[Exit]

Zitat von Andydeluxe

Und warum habe ich für ein WU was auch immer das ist nur 15 Punkte und stingr4y 343 Punkte? Und kann ich das auch irgendwie sehen für was meine CPU eigendlich Rechnet. z.b. für das zählen der O2 Atome oder so?

Weil die halt unterschiedlich gross waren!

Deine hatte 150 Frames - so wie meine auch. Bei der bekam man halt 15 Punkte - siehste ja, dass ich mit meiner ersten WU auch genau diese 15 Punkte bekam.

Meine aktuelle hat wie gesagt 1500 Frames - ist also 10mal so groß - und frisst auch deutlich mehr Zeit.

Die Punkte werden vorher bestimmt.

How do you decide how much credit a work unit is worth? How do you determine how many points a work unit is worth?


Before putting out any new work unit, we benchmark it on a dedicated 2.8GHz Pentium 4 machine with SSE2 disabled (more specifically, as reported by /proc/cpuinfo on linux: vendor_id : GenuineIntel, cpu family : 15, model : 2, model name : Intel(R) Pentium(R) 4 CPU 2.80GHz, stepping : 9, cpu MHz : 2806.438, cache size : 512 KB). This machine runs linux, so all WUs are benchmarked with the linux core.
We plug the results of this into the following formula:
points = 110 * (daysPerWU)
where daysPerWU is the number of days it took to complete the unit. This equation was chosen to match the points for previous Gromacs WUs to the previous point system. The upshot is that Tinker WUs will be worth more than before we set up the new points (i.e. before April 2004).
Please note that the very concept of a reference machine will mean that some WU benchmarking will vary from the performance on your machine. Even between P4s, there are significant differences in architectures over the years. Moreover, variations between FAH WUs can also lead to differences in benchmarking points.
Our goal is consistency within a given definition of a reference machine setup (described above), but beyond that the natural variation from machine to machine and WU to WU will never allow any point system to perfectly reflect what you get on your machine.

Und Deadlines haben die auch noch

How do you set the deadlines for the work units?


Each work unit is benchmarked on a dedicated 2.8 GHz Pentium 4 machine with SSE2 disabled. For most work units (although there may be exceptions, described in the next paragraph), we apply this equation
timeout = 20 * (daysPerWU) + 2 deadline = max(30* (daysPerWU) + 2,10)
where daysPerWU is the number of days it took to complete the unit. The "+2" days is there to give an additional buffer for fast WUs (to allow for servers down, etc). If 30*daysPerWU is less than 10 days, we set the deadline to 10 days, as a minimum time for all projects. The timeout is the time at which the WU is resent to another client and the deadline is the last time which we will give stats credit for the WU.
Occasionally, deadlines may be set shorter or longer than the above calculation indicates, but the reason for having deadlines at all is that the sooner we get back work units, the sooner we can put the results to good use. Also, different projects have different requirements server-side and may require shorter or allow longer deadlines (eg "pfold" calculations can often be run without any deadlines, whereas MREMD calculations work best with very tight deadlines). The assignment server does take machine performance into account in making assignments, thereby allowing slower machines to receive more appropriate work units.