Raistmer, it was a special moment...

I just installed BOINC 6.6.20.
Then I installed your new v10/11 with no affinity and no CPU->GPU.
Then I ran it for 30 minutes.

Then I got this validated task result:

http://setiathome.berkeley.edu/workunit.php?wuid=429443399

It was a WU that validated with your machine.

Isn't that an amazing service Raistmer provides? He works on Opti Apps, and he validates your WU's for you, too!

The new software is running perfect. Thanks for all the work you did on it.

Bob

That's hilarious.. Bob.. would you mind handing over some of that RAW power... woot... :P It still takes me 40 minutes to complete a similar task.. Giggle..:P

And Raistmer.. it's running perfect for me too.. hee hee..

"still giggling"...

G

---

Boinc Wiki




"Great spirits have always encountered violent opposition from mediocre minds." -Albert Einstein

What could be here the prob?

With Raistmer's V7 kill mod I didn't had this prob.

Now with Raistmer's V7 .dll's and V10 CUDA mod app.

The 2. time I got this error:
http://setiathome.berkeley.edu/result.php?resultid=1196613052

'Out Of Memory' ?

My CPU/RAM/mobo isn't OCed.
My GPUs are OCed from manufacturer.

---

Please, keep posting stderr output about "-12 triplet error", it contains some useful info along with error message.

Here's one from my rig with V10 (task #1197925306):

Exception detected inside cudaAcc_find_triplets, dumping client state
icfft=73222, PoT_activity=0, PoT_freq_bin=-1SETI@home error -12 Unknown error
cudaAcc_find_triplets erroneously found a triplet twice in find_triplets_kernel
File: ..\analyzePoT.cpp
Line: 348

---

What could be here the prob?

With Raistmer's V7 kill mod I didn't had this prob.

Now with Raistmer's V7 .dll's and V10 CUDA mod app.

The 2. time I got this error:
http://setiathome.berkeley.edu/result.php?resultid=1196613052

'Out Of Memory' ?

My CPU/RAM/mobo isn't OCed.
My GPUs are OCed from manufacturer.

Try the .dll's from Raistmer's original V10 download,
I'm using them with no problem.

Claggy

What could be here the prob?

With Raistmer's V7 kill mod I didn't had this prob.

Now with Raistmer's V7 .dll's and V10 CUDA mod app.

The 2. time I got this error:
http://setiathome.berkeley.edu/result.php?resultid=1196613052

'Out Of Memory' ?

My CPU/RAM/mobo isn't OCed.
My GPUs are OCed from manufacturer.

CPUID: AMD Processor model unknown

Cache: L1=64K L2=512K
strange, unknown CPU....

Please, keep posting stderr output about "-12 triplet error", it contains some useful info along with error message.

Here's one from my rig with V10 (task #1197925306):

Exception detected inside cudaAcc_find_triplets, dumping client state
icfft=73222, PoT_activity=0, PoT_freq_bin=-1SETI@home error -12 Unknown error
cudaAcc_find_triplets erroneously found a triplet twice in find_triplets_kernel
File: ..\analyzePoT.cpp
Line: 348

thanks

What could be here the prob?

With Raistmer's V7 kill mod I didn't had this prob.

Now with Raistmer's V7 .dll's and V10 CUDA mod app.

The 2. time I got this error:
http://setiathome.berkeley.edu/result.php?resultid=1196613052

'Out Of Memory' ?

My CPU/RAM/mobo isn't OCed.
My GPUs are OCed from manufacturer.

CPUID: AMD Processor model unknown

Cache: L1=64K L2=512K
strange, unknown CPU....

Nothing specially.. or secret.. ;-D

I have the 'old' BIOS (V1.5) on my Mobo. [MSI K9A2 Platinum]
In BIOS V1.7 I should/would have the new CPU Name..
It's the new AMD Phenom II X4 940 BE.. :-)

But.. ;-) ..you could guess/know why I got this error [to now 2. times] ?


..with your V7 mod I got not this error..

---

Iv'e got these lately, they both completed after falling back to host cpu.
with run time error not enough memory
using v10
this one
and this one

Also to my upper question.. :-)

..

Since it's not allowed to ask CUDA related questions in the NC forum..

Profis - please, have a look in the Questions & answers : CUDA :

> BOINC V6.6.11 and 1/2 CUDA performance - thread.


Thanks a lot!

---

This are the two 'Out Of Memory' errors:
http://setiathome.berkeley.edu/result.php?resultid=1195475990


[This error I posted already in this thread]
http://setiathome.berkeley.edu/result.php?resultid=1196613052


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


http://setiathome.berkeley.edu/result.php?resultid=1195495666
Exception detected inside cudaAcc_find_triplets, dumping client state
icfft=86040, PoT_activity=0, PoT_freq_bin=-1SETI@home error -12 Unknown error
cudaAcc_find_triplets erroneously found a triplet twice in find_triplets_kernel
File: ..\analyzePoT.cpp
Line: 348



http://setiathome.berkeley.edu/result.php?resultid=1190297258
Exception detected inside cudaAcc_find_triplets, dumping client state
icfft=94365, PoT_activity=0, PoT_freq_bin=-1SETI@home error -12 Unknown error
cudaAcc_find_triplets erroneously found a triplet twice in find_triplets_kernel
File: ..\analyzePoT.cpp
Line: 348



[This error I posted already in this thread]
http://setiathome.berkeley.edu/result.php?resultid=1195160891
Exception detected inside cudaAcc_find_triplets, dumping client state
icfft=86665, PoT_activity=0, PoT_freq_bin=-1SETI@home error -12 Unknown error
cudaAcc_find_triplets erroneously found a triplet twice in find_triplets_kernel
File: ..\analyzePoT.cpp
Line: 348

---

Iv'e got these lately, they both completed after falling back to host cpu.
with run time error not enough memory
using v10
this one
and this one

Total GPU memory 335216640 free GPU memory 64638976
nothing unusual - your host had too small free memory amount indeed.
True question is not why it fallback to CPU but why on starting time it had so low amount of free GPU memory available?

Also to my upper question.. :-)

..

Since it's not allowed to ask CUDA related questions in the NC forum..

Profis - please, have a look in the Questions & answers : CUDA :

> BOINC V6.6.11 and 1/2 CUDA performance - thread.


Thanks a lot!

NC == Number crunching. Who thinks that GPU doesn't do number crunching ?? GPU performance topics look absolute relevant to this forum as much as Cell blade performance, CPU performance and motherboard/RAM performance....
Monitoring many different places is boring actually...

[This error I posted already in this thread]

Thanks.
It seems this error occurs in different places. More statistic will be good.
I start to put such errors in table.
Only line with state dump (like icfft=86665, PoT_activity=0, PoT_freq_bin=-1SETI@home error -12 Unknown error) worth to post.
Thanks again to all who participate in collection of these bug reports.

NC == Number crunching. Who thinks that GPU doesn't do number crunching ??

Admin/Mods apparently.

---

me@rescam.org

NC == Number crunching. Who thinks that GPU doesn't do number crunching ??

Admin/Mods apparently.

Not me.

But the reason for the CUDA forum is permissions. People who may have just installed BOINC and have no RAC can't post over here. The entire Q&A section of the message boards allows people with no RAC to post if they have a problem.

---

O.K., but there in the 'Questions & answers : CUDA :' - area are only 3 or 4 people around which could have knowledge - ..or not.

Here in the NC are more people around which could have knowledge because of special topics or problems.

..therefore it would be better to allow CUDA related topics here in the NC.


..or make a new NC-CUDA under-forum..

---

O.K., but there in the 'Questions & answers : CUDA :' - area are only 3 or 4 people around which could have knowledge - ..or not.

Here in the NC are more people around which could have knowledge because of special topics or problems.

..therefore it would be better to allow CUDA related topics here in the NC.


..or make a new NC-CUDA under-forum..

It depends on the nature of the CUDA question. If we don't allow CUDA questions here, as you suggest, then why are Raistmer's threads here? The more technically advanced posts stay here in Number Crunching. The simple questions go into the CUDA forum. There is no need for an NC-CUDA forum.

---

NC == Number crunching. Who thinks that GPU doesn't do number crunching ??

Admin/Mods apparently.

Not me.

But the reason for the CUDA forum is permissions. People who may have just installed BOINC and have no RAC can't post over here. The entire Q&A section of the message boards allows people with no RAC to post if they have a problem.

The post permissions were only supposed to be imposed on the Cafe. That's where the problems were when it was implemented.

It depends on the nature of the CUDA question. If we don't allow CUDA questions here, as you suggest, then why are Raistmer's threads here? The more technically advanced posts stay here in Number Crunching. The simple questions go into the CUDA forum. There is no need for an NC-CUDA forum.

That's all relative. Who's to say what is simple and what is advanced? Most likely the poster. So you're right back to the only criteria being the post permission.

---

me@rescam.org

NC == Number crunching. Who thinks that GPU doesn't do number crunching ??

Admin/Mods apparently.

Not me.

But the reason for the CUDA forum is permissions. People who may have just installed BOINC and have no RAC can't post over here. The entire Q&A section of the message boards allows people with no RAC to post if they have a problem.

The post permissions were only supposed to be imposed on the Cafe. That's where the problems were when it was implemented.

Its not up to me on the post permissions. I do not have control over that. All I know is that the entire board except all of Q&A need to have a RAC of 1 to create a post.

It depends on the nature of the CUDA question. If we don't allow CUDA questions here, as you suggest, then why are Raistmer's threads here? The more technically advanced posts stay here in Number Crunching. The simple questions go into the CUDA forum. There is no need for an NC-CUDA forum.

That's all relative. Who's to say what is simple and what is advanced? Most likely the poster. So you're right back to the only criteria being the post permission.

Which is why I mod based on personal judgement. So far I've had no complaints. Then again, this is all of topic in Raistmer's thread, and I doubt he'd take kindly to us hijacking it for this purpose.

---

O.K., but there in the 'Questions & answers : CUDA :' - area are only 3 or 4 people around which could have knowledge - ..or not.

Here in the NC are more people around which could have knowledge because of special topics or problems.

..therefore it would be better to allow CUDA related topics here in the NC.


..or make a new NC-CUDA under-forum..

It depends on the nature of the CUDA question. If we don't allow CUDA questions here, as you suggest, then why are Raistmer's threads here? The more technically advanced posts stay here in Number Crunching. The simple questions go into the CUDA forum. There is no need for an NC-CUDA forum.

CUDA Questions-PLEASE READ - sticky thread in the NC forum.

Hmm.. it's not a mod-team idea.. only from Blurf?

---