Update:

*jumping for joy*

*performing impossible-looking gymnastic routine worthy of a perfect score and gold medal at the Olympics*

Holy Cow!

...or maybe Holy Kitty fits better!

Cruiser's over half-way done with its first all-optimized Astropulse unit, on track for a completion time of less than 31 hours!!!

Compare to non-optimized Astropulse completion time of just less than 98 hours!

Holy speedy completion, Batman!

---

Cruiser
Gateway GT5692 L-f-S Edition
-Phenom X4 9650 CPU
-4GB 667MHz DDR2 RAM
-500GB SATA HD
-Vista x64 SP1
-BOINC 6.2.19 32-bit client
-SSE3 optimized 32-bit apps

Thank you! That was sorely needed and will help SETI's progress!

I've thrown the binaries in there, edited my app_info.xml, stopped and restarted boinc, and it's crunching away!

Looks like my desktop will come in around 23 hours or so compared to 44 with the standard app.

---

I refuse to give in. I will not install 7zip or any other zip like application. I would love to try this out but the format is not a valid zip file. Please post a real zip file.

I refuse to give in. I will not install 7zip or any other zip like application. I would love to try this out but the format is not a valid zip file. Please post a real zip file.

Wow. They offer something for free out of their own time and you want to be stubborn and make demands?

AP v4.35 Rev24b54 SSE3
AP v4.35 Rev30b57 SSE

You're lucky I'm nice. ;)

[Edit] Note that my site only has a 2Mbit upload speed (which would be the maximum others can download it; i.e. my upload is your download), so please don't expect super fast times and don't expect it to be available if everyone suddenly decides to download it all at once (assuming more than one person is interested in the .ZIP format).

I refuse to give in. I will not install 7zip or any other zip like application. I would love to try this out but the format is not a valid zip file. Please post a real zip file.

The nice thing about standards is that there are so many to choose from.

---

One of my boxes came up with an AP error.

WU 343418597 after 9 hours of processing. How it got uploaded and reported with the upload server not working is beyond my understanding. I believe I saw another errored AP workunit by the same computer earlier today.

Yes.......343418593 also uploaded and reported without the benefit of the upload server running. Don't ask me how it did, but it did.

Note.......probably cpu is everheated. Will clock it down a bit.

---

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

One of my boxes came up with an AP error.

WU 343418597 after 9 hours of processing. How it got uploaded and reported with the upload server not working is beyond my understanding. I believe I saw another errored AP workunit by the same computer earlier today.

Yes.......343418593 also uploaded and reported without the benefit of the upload server running. Don't ask me how it did, but it did.

Note.......probably cpu is everheated. Will clock it down a bit.

Your first result failed most likely from this:

In ap_fileio.cpp, Statefile::write, statefile is 0'd, trying again: iteration 1

It cant write state file at checkpointing time. Check your HDD space. (OCing can be the reason too.)
Second task crashed with 0xc0000005 exception. It's excessive OCing most likely.
This result suffered from same error on your host:
http://setiathome.berkeley.edu/result.php?resultid=1002555794
It's SETI MB task.

What about validation of your results? Were they validated OK ?

This one did:


http://setiathome.berkeley.edu/result.php?resultid=1007599411

---

Goodbye Seti Classic

I've just managed to trash all my AP work units on 2 machines now after trying to install the opt app. Apologies to my wingmen. They all complete with a 'computation error' immediately BOINC starts up. I was previously running stock astro with opti seti. My installation process:

• Shut down BOINC
  
• Copied my app_info.xml into projects\setiathome folder
  
• Copied ap_435rev24b54_sse3.exe into projects\setiathome folder (libfftw was already there)
  
• Restarted BOINC
  

There is nothing in the log to indicate why, so I assume its my app_info being wrong. Can someone help me spot anything wrong with this:

<app_info>
<app>
<name>astropulse</name>
</app>
<file_info>
<name>ap_4.35rev24b54_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>435</version_num>
<file_ref>
<file_name>ap_4.35rev24b54_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_version>
<app_name>astropulse</app_name>
<version_num>436</version_num>
<file_ref>
<file_name>ap_4.35rev24b54_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>AK_v8_win_SSSE3x.exe</name>
<executable/>
</file_info>
<app_version>
<app_name>setiathome_enhanced</app_name>
<version_num>517</version_num>
<file_ref>
<file_name>AK_v8_win_SSSE3x.exe</file_name>
<main_program/>
</file_ref>
</app_version>
<app_version>
<app_name>setiathome_enhanced</app_name>
<version_num>527</version_num>
<file_ref>
<file_name>AK_v8_win_SSSE3x.exe</file_name>
<main_program/>
</file_ref>
</app_version>
<app_version>
<app_name>setiathome_enhanced</app_name>
<version_num>528</version_num>
<file_ref>
<file_name>AK_v8_win_SSSE3x.exe</file_name>
<main_program/>
</file_ref>
</app_version>
</app_info>


Here is the bit from the log:

12-Oct-2008 22:29:25 [SETI@home] Starting 19au08ab.3161.20931.13.8.5_1
12-Oct-2008 22:29:25 [SETI@home] Starting task 19au08ab.3161.20931.13.8.5_1 using setiathome_enhanced version 528
12-Oct-2008 22:29:25 [SETI@home] Starting 26au08ad.3200.9070.8.8.38_1
12-Oct-2008 22:29:25 [SETI@home] Starting task 26au08ad.3200.9070.8.8.38_1 using setiathome_enhanced version 528
12-Oct-2008 22:29:25 [SETI@home] Starting 26au08af.2519.10297.5.8.107_0
12-Oct-2008 22:29:25 [SETI@home] Starting task 26au08af.2519.10297.5.8.107_0 using setiathome_enhanced version 528
12-Oct-2008 22:29:25 [SETI@home] Starting 19au08ab.3161.20931.13.8.15_1
12-Oct-2008 22:29:25 [SETI@home] Starting task 19au08ab.3161.20931.13.8.15_1 using setiathome_enhanced version 528
12-Oct-2008 22:29:27 [SETI@home] Computation for task ap_19se08ad_B0_P1_00194_20081009_09250.wu_0 finished
12-Oct-2008 22:29:27 [SETI@home] Computation for task ap_19se08ad_B1_P1_00040_20081009_11020.wu_0 finished
12-Oct-2008 22:29:27 [SETI@home] Computation for task ap_19se08ad_B1_P1_00209_20081009_11020.wu_2 finished
12-Oct-2008 22:29:27 [SETI@home] Computation for task ap_10se08ab_B2_P1_00256_20081010_28635.wu_1 finished
12-Oct-2008 22:30:17 [---] Exit requested by user

---

BOINC blog

MarkJ, delete the spaces in the tags in red :

<file_info>
<name>ap_4.35rev24b54_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>435</version_num>
<file_ref>
<file_name>ap_4.35rev24b54_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_version>
<app_name>astropulse</app_name>
<version_num>436</version_num>
<file_ref>
<file_name>ap_4.35rev24b54_SSE3.exe</file_name>
<main_program />
</file_ref>

---

_\|/_
U r s

Could it be the spaces in the new entries?

Example: <main_program /> or <executable />

My entries don't have the space before the slash.

Otherwise...

When you shut down BOINC, did you also end the process in Task Manager? When I exit BOINC, it still leaves a process running. When I upgraded to the optimized Astropulse application, I also manually ended the process before I installed the new app.

A quote from my previous post about the upgrade:

Suspended the work in progress, shut down BOINC completely (including ending the process), modified the app_info file, copied the new app into the projects directory, restarted BOINC and hit Resume. After a second or two delay (for copying the new app into the slots folders I assume), the progress percentages started jumping much faster than normal. I was quite surprised that it worked - the last time I tried something similar, the work units in progress had errors.

Edit: scooped! Oh well...

---

Cruiser
Gateway GT5692 L-f-S Edition
-Phenom X4 9650 CPU
-4GB 667MHz DDR2 RAM
-500GB SATA HD
-Vista x64 SP1
-BOINC 6.2.19 32-bit client
-SSE3 optimized 32-bit apps

Copied my app_info.xml into projects\setiathome folder

It's recommended not to edit anything inside MB or AP parts of app_info.xml
All you need is just to merge MB part and AP part provided with opt app.
All merged parts should be inside
<app_info>
</app_info>
tags.
Provided app_info.xml is enough to start crunching AP w/o any MB tasks.
All other cases require merging.

Could it be the spaces in the new entries?

Example: <main_program /> or <executable />

My entries don't have the space before the slash.

Otherwise...

When you shut down BOINC, did you also end the process in Task Manager? When I exit BOINC, it still leaves a process running. When I upgraded to the optimized Astropulse application, I also manually ended the process before I installed the new app.

A quote from my previous post about the upgrade:

Suspended the work in progress, shut down BOINC completely (including ending the process), modified the app_info file, copied the new app into the projects directory, restarted BOINC and hit Resume. After a second or two delay (for copying the new app into the slots folders I assume), the progress percentages started jumping much faster than normal. I was quite surprised that it worked - the last time I tried something similar, the work units in progress had errors.

Edit: scooped! Oh well...

It is much cleaner to stop the Boinc process by issuing the "net stop Boinc" command from a cmd window (opened with the "run as Administrator" option if in Vista). This will allow it to close gracefully, whereas crashing it from Task Manager can have undesirable consequences.

F.

---

Well spotted, Urs and L-f-S. "Claggy's Space" (â„¢) is certainly known to cause this problem, ever since Claggy first reported it at Beta.

The question is, where is it coming from? The testers are well aware of the problem, and wouldn't include it in a sample app_info.xml: and I've checked the Lunatics download package, and it isn't there either.

MarkJ, would you mind telling us where you got your Astropulse app_info fragment from (since the spaces aren't in the other half), and which editor you used to make the splice?

Well spotted, Urs and L-f-S. "Claggy's Space" (â„¢) is certainly known to cause this problem, ever since Claggy first reported it at Beta.

The question is, where is it coming from? The testers are well aware of the problem, and wouldn't include it in a sample app_info.xml: and I've checked the Lunatics download package, and it isn't there either.

MarkJ, would you mind telling us where you got your Astropulse app_info fragment from (since the spaces aren't in the other half), and which editor you used to make the splice?

I opened it using IE (the default for XML files). I cut it from IE and pasted it into notepad and saved it as a text file. I then used notepad to add the astropulse section into my existing app_info replacing the bit in there using the stock app.

I'll go and update one of the machines that trashed its AP work and see if it will behave, although i'll have to wait until it gets some more AP work units.

The libfftw is that the same version from the stock app? As I noted above I didn't replace it as the one from the stock app was the same size but more up to date (August 2008).

[update]
I've just edited the app_info and removed the offending spaces. I checked it by opening it with IE (which shows it all nicely formatted) and it has added the space before the slash for the "main program" and "executable". Something else we can blame Microsoft for.

Thanks for the prompt replies guys.

---

BOINC blog

Well spotted, Urs and L-f-S. "Claggy's Space" (â„¢) is certainly known to cause this problem, ever since Claggy first reported it at Beta.

The question is, where is it coming from? The testers are well aware of the problem, and wouldn't include it in a sample app_info.xml: and I've checked the Lunatics download package, and it isn't there either.

MarkJ, would you mind telling us where you got your Astropulse app_info fragment from (since the spaces aren't in the other half), and which editor you used to make the splice?

I used Wordpad to just open and the save the file (something to do with it understanding the CR and translating to CR LF - from another message thread on the optimized AP). I saved it as a text file. I then used notepad to add the astropulse section into my existing app_info replacing the bit in there using the stock app.

I'll go and update one of the machines that trashed its AP work and see if it will behave, although i'll have to wait until it gets some more AP work units.

The libfftw is that the same version from the stock app? As I noted above I didn't replace it as the one from the stock app was the same size but more up to date (August 2008).

I think you'll find that most Windows developers (including Lunatics) supply app_info files with the <CR><LF> convention, so the intermediate Wordpad stage shouldn't be necessary. It's just people whose primary development platform is Linux (and hence use Linux workstations for editing) who add that extra layer of confusion: I've had to point it out to Josh at Berkeley, and Bernd at Einstein (you were active in that thread too, so you may have read it there).

But I wouldn't have expected any version of Wordpad to be 'XML aware' and correct what it thinks to be an XML syntax error (which was the possibility in my mind when I asked about editors). Which takes us back to the first part of the question: who prepared the AP fragment in the first place, and can you remember where you downloaded it from?

Oh - and yes, I think all the copies of libfftw3f-3-1-1a_upx.dll are the same and interchangeable, whatever the datestamp.

Well spotted, Urs and L-f-S. "Claggy's Space" (â„¢) is certainly known to cause this problem, ever since Claggy first reported it at Beta.

The question is, where is it coming from? The testers are well aware of the problem, and wouldn't include it in a sample app_info.xml: and I've checked the Lunatics download package, and it isn't there either.

MarkJ, would you mind telling us where you got your Astropulse app_info fragment from (since the spaces aren't in the other half), and which editor you used to make the splice?

I used Wordpad to just open and the save the file (something to do with it understanding the CR and translating to CR LF - from another message thread on the optimized AP). I saved it as a text file. I then used notepad to add the astropulse section into my existing app_info replacing the bit in there using the stock app.

I'll go and update one of the machines that trashed its AP work and see if it will behave, although i'll have to wait until it gets some more AP work units.

The libfftw is that the same version from the stock app? As I noted above I didn't replace it as the one from the stock app was the same size but more up to date (August 2008).

I think you'll find that most Windows developers (including Lunatics) supply app_info files with the <CR><LF> convention, so the intermediate Wordpad stage shouldn't be necessary. It's just people whose primary development platform is Linux (and hence use Linux workstations for editing) who add that extra layer of confusion: I've had to point it out to Josh at Berkeley, and Bernd at Einstein (you were active in that thread too, so you may have read it there).

But I wouldn't have expected any version of Wordpad to be 'XML aware' and correct what it thinks to be an XML syntax error (which was the possibility in my mind when I asked about editors). Which takes us back to the first part of the question: who prepared the AP fragment in the first place, and can you remember where you downloaded it from?

Oh - and yes, I think all the copies of libfftw3f-3-1-1a_upx.dll are the same and interchangeable, whatever the datestamp.

I've updated my reply.

You are correct about the CR LF issue, which was discussed over at Einstein, although I was getting the two mixed up (hence the update to my prior reply).

As I have now found out it would appear IE adds the space. The original fragment came out of the archive with the exe and dll files. I then cut and pasted it from IE into notepad. I have since fixed the app_info via notepad and then reopened it IE just to check it.

---

BOINC blog

Thanks, Mark, that explains the conundrum perfectly - I've posted the solution to my bug-reporting thread app_info.xml parsing problem at BOINC Development.

I'm going to have to update my trademark to "Claggy'n'Mark's Space"!

I refuse to give in. I will not install 7zip or any other zip like application. I would love to try this out but the format is not a valid zip file. Please post a real zip file.

Wow. They offer something for free out of their own time and you want to be stubborn and make demands?

AP v4.35 Rev24b54 SSE3
AP v4.35 Rev30b57 SSE

You're lucky I'm nice. ;)

[Edit] Note that my site only has a 2Mbit upload speed (which would be the maximum others can download it; i.e. my upload is your download), so please don't expect super fast times and don't expect it to be available if everyone suddenly decides to download it all at once (assuming more than one person is interested in the .ZIP format).

Sorry, maybe that sounded odd, but the last time I installed 7 zip it really messed up my system. So I downloade another copy from the home site. Same mess. Shoot, considering what winzip did to my 64bit system, I won't install that either. Thank you, for the zip file. Would you like me to convert it to a self installing MSI?
As for the work they did, I think it is nothing short of fantastic. I applaud them for all of thier hard work. Which makes using 7zip even more courious, like frosting a cake with stale jello.

---