Print

Please login or register.
1 Hour 1 Day 1 Week 1 Month Forever
Login with username, password and session length

[Diamond]

Basically, this
http://store.steampowered.com/app/384300/

For single-threaded games: Before the application may have had 70-90% access to a CPU processor, it now has near 100% access to a processor. This enables your game to take full advantage of your CPU core, thus enabling it to reach it's highest FPS potential on your system. The net benefit can be up to 30% more CPU power for that game!

I wonder if this can work for DF (probably possible through adding DF to Steam as non-Steam game) and what sort of performance increase there would be. Can someone perhaps test it?
(I am not promoting it, just curious if it will be worth the price)

[dorf]

From what I've read what the utility does, you can replicate this behaviour by putting DF on one core and everything else on others.
Set DF's process priority on higher settings and other non-OS  or non-critical processes to lower settings.
Minimize (or even better: close) any and all programs apart from DF.
Temporarily stop non-critical system services.

This will give you a similar experience as the utility barring Steam integration and some automation.

Which means you won't see a difference in FPS.
But you can try it for yourself. Steam is good for refunds, I hear?

[Admiral Obvious]

I don't usually notice much of a difference.

I've put DF onto a core of its own, and only noticed a very marginal FPS increase on a public PC that has a relatively old dual core in it, after I migrate the OS over to core one. Average FPS went from around 40 to about 50.

I tried the same again at home with my (almost) gaming PC, and didn't notice any FPS change at all, but I've not tweaked the max FPS values, nor have I run a fort for much more than 5 in game years before something !FUN! happens that causes it to crumble into failure.

[Miuramir]

(I am not promoting it, just curious if it will be worth the price)

That product has several issues:

* Very little (if any) of what it does can't be done manually for free.  Whether "paying them to avoid learning a bit more about how computers work" is a good plan is a matter of opinion. 

* It doesn't recognize the way modern processor scheduling works, and how it in particular interacts with modern cooling. 

* The advertising leans heavily on easily-misunderstood hype rather than actual performance. 

My quick evaluation, without paying for it, is that on a particularly poorly set up PC from several generations back, it might help a bit; but not in any way you couldn't do for free.  On a merely slightly older PC that you're multitasking a lot (watching video, posting on forums, playing DF, etc. all at once) it may help a bit.  On a decent modern PC, there's a real chance that it will actually *slow down* operation, as it keeps the scheduler from doing proper thermal cycling.  (The limiting factor on most modern cores is not ultimate max speed, but thermal limits.) 

[Putnam]

all software that promises better performance in other non-specific software is bullshit

[Grimlocke]

Its selling thing we already have, with a helping of snake oil.

On core isolation in general, cooling shouldn't be an actual issue since you processor should be able to run full power on all cores for longer periods of time without cooking itself, always. If it cannot do that then its a bad overclock or possibly time to clean some dust out of the heatsink.

You can usually gain more performance by increasing the process priority, particularly on a processor with a high number of cores. Your OS may not always recognize that DF is a high priority task when for instance, one out of eight cores are busy. It would constitutes only 12.5% load, causing the OS to think your not doing anything intensive and run the CPU is low power mode (meaning, lower clock speeds).

...

So no, no need to spend perfectly good money for a thing you can already do, which likely won't really help anyway.