k less than what your daily output is. From what I can see, looks like you're running a 6670 so that number seems somewhat accurate. Generally it will take a new machine two-three weeks to really settle into its max RAC. I'd think that getting it to close to 9k shouldn't be too tough. Three CPU cores should be good for 4k or so, the GPU should easily do that much as well.
 
awesome, looks like ended up crunching around 10k yesterday in total sounds like 7-8k will be around what I hit then

thanks again for the help, doublechecked back with my dad when visiting over there earlier with our daughter, it was him that I remembered running seti or something similar many many years ago on all of our linux builds we had around the house during idling time although it was a much much different look and a whole lot less processing power back then that is for sure! says the app went public in 1999 but I distinctly remember him running it (or whatever it was at the time?) prior to that closer to 97/98? who knows maybe it was 1999 and I'm just forgetful! I remember the app running what visually looked like a standard disc defrag running across the screensaver.

first thing I thought once I got this new rig up and running was, hey I've got some excess power now so had to fish around and try to find what it was he use to run. was cool getting the verification today my memory was correct!

anyway thanks for the help here, I did a bit of browsing around the forums and reading up on things here when deciding/putting together this system figured I'd try to help out a lil.
 
Back