> Recently I've had several workunits abort with a message such as this:
>
> Aborting result H1_0267.9__0268.1_0.1_T08_Test02_2: exceeded memory limit
> 40000000.000000
>
> This can be seen at http://einstein.phys.uwm.edu//task/1094409
> Note that was from the newest application version for Linux - 4.76. Others
> that had the same error under the earlier version (4.73 I belive) are:
> http://einstein.phys.uwm.edu//task/1093263
> http://einstein.phys.uwm.edu//task/1093262
> http://einstein.phys.uwm.edu//task/1092755
> http://einstein.phys.uwm.edu//task/1089500
> http://einstein.phys.uwm.edu//task/1076183
> http://einstein.phys.uwm.edu//task/1067551
> http://einstein.phys.uwm.edu//task/1062249
>
>
> I didn't see any other reports of this kind of error.
>
> I do compile my own Linux Boinc core client so that I have an optimized
> version. Don't know if this would cause the EAH application to produce this
> error or not.
>
> Maybe someone at UWM wants to take a look. I'll be glad to supply other data
> as needed.
Charlie,
This was a bad oversight on my part. The memory requirements of the new 'coincidence' code are greater than that of the old code, and even there I hadn't left enough headroom. I've bumped the memory requirements up to 70MB, which I think should now be sufficient.
Bruce
