Same here. 10 seconds, restart. 10 seconds, restart.
Running version 7.2.42 (x64) of BOINC, installed as a service.

AMD Phenom CPU.

Same here. 10 seconds, restart. 10 seconds, restart.
Running version 7.2.42 (x64) of BOINC, installed as a service.

AMD Phenom CPU.

"Installed as a service" is the key here. Bingo. First direct confirmation that the application was compiled against the buggy BOINC API that was in circulation from July to November last year. If you can, at least temporarily, re-install that version of BOINC in user mode, the model should run normally - and provide the proof to confirm the diagnosis.

If you can, at least temporarily, re-install that version of BOINC in user mode, the model should run normally - and provide the proof to confirm the diagnosis.

I'm running in user mode (v7.2.42) (for vLHC and ATLAS) and half my short WUs are erroring out

There have been reports that some anti-virus programs have been interfering with the successful download of that file. Check in your anti-virus or anti-malware console that nothing is requesting your permission to download the file anyway. We are as certain as we can be that any virus/malware warning on that particular file is a false alarm, but proceed at your own risk.

And this morning, the batch of short models seems to be finished so except for the re-issues which presumably will have a greater percentage of failures this one can have a rest till we find what the next batch brings.

Well I don't know if this will help anyone or not. I am running Windows 8.1 and this probably applies to Windows 8 as well (maybe even Vista). What I have discovered for my own installation is that .net framework 3.5 was needed in order to solve my problems with work units crashing after 10 sec or immediately upon startup (seems to have cleared up many computation errors as well).

Hope this is of help to others.

.net framework 3.5 can be found here

After installation run Microsoft update to get the latest patches.

An article in the Microsoft database deals with installation of .net framework 3.5 on Windows 8 and 8.1 and can be found here.

Unfortunately there seems to be a bug in the Microsoft code that misses catching applications that you might install that require .net framework 3.5, so don't rely upon Windows to inform you that you need to install it on demand. You are better to just download it and install manually.

Lastly don't forget to reboot to activate any changes (.net framework 3.5 will automatically be enabled upon installation, no need to enable it through Control Panel).

Note to the Moderators: maybe this should be a sticky (but then again perhaps it needs to be bundled with Boinc which is not our problem).

---

6,000?? Give it a rest.

G�bekli Tepe is more than 10,000 years old. And quite intricate I might add.

Explain that!

Upon further reading at Microsoft this is only an issue with Windows 8 and 8.1. It seems that .net framework 3.5 was excluded from both but are included in previous versions right up to Windows 7.

The bottom line is this: you need .net framework 3.5 installed if you are running Windows 8 or 8.1. In this case check through {Control Panel>Programs and Features>Turn Windows features on or off} to verify that it is installed and active.


Have fun!!

---

6,000?? Give it a rest.

G�bekli Tepe is more than 10,000 years old. And quite intricate I might add.

Explain that!

And as a heads up for anyone who might happen to be running Windows 10 Technical Preview, .net framework 3.5 is not included either but can be enabled and downloaded through use of the Control Panel feature as described in my previous post.

---

6,000?? Give it a rest.

G�bekli Tepe is more than 10,000 years old. And quite intricate I might add.

Explain that!

As a short update. I have solved the runtime problem but computation errors still occur on the CM3 shorts. Mind you I have not had one of these actually crash after 10 sec but rather computation errors in the 20 min range. I believe the computation errors for the CM3 shorts have been explained adequately elsewhere in this forum.

---

6,000?? Give it a rest.

G�bekli Tepe is more than 10,000 years old. And quite intricate I might add.

Explain that!