This week we started with some delay a number of new runs with the objective of fitting model parameters of a new type of transmission models. These for the first time allow for heterogeneity in simulated transmission. This will be important to get an accurate description of areas with relatively low malaria transmission intensities.
A consequence of this is that we will need slightly more computing power in the future. We will therefore enable account creation next Monday to allow more users to participate.
In addition, we have optimized the configuration of the scheduler some time ago. It now does a good job at choosing a reliable host (one with a short turnaround time and a high success rate) when resending a result because of an error or timeout. This significantly reduces the number of results that have to be sent a third time, and improves the performance of our optimization algorithm. We can therefore make much better use of the available computing power.
____________
Alain Studer
Swiss Tropical Institute
/ 
The initial runs sure are made up of loooong wu\'s. :)
Live long and BOINC!
____________
Paul
(S@H1 8888)
/ 
The initial runs sure are made up of loooong wu\'s. :)
Live long and BOINC!
Yes the \'62\' units are VERY long! I used to think the \'42\' units were long, ha was I fooled! I am currently at 5hours and 35 minutes and in the upper 99% range on my nearly fastest machine, a 3.2ghz dual core with HT on each core.
____________
/