IFRS 的帖子

141) 留言板 : Number crunching : Anyone else have issues with the new Nvidia driver? (消息 924747)
发表于：8 Aug 2009 作者： IFRS
Post:
Im having the downclocking problem on my 295 (not on the 250s or the 9600s), but im pretty sure its not due to 190.38, as it seems to happen with previous drivers too.
GTX 295/WinXP 32bit/6.6.20 BOINC
Does any one have a solution for this problem?

142) 疑难解答 : GPU applications : How to install CUDA 2.3? (消息 920712)
发表于：23 Jul 2009 作者： IFRS
Post:
Ty popandbob, i did the update, its seems to working fine, but I still had the message saying that CUDA is 1.1 version...?
143) 疑难解答 : GPU applications : How to install CUDA 2.3? (消息 920693)
发表于：23 Jul 2009 作者： IFRS
Post:
Noob question: I just update my drivers to 19038 but im still getting this msg:
23/7/2009 12:56:06 CUDA devices: GeForce GTS 250 (driver version 19038, CUDA version 1.1, 512MB, est. 89GFLOPS), GeForce 9600 GT (driver version 19038, CUDA version 1.1, 512MB, est. 43GFLOPS)

What to have CUDA 2.3 working? Where can I Dl the needed files?
144) 留言板 : Number crunching : Panic Mode On (20) Server problems (消息 917990)
发表于：15 Jul 2009 作者： IFRS
Post:
I don´t mind if I can´t upload my results for over a week or more. But having idle machines that can´t download more work because the out pile is full, while work to distribute and process increase, that don´t make any sense to me.
There is any work around this "feature"?

145) 留言板 : Number crunching : Astropulse on CUDA? (消息 916921)
发表于：11 Jul 2009 作者： IFRS
Post:
I didn´t find any post about it, but it is a possibility at all? There is any ongoing work on it, or it will never happen?
That will give some help for servers.

If I'm not mistaken, Nvidia did the port for MB.

NVIDIA certainly contributed a lot of resources - see the AUTHORS file downloaded with your copy of the CUDA application:

CUDA development by:
Przemyslaw Zych (NVIDIA)
Scott Le Grand (NVIDIA)
Bob Johnston (NVIDIA)

But I doubt they could have "done" it - a unique task, for a unique platform, that needed to be extended at the same time to handle the unique processor - would have needed input from the resident 'house' team too.

Maybe the house team could do the next port themselves - or maybe they would draw on the optimisation crews instead/as well. But as Ned says, I doubt it's top priority right now.



I would love to get rid of the small units. Faster computers (i.e. multiple CUDA processors) should be able to choose to crunch only big units, thus helping on server traffic.
146) 留言板 : Number crunching : Astropulse on CUDA? (消息 916590)
发表于：10 Jul 2009 作者： IFRS
Post:
I didn´t find any post about it, but it is a possibility at all? There is any ongoing work on it, or it will never happen?
That will give some help for servers.
147) 留言板 : Number crunching : AP ready to send? (消息 913600)
发表于：3 Jul 2009 作者： IFRS
Post:
Richard´s tip made the trick. Im glad to be back at the big units, no matter what credit they give. I hate hitting bruno over and over and over again with the small ones.
I remember the time when the computers will hit the server once in several days to send results and have more work.
With multiple CUDA beasts and and i7's, you should be able to chose to process units 5, maybe 10 times bigger than AP, in my oppinion. I will gladly process units that can take 5 days to process on my i7, that will be a challenge.
Processing AP on CUDA will help the servers with traffic too.
148) 留言板 : Number crunching : AP ready to send? (消息 913453)
发表于：3 Jul 2009 作者： IFRS
Post:
Great Richard, thank you very much. I already fix the error you point out. Hopefully I will get more 5.05 units to test that, as soon i can flush out all that stucked work...
149) 留言板 : Number crunching : AP ready to send? (消息 913391)
发表于：2 Jul 2009 作者： IFRS
Post:
So if any1 can spot the error (or the errors) on my app_info, it will be very appreciated :)


If your app_info is messed up, you'd probably recieve something like [error] State file error: bad application name...,
in which case you wouldn't be able to get any of those two 505s.
You'll find an answer for your 'problem' in this thread posted year ago.
Only thing that I can notice is that you're missing <flops> adjustment which shouldn't affect processing in any way.
You could check famous App_info thread for more info on that.

Cheers.


Ohhh i miss that post, really. Ty for all your info guys :)
150) 留言板 : Number crunching : AP ready to send? (消息 913357)
发表于：2 Jul 2009 作者： IFRS
Post:
It looks like you used IE to paste in the 505 area, IE adds all those extra dashes that caused the problem.

Use either notepad or wordpad to open the app_info.xml file and paste in the corrected code.


Yep you are right.... kinda. I use the IE to cut and paste here on my reply, but noes, I use the notepad to edit my real file, that will be this:

<app_info>
<app>
<name>setiathome_enhanced</name>
</app>
<file_info>
<name>AK_v8b_win_SSE41.exe</name>
<executable/>
</file_info>
<app_version>
<app_name>setiathome_enhanced</app_name>
<version_num>528</version_num>
<file_ref>
<file_name>AK_v8b_win_SSE41.exe</file_name>
<main_program/>
</file_ref>
</app_version>
<app_version>
<app_name>setiathome_enhanced</app_name>
<version_num>603</version_num>
<file_ref>
<file_name>AK_v8b_win_SSE41.exe</file_name>
<main_program/>
</file_ref>
</app_version>

<app>
<name>astropulse_v5</name>
</app>
<file_info>
<name>ap_5.03r112_SSE3.exe</name>
<executable/>
</file_info>
<app_version>
<app_name>astropulse_v5</app_name>
<version_num>503</version_num>
<file_ref>
<file_name>ap_5.03r112_SSE3.exe</file_name>
<main_program/>
</file_ref>
</app_version>

<app>
<name>setiathome_enhanced</name>
</app>
<file_info>
<name>MB_6.08_mod_CUDA_V11_VLARKill_refined.exe</name>
<executable/>
</file_info>
<file_info>
<name>cudart.dll</name>
<executable/>
</file_info>
<file_info>
<name>cufft.dll</name>
<executable/>
</file_info>
<file_info>
<name>libfftw3f-3-1-1a_upx.dll</name>
<executable/>
</file_info>

<app_version>
<app_name>setiathome_enhanced</app_name>
<version_num>608</version_num>
<plan_class>cuda</plan_class>
<avg_ncpus>0.040000</avg_ncpus>
<max_ncpus>0.040000</max_ncpus>
<coproc>
<type>CUDA</type>
<count>1</count>
</coproc>
<file_ref>
<file_name>MB_6.08_mod_CUDA_V11_VLARKill_refined.exe</file_name>
<main_program/>
</file_ref>
<file_ref>
<file_name>cudart.dll</file_name>
</file_ref>
<file_ref>
<file_name>cufft.dll</file_name>
</file_ref>
<file_ref>
<file_name>libfftw3f-3-1-1a_upx.dll</file_name>
</file_ref>
</app_version>

<app>
<name>astropulse_v505</name>
</app>
<file_info>
<name>ap_5.05r168_SSE3.exe</name>
<executable />
</file_info>
<app_version>
<app_name>astropulse_v505</app_name>
<version_num>505</version_num>
<file_ref>
<file_name>ap_5.05r168_SSE3.exe</file_name>
<main_program />
</file_ref>
</app_version>

</app_info>
151) 留言板 : Number crunching : AP ready to send? (消息 913337)
发表于：2 Jul 2009 作者： IFRS
Post:
[quote]

Can you also post your app_info.xml file - possibly a typo somebody might spot?


Sure, here it is:
- <app_info>
- <app>
<name>setiathome_enhanced</name>
</app>
- <file_info>
<name>AK_v8b_win_SSE41.exe</name>
<executable />
</file_info>
- <app_version>
<app_name>setiathome_enhanced</app_name>
<version_num>528</version_num>
- <file_ref>
<file_name>AK_v8b_win_SSE41.exe</file_name>
<main_program />
</file_ref>
</app_version>
- <app_version>
<app_name>setiathome_enhanced</app_name>
<version_num>603</version_num>
- <file_ref>
<file_name>AK_v8b_win_SSE41.exe</file_name>
<main_program />
</file_ref>
</app_version>
- <app>
<name>astropulse_v5</name>
</app>
- <file_info>
<name>ap_5.03r112_SSE3.exe</name>
<executable />
</file_info>
- <app_version>
<app_name>astropulse_v5</app_name>
<version_num>503</version_num>
- <file_ref>
<file_name>ap_5.03r112_SSE3.exe</file_name>
<main_program />
</file_ref>
</app_version>
- <app>
<name>setiathome_enhanced</name>
</app>
- <file_info>
<name>MB_6.08_mod_CUDA_V11_VLARKill_refined.exe</name>
<executable />
</file_info>
- <file_info>
<name>cudart.dll</name>
<executable />
</file_info>
- <file_info>
<name>cufft.dll</name>
<executable />
</file_info>
- <file_info>
<name>libfftw3f-3-1-1a_upx.dll</name>
<executable />
</file_info>
- <app_version>
<app_name>setiathome_enhanced</app_name>
<version_num>608</version_num>
<plan_class>cuda</plan_class>
<avg_ncpus>0.040000</avg_ncpus>
<max_ncpus>0.040000</max_ncpus>
- <coproc>
<type>CUDA</type>
<count>1</count>
</coproc>
- <file_ref>
<file_name>MB_6.08_mod_CUDA_V11_VLARKill_refined.exe</file_name>
<main_program />
</file_ref>
- <file_ref>
<file_name>cudart.dll</file_name>
</file_ref>
- <file_ref>
<file_name>cufft.dll</file_name>
</file_ref>
- <file_ref>
<file_name>libfftw3f-3-1-1a_upx.dll</file_name>
</file_ref>
</app_version>
- <app>
<name>astropulse_v505</name>
</app>
- <file_info>
<name>ap_5.05r168_SSE3.exe</name>
<executable />
</file_info>
- <app_version>
<app_name>astropulse_v505</app_name>
<version_num>505</version_num>
- <file_ref>
<file_name>ap_5.05r168_SSE3.exe</file_name>
<main_program />
</file_ref>
</app_version>
</app_info>

And the machine where i got the 505 fault is my i7:
http://setiathome.berkeley.edu/show_host_detail.php?hostid=4530722

About the errors in that GT9600, its overclocked related. I found that the crapy Lenovo quad where I install it seems not to handle the power needed by the video card, as i had to greatly decrease the clock on that card that i used to run flawlessy on the i7 earlier...

So if any1 can spot the error (or the errors) on my app_info, it will be very appreciated :)
152) 留言板 : Number crunching : AP ready to send? (消息 913295)
发表于：2 Jul 2009 作者： IFRS
Post:
I finally got 2 AP 505, but they dont even start, they turn in a computation error. I got this:

2/7/2009 12:21:04 SETI@home Computation for task ap_22no08ac_B3_P0_00102_20090701_31764.wu_1 finished
2/7/2009 12:21:04 SETI@home Output file ap_22no08ac_B3_P0_00102_20090701_31764.wu_1_0 for task ap_22no08ac_B3_P0_00102_20090701_31764.wu_1 absent
2/7/2009 12:21:04 SETI@home Computation for task ap_22no08ab_B4_P0_00218_20090702_10123.wu_0 finished
2/7/2009 12:21:04 SETI@home Output file ap_22no08ab_B4_P0_00218_20090702_10123.wu_0_0 for task ap_22no08ab_B4_P0_00218_20090702_10123.wu_0 absent

I "think" i have the correct 505 on my app_info, and ap_5.05r168_sse3.exe from lunatics site on my seti folder. What went wrong?
153) 留言板 : Number crunching : AP ready to send? (消息 913118)
发表于：2 Jul 2009 作者： IFRS
Post:
Yes, its all marked mate.. ty for your input. But still, I think there it´s something wrong, because they are not coming in....
154) 留言板 : Number crunching : AP ready to send? (消息 913074)
发表于：1 Jul 2009 作者： IFRS
Post:
Hi,

Did you add 5.05 to the app info file on your pc's?


Yep, I did it.
155) 留言板 : Number crunching : AP ready to send? (消息 913065)
发表于：1 Jul 2009 作者： IFRS
Post:
Now I see on the server status page that there is 1,362 AP units ready to send. But Im not having any on any of my computers. I even set AP 5.05 on some of them, just in case, and no AP to me. Why is that?
156) 留言板 : Number crunching : I'm quitting Seti 'cause of increased power cost coming from "Cap and Trade" (消息 912078)
发表于：27 Jun 2009 作者： IFRS
Post:
Seems like the feelings of joy, pain and fear are the same all around the world. We all like to crunch, and suffer with the hardware and power bill costs. I cut my efforts and then came back later with all i got so many times on this 10 years. Maybe its because of the pain that we are a bit addicted on it? I don´t know...i am a bit addicted on other things aswell. Well some of them may cause cirrhose and kids, instead of melting of polar ice caps :)
But all of them cause me joy, pain and fear, so im pretty sure im addicted on SETI. Until the point im forced to pull back.
157) 留言板 : Number crunching : OK, Lucy, 'splain this one to me. (消息 908050)
发表于：16 Jun 2009 作者： IFRS
Post:
Plus check your client_state and client_state_prev files for <p_ncpus>5</p_ncpus> line and change it to 4. Sometimes, even if you delete your previous cc_config line, that files didnt change and will force one of your cores to process 2 MB units at same time.
158) 留言板 : Number crunching : Overclocking both cards in SLI? (消息 900309)
发表于：28 May 2009 作者： IFRS
Post:
Really? Though that core clock will do the biggest boost.... ty Mike, I will do some more test with shaders clocks :)
159) 留言板 : Number crunching : Overclocking both cards in SLI? (消息 900013)
发表于：27 May 2009 作者： IFRS
Post:
Nice bad actor, with your tip I realize that I dind´t have the last Precision Tool version. Upgrading to the 171 did the trick. Thank you.
160) 留言板 : Number crunching : Overclocking both cards in SLI? (消息 899869)
发表于：27 May 2009 作者： IFRS
Post:
I just add another 9600GT card to my rig, but I cant overclock both of them. Im using EVGA precision tool, and the NVIDIA control panel show that just one of them is overclocked. Times reflect that aswell, since CUDA is 10% faster in the overclocked card than in the other. Can I overclock both cards?


前 20 · 后面 20


 
©2020 University of California
 
SETI@home and Astropulse are funded by grants from the National Science Foundation, NASA, and donations from SETI@home volunteers. AstroPulse is funded in part by the NSF through grant AST-0307956.