The largest file in the download is shared by all workunits within a 'batch' so you already had this file on your computer and did not need to download it again.

We are looking at the errors that have been returned now.

Yeah, graphics are omitted in the BETA tests. You should still get a progress bar on the BETA that will progress. Are you seeing this?

-Uplinger

I posted and deleted in the Dengue forum:
The problem isn't the Dengue task, it is task BETA_ach1_2_14_9 using beta2 version 513.

I isolated the problem by allowing no new tasks and suspending jobs in queue. Here are some sequential screenshots:

http://img463.imageshack.us/img463/7641/dengueex9.jpg

http://img472.imageshack.us/img472/8648/dengue2ao7.jpg

http://img482.imageshack.us/img482/5374/dengue3dk0.jpg <---all heck broke loose here when writing checkpoint files

http://img482.imageshack.us/img482/5945/dengue4xd6.jpg

http://img463.imageshack.us/img463/5475/dengue5tx6.jpg <---so bad that CPU utilization is less than 100%

The task is currently at 90% complete. The behavior is cyclic. At its worst, the computer is unresponsive.

[edit] To help it along, I've chosen to use only one processor as shown below.

http://img472.imageshack.us/img472/3861/11790936br9.jpg [/edit]

Vester, not sure, but multiple processes like crunching and DDDT+ACAH simultaneous with only 1gb ram will tax a system. The System Requirements for AC@H are steep (hook up to a future CPDN project and you'll discover what really steep is!).

I'm testing on a 6 year old HT machine which allows only 1 thread and the job is running perfectly fine, with barely any interference, but that OOS Calc is slow and now extra slow (cause it seems to use alot of disk i/o(?)

I've allowed BOINC to only write once per 600 seconds to cut on the disk i/o. Checkpoints started out with 25 minutes apart and advanced so far, 60% into the job to 19 minutes apart. Peaks RAM read sofar were 541mb and 250mb VM. The amount of data read/write with such big models are to be expected large and so they are.

Added: If reading the log right, it took 5 minutes to upload your result. Not as bad as i thought to read elsewhere. WCG must be tweaking, or your upload bandwidth is good. I'll know how it fares in a few hours from the outpost.

My recommendation is to get extra 1gb. Even in the extortionate pricey realms I live, saw last week a 1gb 667 stick did 79,90 Euro, down from 199 Euro 2 months ago. Tempting as going from 1 to 1.5gb took already most of the performance issues away. A C2D running here also has 1.5gb running 2 threads and never/very rarely has any sluggishness.

The one thing for as long as i can remember are touch-pads on laptops, they are the one element affected by anything busy. For the finer/tense 'art'work, i just do a quick ThreadMaster tweak to a lower percent. There's a 'how to' in the Start Here forum for that and works a treat. I know why I'm doing this, so accept that occasional freeze and jump. WCG want the idle / spare cycles which at 60% are just as precious as full throttle.

I've got a Vista with C2D and 4GB of RAM - it runs smooth as silk. I wouldn't run it with anything less than 3GB though (tried 2GB and it was sluggish).