Chance of WU Corruption?

[compdewd]

Hello everyone!

I've been wondering how great the chances are of causing miscalculations of work by running other computer processes? And if this happens and the work is submitted to the server and the next work unit is created from it, is there a way to catch this before more incorrect work goes out and causes a completely incorrect end simulation? Will the original unit not upload if there is a slight miscalculation? I know that some corrupt units don't upload, but I assume that that is because there is a corruption in the file format rather than with the work itself...

Just a little curious lol. Thanks in advance for your answers!

[Jesse_V]

Don't worry about it. Run your other computer processes at will, there's no risk of contaminating the WU. Errors occur for a variety of reasons, including heat, excessive overclocking, unsupported hardware/drivers, or the WU simple leads itself into an impossible atomistic situation. It may have little to do with a file format. There is built-in error checking, so if something goes wrong, the error is caught and properly handled. In some cases, the servers can use several donors to try to complete an errored-out WU, in the hopes that one of them will be able to successfully complete it. If not, then it's marked as bad and a script pulls it from distribution. Then the PG scientist in charge of the WU's project will try to fix the problem. The whole system is pretty stable, you don't need to worry.

[compdewd]

Okay thanks!