Print

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

[Jaqie Fox]

Perhaps it would be a good idea for us users/fans to get a good DF benchmarking set up for systems, so there is a definitive imperical reference for what kind of computer parts will best make DF perform better.

good idea/bad idea?

I'm willing to do most of the work myself but I would need *some* help from others that have systems I don't have access to.

[TotalPigeon]

Sounds good to me. Someone needs to sort out some way for us to compare how DF runs on each pc. I'm thinking we pick some fort as a standard, host it somewhere and get people to run it in a clean setup of DF. Then they post their average framerate on the fort, lowest framerate when filling the forts moat/farm/world (I guess we'd need to pick a fort with some sort of water/lava feature) and the average framerate at the start of a new fort.

I can't really think of any other good ways to screw with DF and get a decent indication of how the pc is handling it... at a push, maybe timing how long the game stops for after a really big bridge is built. But thats kind of lame.

So, which fort to use?

I seem to recall Boatmurdered having a "flood the world with lava" device...

[Jaqie Fox]

the fort isnt the problem. getting an imperical set of measurements is.

Using FRAPS is one of the things that would be perfect for this - you can record a graph of the framerate of an app - we would need to take out all possibility for random changes to interfere with good results... I am hoping that the record/playback system takes care of that, and this is all stuff I can find out myself first hand, but once Ive got a method of getting reliable benchmark results I will simply need a wider sample of hardware to run it on then I have personal access to.

[Elshar]

If you guys can come up with some sort of universal test package, I can test it. I have access to everything from 486's (Sometimes, when people want to get rid of their really old machines) to multi-cpu systems (I have a dual dual-core Xeon sitting on my bench at work atm, along with a dual core Opteron, and random P4s).

I think possibly a few good tests would be something like:

- A new fort undug in various climates (Tundra/none/none, Tropical/Heavily/Thick etc)

- A typical 200-dwarf fort doing normal stuff

- A fort under siege

- A fort with permaflood

Y'know, a few well setup test scenarios.

[I3erent]

I think engraving could be a good test my rig slows miserably when i have a bunch of engravers goin.... 2.7 p4 nvidea 5200 512 rambus ram...