Why does Einstein pushes much more data than I can process timely?
23 Jan 2014 22:37:15 UTC
Topic 197353
(moderation:
Today I got 2 WU's from Einstein for 56 hrs and 126 hrs to be finished before 30-01 and 06-02, which means an average CPU-time of over 6 resp. 9 hrs/day. As I don't use my private laptop for so many hours a day it will be impossible to finish the WU's timely. Besides this I see more and more that the estimated run-times are much bigger then estimated making it even more difficult to finish the WU's timely. Why does Einstein push so much data to my laptop? From the past it should be very easy to estimate which load my laptop can handle. Why not push more smaller WU's of let's say 10 hrs/piece? Rosetta works like this and never ever caused the problems I see now again!
P.S.: today I finished PA0046_08691_240_1 with almost 100 hrs of CPU-time and 0 credits - too late. As you'll understand this WoW is not making me happy at all!
Why does Einstein pushes much more data than I can process timel
)
First: it's you BOINC which decides for how much work to ask. And then the Einstein server decides what to send you to best match that request.
Looking at the tasks of your Core 2 Duo laptop I can see that "PA0046_08691_240_1" was a CUDA WU, i.e. processed by your GPU. The 9200M GS is the absolute bottom line of GPUs, it has 8 shaders of an old architecture, whereas current models have a few 100 to a few 1000's. I don't think it makes any sense to run Einstein tasks on this.
After that WU was returned you were senT another one for the GPU. Your BOINC asked for work, because the GPU was running dry, and Einstein couldn't send you anything less than a single WU. That's why your GPU will be overloaded again by this WU.
The CPU should fare better, but you might want to switch this host to a project with smaller WUs nevertheless. Your i5 should be fine, though.
MrS
Scanning for our furry friends since Jan 2002