• Some users have recently had their accounts hijacked. It seems that the now defunct EVGA forums might have compromised your password there and seems many are using the same PW here. We would suggest you UPDATE YOUR PASSWORD and TURN ON 2FA for your account here to further secure it. None of the compromised accounts had 2FA turned on.
    Once you have enabled 2FA, your account will be updated soon to show a badge, letting other members know that you use 2FA to protect your account. This should be beneficial for everyone that uses FSFT.

Benchmark utility

N

Nathan_P

[H]ard DCOTM x3
Joined
Mar 2, 2010
Messages
3,522
There has been quite a debate going on over at FF about -bigadv and how they are fairly allocated to machines that will run them quickly. As you may well be aware there are machines out there running -bigadv that probably should not be whilst other machines that can do so comfortably end up with normal SMP.

A lot of the debate has centered on how to fairly assess each individual machine as to its suitablility for -bigadv/-bigbeta work. My suggestion was based on actual core count but most others agree that some sort of benchmark should be run before each WU and in retrospect I agree that this is probably best.

Therefore my question is would Haitch and the rest of the benchmark team consider the possibility of your benchmark utility being put forward as a possible option for use in such a role?
 
Another thought: could something similar to HFM's TPF history be implemented in the official client? That way, if there's more requests for a certain type of unit than units available, the servers could query for that machine's previous performance on that or similar units to determine which machines should receive them, and if there's plenty of units available, the current system seems to work fine.

Just my .02
 
Seems like the necessary data for such a thing should already be available given that the Machine ID of the client requesting a WU is known to the AS coupled with the fact that the client will calculate a performance fraction based on the percent of time to deadline remaining at the end of each unit. Using the average performance fractions would also avoid the pitfall of an average TPF calculation that does not take into account machines that do not fold at 100% 24/7. I don't know what the PF data is used for though, if anything, and whether or not it is available to the AS or if it is only available locally. Nevertheless, the capacity for implementing unit assignments based on a performance meritocracy appears to already exist.
 
the problem is convincing PG to do it and quite frankly it takes them years to get anything done on the software side. while i'd like to see it happen, we all know it never will.
 
You must log in or register to reply here.
Back
Top