Done. When I was running the CUDA32 it was finishing up what I had & just came down just running 2 is that the idea? I just have a regular one & that Astropulse & sure is running allot faster that's for sure thanks for the help.

---

Done. When I was running the CUDA32 it was finishing up what I had & just came down just running 2 is that the idea? I just have a regular one & that Astropulse & sure is running allot faster that's for sure thanks for the help.

Best to run 1 WU at a time to see what the card is capable of.
You can also make use of command line variables to improve performance. Once you've got it running 1 at a time as it's best possible speed, then see how 2 at a time goes.
With SoG each GPU WU being processed needs a CPU core to support it for maximum performance.

Reserving 1 CPU core will also help reduce any impact on CPU processing times by not over committing the system.

Edit- make sure you have selected the AVX application for the CPU- will get another 30%+ boost in CPU performance using it.

To reserve a CPu core for each GPU WU being processed.
In your
C:\ProgramData\BOINC\projects\setiathome.berkeley.edu
create a new file (with Notepad) and name it
app_config.xm
and put the following contents in it.

<app_config>
 <app>
  <name>setiathome_v8</name>
  <gpu_versions>
  <gpu_usage>1.00</gpu_usage>
  <cpu_usage>1.00</cpu_usage>
  </gpu_versions>
 </app>
 <app>
  <name>astropulse_v7</name>
  <gpu_versions>
  <gpu_usage>0.5</gpu_usage>
  <cpu_usage>1.0</cpu_usage>
  </gpu_versions>
 </app>
</app_config>

In the BOINC manager, use Options, Read config files to get it to use those settings.

---

Grant
Darwin NT

He should be able to run 2 at a time on that card if he doesn't mind the noise and Heat on it. But he will have to run a command line to get the most out of it. Remember the 970 only have 3.5 GB of usable memory. I'd also probably use a max concurrent to free up a core or 2 for the GPU. We can talk about that later. For now, let it run and finish up what he has. Don't know when we will get more work so he can fine tune it

---

Done. When I was running the CUDA32 it was finishing up what I had & just came down just running 2 is that the idea? I just have a regular one & that Astropulse & sure is running allot faster that's for sure thanks for the help.

Best to run 1 WU at a time to see what the card is capable of.
You can also make use of command line variables to improve performance. Once you've got it running 1 at a time as it's best possible speed, then see how 2 at a time goes.
With SoG each GPU WU being processed needs a CPU core to support it for maximum performance.

Reserving 1 CPU core will also help reduce any impact on CPU processing times by not over committing the system.

Edit- make sure you have selected the AVX application for the CPU- will get another 30%+ boost in CPU performance using it.

To reserve a CPu core for each GPU WU being processed.
In your
C:\ProgramData\BOINC\projects\setiathome.berkeley.edu
create a new file (with Notepad) and name it
app_config.xm
and put the following contents in it.

<app_config>
 <app>
  <name>setiathome_v8</name>
  <gpu_versions>
  <gpu_usage>1.00</gpu_usage>
  <cpu_usage>1.00</cpu_usage>
  </gpu_versions>
 </app>
 <app>
  <name>astropulse_v7</name>
  <gpu_versions>
  <gpu_usage>0.5</gpu_usage>
  <cpu_usage>1.0</cpu_usage>
  </gpu_versions>
 </app>
</app_config>

In the BOINC manager, use Options, Read config files to get it to use those settings.

Since he's running a newer BOINC he might want to try the newer app_config.xml

<app_config>
  <app_version>
    <app_name>setiathome_v8</app_name>
    <plan_class>opencl_nvidia_SoG</plan_class>
    <avg_ncpus>1</avg_ncpus>
    <ngpus>0.5</ngpus>
    <cmdline>-sbs 768 -period_iterations_num 10 -hp -high_prec_timer -high_perf -tt 500</cmdline>
  </app_version>
   <app_version>
     <app_name>astropulse_v7</app_name>
     <plan_class>opencl_nvidia_100</plan_class>
     <avg_ncpus>1</avg_ncpus>
      <ngpus>0.5</ngpus>
      <cmdline>-unroll 28 -oclFFT_plan 256 16 256 -ffa_block 12288 -ffa_block_fetch 6144 -tune 1 64 4 1 -tune 2 64 4 1 -hp </cmdline>
   </app_version>
 <project_max_concurrent>7</project_max_concurrent>
</app_config>

---

Since he's running a newer BOINC he might want to try the newer app_config.xml

I prefer to keep it simple.
:-)

---

Grant
Darwin NT

I did the config file & saved both of them. Many thanks...

P.S. I can run 2 jobs with no issues. Now even though I have 7 fans that include the CPU fan & video card fans. It never overheats at all.

---

Also that when some people choose to shout or even scream, there always should be a reason for why there could be no response back in return,
and for such a thing I could blame project management, if not the project itself.

Making it such a thing and it still could be the user experience, but next you could also be left with the other alternative, who chooses to keep silent for a similar reason,
and therefore not being any responsive at all.

I did the config file & saved both of them. Many thanks...

P.S. I can run 2 jobs with no issues. Now even though I have 7 fans that include the CPU fan & video card fans. It never overheats at all.

Times look good. I can see that you finished 2 OpenCl before Seti went down. See commandlines took in the last work unit that reported.

---

Hopefully, I can run it 12hrs straight get a real RAC out of it as long the splitters do not go down again.

---

Hopefully, I can run it 12hrs straight get a real RAC out of it as long the splitters do not go down again.

Looking much better!
Don't expect RAC to move all that quickly. It often takes a week or so to see significant changes, due to how it's calculated. Don't quote me, but I think it's like a 30 day running average? (Anyone feel free to correct/clarify)

---

Hopefully, I can run it 12hrs straight get a real RAC out of it as long the splitters do not go down again.

Looking much better!
Don't expect RAC to move all that quickly. It often takes a week or so to see significant changes, due to how it's calculated. Don't quote me, but I think it's like a 30 day running average? (Anyone feel free to correct/clarify)

Without major system outages, major hardware changes take about 2 months for RAC to "stabilise."

---

Grant
Darwin NT

Times look good.

Yeah, 45min instead of 1hr 22min for a CPU WU, with only a few seconds difference between Run time & CPU time instead of 10min I would consider to be a significant improvement.

---

Grant
Darwin NT

Thanks to Zalsters help & others the Lunatics SoG Build & custom XML App_Config has really made it go allot faster my RAC has increased & according to Zalster I avg about 7 minutes on GPU. (Geforce 970) Just one PC try to run it between 8-12hrs a day.

---

That is what I am thinking here & the question is why are they are doing it? You all have a good argument here. I realize we all here to donate our time & our High-End PC for these projects & not all about credits but yes, you deserve credit, where it takes power to get it done some of us, have servers going or just one PC. Newbie me 643 Tasks done getting close to a 1,000 avg only 55 credits so yeah, I understand the sour taste in everyone mouth.

---

CreditNew v2

I hereby propose that Seti@home volunteers adopt for themselves CreditNew v2 whereby the true RAC for Seti@home from henceforth (until things change) shall be the current reported Credit or RAC multiplied by 10.

To calculate your new RAC or Credits per WU simply add a zero after all the reported credits or RAC so a Credit of 55 per WU is now 550 and a RAC of 4,526 is actually a RAC of 45,260.

You will immediately feel better and more valued after you apply CreditNew v2 to your existing Credits and RAC.

Disclaimer: Participation in CreditNew v2 is entirely voluntary and it may not work for everyone.

CreditNew v2

I hereby propose that Seti@home volunteers adopt for themselves CreditNew v2 whereby the true RAC for Seti@home from henceforth (until things change) shall be the current reported Credit or RAC multiplied by 10.

To calculate your new RAC or Credits per WU simply add a zero after all the reported credits or RAC so a Credit of 55 per WU is now 550 and a RAC of 4,526 is actually a RAC of 45,260.

You will immediately feel better and more valued after you apply CreditNew v2 to your existing Credits and RAC.

Disclaimer: Participation in CreditNew v2 is entirely voluntary and it may not work for everyone.

ROFLMAO. Good one!
I shall adopt your new credit scheme immediately.
There, I feel better already............

Meow!

---

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

the bigger you have, the better you feel


lol

---

the bigger you have, the better you feel


lol

That's what sh.. oops... Sorry, that one almost slipped out... ;-)

---

What if I happened to say such a thing?

I guess a number of us gents would probably just look down at the floor, shuffle our feet and mumble something incoherent? :-)

or maybe just chuckle and smile possibly? It depends on ones sense of humor I suppose, but as it's all in good fun, I would hope no one would take offense.

---