Print

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

[Beeskee]

~30 FPS on a dual core 2.2 ghz cpu and 3gb of ram.

No settings were changed.

[Nolor]

20FPS on a 3.2Ghz Pentium IV.
2 gigs of ram, Windows 7 32-bit.
Radeon 3870, 512MB. Though my processor's clearly the bottleneck here.

[diriel]

Asus Rampage Formula Mobo
Q6600 OC/d to 3.01
8 gigs of 1066 corsair
2x WD 640 gig hdd
ATI 3870
Onboard sound solution.

1 minute test=40fps +/- 3fps fluctuation.
I could only imagine 150 dwarves / 5-7 years in.

Have FUN
Edit: Windows XP 64.

[Vigilant]

Are you serious? Isn't DF all about CPU power? I mean, will there be an FPS difference if you play with a GTX280 instead of a 9800GT?
While we are at it, what are the best init settings for an average system? [4Gb RAM, 3Ghz dual core, GTX260, OS: Vista x64]

Also, newer processors are just well... better a lot of the time. The clock speed may be the same, but the number of clock cycles it takes to perform an instruction can vary. Take something simple like adding. CPU 1 finishes an add in 5 cycles, CPU 2 finishes an add in 2 cycles. Even if they have the same clock speed CPU 2 will be much faster.

DF doesn't need processing power, it needs a cpu that can move bytes around quickly. On a decent fort it is doing a scatter read/write on almost 500 megs of ram. Newer processors just happen to have better memory access.

You missunderstand. Pathfinding is series of instructions and checks. Clock speed is more instructions being able to be executed in a certain timespan yes. But CPI (Cycles per Instruction) represent how long a certain instruction TAKES. Newer processors tend to have slightly better CPI, even if their clock speed is around the same as older ones. We've actually hit a bit of a ceiling on clock speed due to power and transmission concerns, which is why newer CPU's are around the same clock speed. So faced with this problem companies have focused on making parallel processors and improving the architexture, the latter leading to better CPI.

[devek]

I don't misunderstand at all lol. Modern CPUs on mundane tasks spend very little time actually processing and most of their time simply waiting on data.

It still takes time to take a byte of memory from ram and put it into your cpu, and I stand by my opinion that DF isn't cpu bound. It spends more time collecting the data to path, than it does pathing. Sorry.

[Vigilant]

I don't misunderstand at all lol. Modern CPUs on mundane tasks spend very little time actually processing and most of their time simply waiting on data.

It still takes time to take a byte of memory from ram and put it into your cpu, and I stand by my opinion that DF isn't cpu bound. It spends more time collecting the data to path, than it does pathing. Sorry.

There is that. The pipeline is probably full of stalls during the LW/SW's, and I can't see pathfinding having large benifits from caching sadly with each dwarf on a different area of the map trying to go somewhere.

[blazzano]

I tried it on a computer from around the turn of the millennium (and coincidentally running Windows Millennium Edition):

850 mhz Athlon Thunderbird
384 MB of PC100 RAM
32MB video card (Nvidia TNT2)

The result?  DNF.    It took an hour to load the save, and another hour passed between me unpausing the game and running out of time to actually see the gameplay FPS - had to Ctrl-Alt-Del it.  At that point it hadn't yet been able to render a new frame.  Despite the incredible slowness of the CPU, the real limit for a map like that is the lack of memory; when the size of the map greatly exceeds the available physical memory, the hard drive needs to be accessed constantly.  Not light hard drive access either, but the sort that makes the entire system unresponsive.  I feel like trying it again later, but ultimately I suspect that it would achieve some small fraction of a FPS in practice.

[Tormy]

I don't misunderstand at all lol. Modern CPUs on mundane tasks spend very little time actually processing and most of their time simply waiting on data.

It still takes time to take a byte of memory from ram and put it into your cpu, and I stand by my opinion that DF isn't cpu bound. It spends more time collecting the data to path, than it does pathing. Sorry.

Either way one thing is for sure. I am going to buy an i7 in this year. I thought that it's pointless to buy these quad core processors for gaming, but it looks like that I was very wrong. 

[axus]

I'd think that Antivirus software would have more of a performance impact than a virus

[Baughn]

That does tend to be the case, yes. Viruses try to stay hidden..

In all seriousness, the #1 performance problem I see - for DF of anything else - is antivirus software. If you ask me, just do without. Common sense (and a lack of IE) will work just as well.

[devek]

I use eset nod32, and I challenge you to show how that program would slow down DF

It is the fastest anti-virus on the market, by far, and it is always in the top ranking in virus detection.

Norton is the slowest and worst anti-virus on the market, even compared to actual viruses.

[Rask]

19 FPS here on the laptop

Athlon X2 DualCore QL-66, 2.2 GHZ
2 GB RAM
Windows 7

[Thief^]

I use eset nod32, and I challenge you to show how that program would slow down DF

It is the fastest anti-virus on the market, by far, and it is always in the top ranking in virus detection.

Norton is the slowest and worst anti-virus on the market, even compared to actual viruses.

Some viruses have been known to clean other viruses and install windows patches. e.g.. I've never seen Norton do either (ok that was a bit harsh). I have a similar opinion of McAffee and AVG. IMO Nod32 is great, and Avast's free home version is pretty good too.


I'm tempted to run this test on my Win7 eeepc...
Brings a new meaning to "low fps".

[Hakar]

My CPU: AMD Athlon 64 X2 4800+
My FPS: 12-20

[dree12]

This isn't an accurate test. I suggest that people use the exact same folder, post their OS, kill all non-needed processes (including explorer.exe on windows), and post either their Windows preformance scores (which are surprisingly accurate) or their memory+speed, CPU model + clock speed + core number, graphics+sound card model and GPU. We can determine better results that way. DF isn't just a CPU task like many suggest it is. Why does decreasing G_FPS_CAP increase speed? Why does turning MUSIC:OFF increase speed? Because DF has to do all that as well. I don't mean to insult anyone who posted, but I feel this information is all useless.