To anyone that may know.

Firstly apologies for posting a new thread but I have searched the forums and not found the answer or a reference to it anywhere.

The question I have is as I understand it Cuda will not work with a optimised app or an app info file (at the moment), if I am wrong please correct me, what I would like to know is how do I get to a vanilla version that will run the gpu application without losing units in progress or units waiting to run so I dont upset any prospective wingmen.

I have already backed up my current work in case its all too hard and want to use both CPU'S instead so is it just a matter of uninstalling and reinstalling Boinc and transferring existing files or is it a bit more involved?

I have already installed the Cuda drivers and the latest Boinc app but I am currently running the following.


<core_client_version>6.4.5</core_client_version>
<![CDATA[
<stderr_txt>
Unrecognized XML in parse_init_data_file: computation_deadline
Skipping: 1229928491.136000
Skipping: /computation_deadline
Unrecognized XML in parse_init_data_file: computation_deadline
Skipping: 1229928491.136000
Skipping: /computation_deadline
Windows optimized S@H Enhanced application by Alex Kan
Version info: SSSE3x (Intel, Core 2-optimized v8-nographics) V5.13 by Alex Kan
SSSE3x Win32 Build 41 , Ported by : Jason G, Raistmer, JDWhale

CPUID: Intel(R) Core(TM)2 Duo CPU E4700 @ 2.60GHz
Speed: 2 x 3008 MHz
Cache: L1=64K L2=2048K
Features: MMX SSE SSE2 SSE3 SSSE3

Work Unit Info:
...............
Credit multiplier is : 2.85
WU true angle range is : 2.701628

Flopcounter: 5215093103983.061500

Spike count: 0
Pulse count: 0
Triplet count: 3
Gaussian count: 0
called boinc_finish

</stderr_txt>
]]>

Any help would be greatly appreciated.

---

My quad host is doing alot of these also.

---

To anyone that may know.

Firstly apologies for posting a new thread but I have searched the forums and not found the answer or a reference to it anywhere.

The question I have is as I understand it Cuda will not work with a optimised app or an app info file (at the moment), if I am wrong please correct me, what I would like to know is how do I get to a vanilla version that will run the gpu application without losing units in progress or units waiting to run so I dont upset any prospective wingmen.

I have already backed up my current work in case its all too hard and want to use both CPU'S instead so is it just a matter of uninstalling and reinstalling Boinc and transferring existing files or is it a bit more involved?

I have already installed the Cuda drivers and the latest Boinc app but I am currently running the following.


<core_client_version>6.4.5</core_client_version>
<![CDATA[
<stderr_txt>
Unrecognized XML in parse_init_data_file: computation_deadline
Skipping: 1229928491.136000
Skipping: /computation_deadline
Unrecognized XML in parse_init_data_file: computation_deadline
Skipping: 1229928491.136000
Skipping: /computation_deadline
Windows optimized S@H Enhanced application by Alex Kan
Version info: SSSE3x (Intel, Core 2-optimized v8-nographics) V5.13 by Alex Kan
SSSE3x Win32 Build 41 , Ported by : Jason G, Raistmer, JDWhale

CPUID: Intel(R) Core(TM)2 Duo CPU E4700 @ 2.60GHz
Speed: 2 x 3008 MHz
Cache: L1=64K L2=2048K
Features: MMX SSE SSE2 SSE3 SSSE3

Work Unit Info:
...............
Credit multiplier is : 2.85
WU true angle range is : 2.701628

Flopcounter: 5215093103983.061500

Spike count: 0
Pulse count: 0
Triplet count: 3
Gaussian count: 0
called boinc_finish

</stderr_txt>
]]>

Any help would be greatly appreciated.

For something as (apparently) unpredictable as the current CUDA build, I would run down the cache by setting NNT then remove (or re-name) the app_info.xml file when the cache is empty.

F.

---

Thanks Fred sometimes the simplest answers are the best.

---

Wonder if you could do an app_info file to crunch MB with the CUDA app and also crunch AP with the remaining cores. As follows.
I do not know if this works, just asking!!

<app_info>	
    <app>
        <name>setiathome_enhanced</name>
    </app>
    <file_info>
        <name>setiathome_6.05_windows_intelx86__cuda.exe</name>
        <executable/>
    </file_info>
    <app_version>
        <app_name>setiathome_enhanced</app_name>
        <version_num>605</version_num>
        <file_ref>
            <file_name>setiathome_6.05_windows_intelx86__cuda.exe</file_name>
            <main_program/>
        </file_ref>
    </app_version>
    <app>
	<name>astropulse</name>
    </app>
    <file_info>
	<name>ap_5.00r69_SSE3.exe</name>
	<executable/>
    </file_info>
    <file_info>
	<name>libfftw3f-3-1-1a_upx.dll</name>
	<executable/>
    </file_info>
    <app_version>
	<app_name>astropulse</app_name>
	<version_num>500</version_num>
	<file_ref>
	    <file_name>ap_5.00r69_SSE3.exe</file_name>
	    <main_program/>
	</file_ref>
	<file_ref>
	    <file_name>libfftw3f-3-1-1a_upx.dll</file_name>
	</file_ref>
    </app_version>
</app_info>

Of course your optimized app for AP may have a different file name and would have to be edited to the file name you have.

as always, crunch your cache down before changing the app_info file

---

Boinc....Boinc....Boinc....Boinc....

I'm pretty sure that there are at least two new DLLs that come with the CUDA client app. I had a look last night at writing an app info file for my main beast, but decided against even testing it because I have a feeling (with the number of compute errors that CUDA clients are kicking up) that the CUDA app is going to get an update sooner rather than later.

Not entirely happy about it either; my main box can, with an optimised app, chew through an AP workunit in around 19 hours. Now it's ability to do that is curtailed as I'm forced to use an unoptimised client, or stop receiving AP units completely (something I don't want to do as my other machines can benefit perfectly well from using 'traditional' SSE optimised apps).

---

Not entirely happy about it either; my main box can, with an optimised app, chew through an AP workunit in around 19 hours. Now it's ability to do that is curtailed as I'm forced to use an unoptimised client, or stop receiving AP units completely (something I don't want to do as my other machines can benefit perfectly well from using 'traditional' SSE optimised apps).

You could try what I do, and set the one that you're running CUDA on to a different location like home or work and set the optimaized one to a different location. Then go into your SETI@home preferences page and disable AP for the location the unoptimized pc is on.

I then use the gpu to crunch the CUDA MB for Seti, and my cpu for the new Einstein opti app.

---

Could someone please post the names of any DLL files that are required for CUDA?

So far I have the name of the CUDA app....setiathome_6.05_windows_intelx86__cuda.exe

But not any DLL's if any are required.

---

Boinc....Boinc....Boinc....Boinc....

You could try what I do, and set the one that you're running CUDA on to a different location like home or work and set the optimaized one to a different location. Then go into your SETI@home preferences page and disable AP for the location the unoptimized pc is on.

I then use the gpu to crunch the CUDA MB for Seti, and my cpu for the new Einstein opti app.

Thanks for the advice; I've just done this. Not sure how long it'll take for Monolith to pick up the new settings, and I can't force it to as the machine is at home and I'm at work. I think I'll just abort the two AP workunits it's working on at the moment too, as (from what I can see from here) they must be working at high priority for some reason, and thus blocking the client from getting more Enhanced WUs.

Could someone please post the names of any DLL files that are required for CUDA?

So far I have the name of the CUDA app....setiathome_6.05_windows_intelx86__cuda.exe

But not any DLL's if any are required.

If noone else has provided this by the time I get home this evening, I'll post them.

---

Names of the DLLs (Ignore the tags please, experimentation in progress)

<file_info>
<name>cudart.dll</file_name>
<executable/>
</file_info>
<file_info>
<name>cufft.dll</file_name>
<executable/>
</file_info>
<file_info>
<name>libfftw3f-3-1-1a_upx.dll</file_name>
<executable/>
</file_info>

---

"Living by the wisdom of computer science doesn't sound so bad after all. And unlike most advice, it's backed up by proofs." -- Algorithms to live by: The computer science of human decisions.

With the help of previous post I have gotten Boinc to work with the optimized Astropulse (ap_5.00r69_SSE.exe) client and CUDA 6.05 with the following app_into.xml file:


<app_info>
<app>
<name>setiathome_enhanced</name>
</app>
<file_info>
<name>setiathome_6.05_windows_intelx86__cuda.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>605</version_num>
<file_ref>
<file_name>setiathome_6.05_windows_intelx86__cuda.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</name>
</app>
<file_info>
<name>ap_5.00r69_SSE3.exe</name>
<executable/>
</file_info>
<file_info>
<name>libfftw3f-3-1-1a_upx.dll</name>
<executable/>
</file_info>
<app_version>
<app_name>astropulse</app_name>
<version_num>500</version_num>
<file_ref>
<file_name>ap_5.00r69_SSE3.exe</file_name>
<main_program/>
</file_ref>
<file_ref>
<file_name>libfftw3f-3-1-1a_upx.dll</file_name>
</file_ref>
</app_version>
</app_info>

---

Arvid Almstrom

Thanks for that, I'll give it a try when I empty my current CUDA cache.

---

With the help of previous post I have gotten Boinc to work with the optimized Astropulse (ap_5.00r69_SSE.exe) client and CUDA 6.05 with the following app_into.xml file:

<snip>

I have a few extra lines in there that seem to make it think its a cuda task. My app_info is below. The bit in bold is the extras (based upon info from Jason Gee and some experimenting)

<app_info>
<app>
<name>astropulse</name>
</app>
<file_info>
<name>ap_5.00r69_SSE3.exe</name>
<executable/>
</file_info>
<file_info>
<name>libfftw3f-3-1-1a_upx.dll</name>
<executable/>
</file_info>
<app_version>
<app_name>astropulse</app_name>
<version_num>500</version_num>
<file_ref>
<file_name>ap_5.00r69_SSE3.exe</file_name>
<main_program/>
</file_ref>
<file_ref>
<file_name>libfftw3f-3-1-1a_upx.dll</file_name>
</file_ref>
</app_version>
<app>
<name>setiathome_enhanced</name>
</app>
<file_info>
<name>setiathome_6.05_windows_intelx86__cuda.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>605</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>setiathome_6.05_windows_intelx86__cuda.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_info>

---

BOINC blog

Went ahead and installed the app info file and AP files posted by MarkJ, now just need to get assigned an AP unit. CUDA running so I assume it won't be given an AP until I finish the existing cache running.

Edit: First task completed with the new app info file added to task already in progress. Instead of 00.00.00 completion times I now have times ranging from 268.48.38-905.45.50 on remaining tasks.

---

Just thought I'd add, with the app info file completion times are decreasing with each task. As of my last start up of a new task, times decreased about 3 hours on the estimate for the remaining shorties I have.

---

I am going to try a different twist to the app_info.xml above....

I added the cuda alongside the MB+AP just to see what happens....

~Bob

---

Do you Good Search for Seti@Home? http://www.goodsearch.com/?charityid=888957
Or Good Shop? http://www.goodshop.com/?charityid=888957

try what i posted here
http://setiathome.berkeley.edu/forum_thread.php?id=50818

I am going to try a different twist to the app_info.xml above....

I added the cuda alongside the MB+AP just to see what happens....

~Bob

Keep us informed; I'll be very interested to hear if you can get a positive result.

---

I am going to try a different twist to the app_info.xml above....

I added the cuda alongside the MB+AP just to see what happens....

~Bob

Keep us informed; I'll be very interested to hear if you can get a positive result.

Well I got some cuda wu's :)
My AKv8 wu's I received before this test are still there and they are not displaying as a CUDA app.. Now its up to the server to see if it will assign any non cuda work... Optimized AP is running alongside cuda as well.

---

Do you Good Search for Seti@Home? http://www.goodsearch.com/?charityid=888957
Or Good Shop? http://www.goodshop.com/?charityid=888957

try what i posted here
http://setiathome.berkeley.edu/forum_thread.php?id=50818

or maybe try that you can run a ap wu along side of cuda ones and lower cpu's to use my one