View Single Post
Old 09-11-2012, 05:57 PM   #17
Deago

Loremaster
Deago's Avatar
 
Join Date: Jul 2008
Posts: 625
Default

Wingrider01 wrote:

Ulrichvon wrote:

Wingrider01 wrote:

Is Very significant backed by controlled testing or just by the TSMB method? Would love to see the comparision graphs and data points

That would depend explicitly on your cpu and the turbocore settings associated with it.

For my machine, when forced to a 'sane' affinity setting, eq2 benefits from another 500mhz of processing power, which equates to about 13 extra FPS while raiding with my client settings.

The difference between 7-8 FPS when locked to core 6, and 19-23 FPS when locked to core 2, is huge IMO.

Now if the turbocore jump on your rig is less, the difference would be less.

All in all, SOE doesn't need to do anything with this change OTHER than remove their over-riding of OS and USER affinity settings.  Over-riding the system and the user is a bad idea to start with and should have never been done that way in the first place.

A few of us tried to get SoE to understand that back when this change was made, but they wouldn't listen then either.  Now with these newer 'turbo core' CPU's the issue is a larger concern and affects people other than just multi-boxers.

running a Core I7 3770K 3.5GHZ

still want to see the data points that you was using to justify the statement, will take them from you configuration. No offense but I take no performence comment at face value unless it is proven - my programmers know this already and will not make statements they do not have the data points to back it up withh

As a software engineer Wing made me lol because I can relate to his feelings on the matter.

I will say that 90% of the time the OS is smarter with affinity.  So I agree with wing that data would need to be summoned for it to be taken seriously. 

__________________
Hello fellow EQ2 gamers!
Deago is offline   Reply With Quote