I've noticed more 'compute errors' in the last week...on a daily basis, on these types of wu's :

HIVPR
GIANNI_pYIpYVkp01
IBUCH

Running Linux64
boinc 6.4.5
177.80 driver

---

Consciousness: That annoying time between naps......

Experience is a wonderful thing: it enables you to recognize a mistake every time you repeat it.

Your tasks show the message:

<message>
process got signal 11
</message>

Anyone knows what this means?

MrS

---

Scanning for our furry friends since Jan 2002

Its a general system wide error message, not BOINC specific, you can get it using DB's for example (MySql etc). Its hard to know what the absolute cause is because its a general error message. It could be because you hit a bug, its possible that a binary or one of the libraries it was linked to is corrupted, badly built, or misconfigured. It could have lost its link to the Project because of that, and fallen over. The error can also be caused by failing hardware. In general terms its indicating that some Link information needed to continue has been lost - and that can happen for a multitude of reasons.

In other threads I have seen over time on this one, general concensous has been - after all the usual sensible checks for "standard" problems - is if it still is repeating the error "signal 11", run the cache dry and do a reset, in the final analysis, detach / reattach which generates new files thus getting over the corruption if it was software based. Clearly that will not resolve it if hardware based, however it seems by-in-large to surface as a software problem.

Sledgehammer to crack a nut approach, however I have seen no other solution to this issue when its surfaced elsewhere.

A real hairy example of whats behind this (UNIX but this applies to other systems) is at signal 11 core explanations

Personally I have no idea what half that link means, however someone out there might and enlighten us all :)

Regards
Zy

Memory segmentation error This may be caused by bad compiles, bad memory, kernal panics, and the most common OC (if they are randomly distributed) ...

If it is a bad compile all tasks will fail at the same point on all systems everywhere.

Bad memory will usually be clustered in the same arena but will affect all projects ...

Kernal panics are rare, but caused by OS errors...

Overclocking ... well, enough said.

I had looked around a bit before posting. And I found similar to what you are mentioning Zydor. But, fresh eyes on an issue sometimes help.

I spent some time looking while I was letting the cache run down. The solution came down to detach, memory issue or lib32 issues, not sure about the last one.....

So, I detached, and will see what happens...and keep the sledge hammer next to the bench...just in case..;)

---

Consciousness: That annoying time between naps......

Experience is a wonderful thing: it enables you to recognize a mistake every time you repeat it.

I had a lot of error initially as well.
I found out that when the wife would switch to her side of the computer I immediately would get a "computer error"
SO I change the BOINC settings in allowing work to be done ONLY on my side and I disabled the "do GPU while computer is in use" and I did not get other errors.
I havent been crunching lately (waiting for a server) GPUGRID but give this a try.

Martin

Thanks for the tip. I actually asked my wife if she had been trying to use that computer for something...Not that she would be using it, but I had to ask. Besides, it's in the dungeon, surrounded by barbed wire, claymores and a pair of guard wraiths...

It is a 24/7 cruncher. And is not used for much except to check the status of wu's, forum and email in the evening.

The detach trick did not work. So I am shutting it down until this weekend. I'll try a driver update and see what that does.

---

Consciousness: That annoying time between naps......

Experience is a wonderful thing: it enables you to recognize a mistake every time you repeat it.

Thanks for the tip. I actually asked my wife if she had been trying to use that computer for something...Not that she would be using it, but I had to ask. Besides, it's in the dungeon, surrounded by barbed wire, claymores and a pair of guard wraiths...

It is a 24/7 cruncher. And is not used for much except to check the status of wu's, forum and email in the evening.

The detach trick did not work. So I am shutting it down until this weekend. I'll try a driver update and see what that does.

If you remote into it that can also cause problems. Remote desktop and user switch ... bad moon rising ...

I've been getting compute errors on files of type 182-IBUCH_KID_shao_ba1-1-100-RND0470_1 only. This has happened on two separate machines. One with 32 bit Vista, 8800GT (O/C), client 6.6.20 & 182.50 driver. The other with 64 bit Vista, 9800 GX2 (Not O/C), client 6.6.20 & 182.50 driver. Any suggestions ?

I powered up my xp64 box to take over for the buggy Linux64 machine. Same GPU and no problems.

---

Consciousness: That annoying time between naps......

Experience is a wonderful thing: it enables you to recognize a mistake every time you repeat it.