client error?

stargazer8913
stargazer8913
Joined: 24 Feb 05
Posts: 3
Credit: 5493
RAC: 0
Topic 192819

Recently my work units have been turning up with client errors. Any ideas as to why this is?

luis vieira
luis vieira
Joined: 20 Apr 07
Posts: 1
Credit: 28585443
RAC: 0

client error?

Quote:
Recently my work units have been turning up with client errors. Any ideas as to why this is?


Dan Butterfield
Dan Butterfield
Joined: 4 Mar 05
Posts: 3
Credit: 214288
RAC: 0

RE: Recently my work units

Quote:
Recently my work units have been turning up with client errors. Any ideas as to why this is?

My guess would be E@H. Since the larger work units came out this seems to be a common problem on my 3 machines. After doing almost 60 hours work on one unit
it goes to Client error @ about 99% done. If I get one more error E@H is done on my Machines.

Brian Silvers
Brian Silvers
Joined: 26 Aug 05
Posts: 772
Credit: 282700
RAC: 0

RE: RE: Recently my work

Message 68009 in response to message 68008

Quote:
Quote:
Recently my work units have been turning up with client errors. Any ideas as to why this is?

My guess would be E@H. Since the larger work units came out this seems to be a common problem on my 3 machines. After doing almost 60 hours work on one unit
it goes to Client error @ about 99% done. If I get one more error E@H is done on my Machines.

@ Dan Butterfield -

For your Pentium 2.8GHz machine, my suggestion would be for you to do a reset project. This will delete any current datapack that you have and the science application. A new Windows science application (4.24) should automatically download and you'll download a new set of data to work on. Version 4.24 has the following changes:

  • * referencing a symbol store on the E@H server (should work now). This means that in case of a (debuggable) client error a debugger will be loaded (by newer BOINC Clients) that will in turn contact the Einstein@Home server in order to download debugging symbols ("phone home"). This means that the PDB is no longer distributed with the App, and symbol information will be downloaded compressed and only when needed.
    *

fixed a bug related to uninitialized memory
* more verbose output in case of errors
* small improvements to the checkpointing
* a diffrent function is used for rounding in sin/cos calculation (ftol() instead of modf()), which should avoid the broken CPU feature detection and give some speedup on AMD and non-SSE2 Intel CPUs.

The two items I made bold may help in your case.

Brian

Brian Silvers
Brian Silvers
Joined: 26 Aug 05
Posts: 772
Credit: 282700
RAC: 0

RE: Recently my work units

Quote:
Recently my work units have been turning up with client errors. Any ideas as to why this is?

@Stargazer -

I'd suggest the same for you (to do a reset project), if you still have problems after the current result you're working on... See my comments to Dan...

Brian

Bikeman (Heinz-Bernd Eggenstein)
Bikeman (Heinz-...
Moderator
Joined: 28 Aug 06
Posts: 3522
Credit: 738052318
RAC: 1310939

RE: RE: Recently my work

Message 68011 in response to message 68010

Quote:
Quote:
Recently my work units have been turning up with client errors. Any ideas as to why this is?

@Stargazer -

I'd suggest the same for you (to do a reset project), if you still have problems after the current result you're working on... See my comments to Dan...

Brian

The new shopuld be distributed to you automatically the next time BOINC syncs with the server. Even if the new app doesn't fix the error, it has advanced debugging features which might help the team to identify and fix the error.

CU

BRM

stargazer8913
stargazer8913
Joined: 24 Feb 05
Posts: 3
Credit: 5493
RAC: 0

The client errors are still

The client errors are still happening with every work unit I receive, this is the only project its occurring with. Any other suggestions?

Ananas
Ananas
Joined: 22 Jan 05
Posts: 272
Credit: 2500681
RAC: 0

Did that problem start with

Did that problem start with Vista or did you already crunch successful with Vista?

If I recall right, 0x40010004 can be some problem with the graphics driver

stargazer8913
stargazer8913
Joined: 24 Feb 05
Posts: 3
Credit: 5493
RAC: 0

I've never run it on any OS

I've never run it on any OS other than vista.

Dan Butterfield
Dan Butterfield
Joined: 4 Mar 05
Posts: 3
Credit: 214288
RAC: 0

RE: RE: RE: Recently my

Message 68015 in response to message 68009

Quote:
Quote:
Quote:
Recently my work units have been turning up with client errors. Any ideas as to why this is?

My guess would be E@H. Since the larger work units came out this seems to be a common problem on my 3 machines. After doing almost 60 hours work on one unit
it goes to Client error @ about 99% done. If I get one more error E@H is done on my Machines.

@ Dan Butterfield -

For your Pentium 2.8GHz machine, my suggestion would be for you to do a reset project. This will delete any current datapack that you have and the science application. A new Windows science application (4.24) should automatically download and you'll download a new set of data to work on. Version 4.24 has the following changes:

  • * referencing a symbol store on the E@H server (should work now). This means that in case of a (debuggable) client error a debugger will be loaded (by newer BOINC Clients) that will in turn contact the Einstein@Home server in order to download debugging symbols ("phone home"). This means that the PDB is no longer distributed with the App, and symbol information will be downloaded compressed and only when needed.
    *
fixed a bug related to uninitialized memory
* more verbose output in case of errors
* small improvements to the checkpointing
* a diffrent function is used for rounding in sin/cos calculation (ftol() instead of modf()), which should avoid the broken CPU feature detection and give some speedup on AMD and non-SSE2 Intel CPUs.

The two items I made bold may help in your case.

Brian


Dan Butterfield
Dan Butterfield
Joined: 4 Mar 05
Posts: 3
Credit: 214288
RAC: 0

RE: RE: RE: Recently my

Message 68016 in response to message 68009

Quote:
Quote:
Quote:
Recently my work units have been turning up with client errors. Any ideas as to why this is?

My guess would be E@H. Since the larger work units came out this seems to be a common problem on my 3 machines. After doing almost 60 hours work on one unit
it goes to Client error @ about 99% done. If I get one more error E@H is done on my Machines.

@ Dan Butterfield -

For your Pentium 2.8GHz machine, my suggestion would be for you to do a reset project. This will delete any current datapack that you have and the science application. A new Windows science application (4.24) should automatically download and you'll download a new set of data to work on. Version 4.24 has the following changes:

  • * referencing a symbol store on the E@H server (should work now). This means that in case of a (debuggable) client error a debugger will be loaded (by newer BOINC Clients) that will in turn contact the Einstein@Home server in order to download debugging symbols ("phone home"). This means that the PDB is no longer distributed with the App, and symbol information will be downloaded compressed and only when needed.
    *
fixed a bug related to uninitialized memory
* more verbose output in case of errors
* small improvements to the checkpointing
* a diffrent function is used for rounding in sin/cos calculation (ftol() instead of modf()), which should avoid the broken CPU feature detection and give some speedup on AMD and non-SSE2 Intel CPUs.

The two items I made bold may help in your case.

Brian

Brian I did what you suggested and the very next day it happened again. I ran through the routine again and now all seems to be ok. Thanks Dan

Comment viewing options

Select your preferred way to display the comments and click "Save settings" to activate your changes.