User of the day

News

Server upgrade.
Aug 29, 2010
I've just moved our WWW server (including BOINC) to a new virtual machine. It is possible that something was overlooked in the process. Please tell, if some features of the site do not work anymore. Otherwise I'm back from vacations and soon will start processing the results of our last run.
Another batch, now with anti-aliasing.
May 13, 2010
The previous run have reproduced the map of ground states in circular nano-cylinder in agreement with theory (which means that the energy is calculated correctly), but its precision was not enough to obtain the map of metastable states. This is because tracing of metastable states involves taking derivatives, which amplifies the errors. To reduce the jitter, associated with assignment of constant charges to finite elements, I have introduced a Monte-Carlo anti-aliasing procedure to calculate the average density of charges across each finite element. I hope this will be sufficient to reduce the jitter. We'll see. If not, there are other, more intensive methods.
New batch of workunits.
November 24, 2008
After finishing processing the previous batch I have launched a new one today. New workunits require about 330 Mb of RAM to complete, which should not be that much. A summary of the preliminary results of the previous run is here.
Planned downtime
August 1, 2008
The project server will be down (or working erratically) during the coming weekend (August 1-2) due to computer hardware upgrade.
New smaller WUs
July 23, 2008
To improve precision on one of the critical lines (clumped in-between of a very few grid points) a new barrage of WUs is dispatched, increasing the resolution on calculated maps. These new points are going to complement the results of our fist round and use the same (lower) number of finite elements. The memory consumption of new WUs is low (but there is no credit bonus as in memory-demanding WUs, just pure FLOPS).
...more

News is available as an RSS feed .