Sorry your browser is not supported!

You are using an outdated browser that does not support modern web technologies, in order to use this site please update to a new browser.

Browsers supported include Chrome, FireFox, Safari, Opera, Internet Explorer 10+ or Microsoft Edge.

DarkBASIC Professional Discussion / FPS V's Machine Specs

Author
Message
JSN
22
Years of Service
User Offline
Joined: 8th Nov 2003
Location: 127.0.0.1
Posted: 3rd Dec 2003 12:47
Link
I was wondering if anyone had some kind of graph that represented the amount of cpu-load against a particular machine spec.

For example...

Having 200 sprites moving from left to right, single pixel increments, would take 1.8 seconds on a machine that has a 1Ghz CPU

The same results on a 2Ghz (for simple maths!), would take .9 seconds (again, for sake of example).

It would be handy to know...

-----

~ J ~
Back to top
Profile PM Website
OSX Using Happy Dude
22
Years of Service
User Offline
Joined: 21st Aug 2003
Location: At home
Posted: 3rd Dec 2003 12:57
Link
Why ?


Mirrors are more fun than television. Well, that was fun, in a not-so-fun sort of way...
Back to top
Profile PM Email Website
JSN
22
Years of Service
User Offline
Joined: 8th Nov 2003
Location: 127.0.0.1
Posted: 3rd Dec 2003 13:03
Link
Because the last thing I want is to create a project and say "Hey guys, (and gals - bit PC there...!), look what I've done, it's got this, that, and t'other" only to have some small print saying it requires a 3.2Ghz machine to run.

I did read somewhere that if I set my sync rate to, say, 40FPS, then whatever I see on my screen (with a CPU running at 3.8Ghz), will be the same as someone who has a 400Mhz, but I've got my doubts...!

-----

~ J ~
Back to top
Profile PM Website
OSX Using Happy Dude
22
Years of Service
User Offline
Joined: 21st Aug 2003
Location: At home
Posted: 3rd Dec 2003 13:19
Link
That's where using a routine to calculate delta movement (aka time based movement) comes in. That way you dont need to know what (possibly) each machine is capable off.


Mirrors are more fun than television. Well, that was fun, in a not-so-fun sort of way...
Back to top
Profile PM Email Website
JSN
22
Years of Service
User Offline
Joined: 8th Nov 2003
Location: 127.0.0.1
Posted: 3rd Dec 2003 13:28
Link
THANKS TCA, I honestly didn't know that's what Delta Time is. I thought it was some 3d thingie-ma-jig.

Right, er, off to learn that...

Thanks again, learnt something new...

-----

~ J ~
Back to top
Profile PM Website
waffle
23
Years of Service
User Offline
Joined: 9th Sep 2002
Location: Western USA
Posted: 3rd Dec 2003 13:46
Link
I think they missed the point... If FPS drops below 10, your game becomes almost unplayable. And yes, it would be nice to get some kind of idea under what conditions that may happen.

so here are some clues to try..
1 - during devolement, set syncrate to 0 or run in windowed mode
and measure FPS or time to update display... anything will work

2 - determine a way to measure how many textures are in use at any one time... Maybe count the number of individual models types in use and so forth.

3 - count polygons actually rendered. I think DB has a method for that.

------------
math stuff
A - Compare texture usage to 3D card rating. If using 50% capacity,
than a card of 1/2 size should be OK depending on polygons

B - If target computer has no card, DBPro might not work or may run at 1/2 to 1/3 speed.... This should still be OK, as I'm doing fine
on a PC running at 0.9Ghz without a 3Dcard. I do have slowdowns from time to time.

C - Compare CPU Speed
If you have a 3GH with 3D and run at about 200fps (windowed)
2GH with 3D expect about 100, without 50
1GH with 3D 50, without 25

assuming your 3D card is maxed out. If your card is at 1/2 capacity
then there would be a much less slowdown from one PC to another

Making a comparison fullscreen exclusive will be unreliable as the max refresh rate is locked to the monitor.

So, set your movement stuff based on the MAX speed and then test with FPS locked at "worst case PC" that you wish to support and see how it plays.

internet gaming group
Back to top
Profile PM Email Website
OSX Using Happy Dude
22
Years of Service
User Offline
Joined: 21st Aug 2003
Location: At home
Posted: 3rd Dec 2003 13:50
Link
You would still need time-based movements as Windows is (loosely anyway) multi-tasking system - it could slow down slightly for any reason at any time.


Mirrors are more fun than television. Well, that was fun, in a not-so-fun sort of way...
Back to top
Profile PM Email Website
JSN
22
Years of Service
User Offline
Joined: 8th Nov 2003
Location: 127.0.0.1
Posted: 3rd Dec 2003 14:17
Link
Quote: "1 - during devolement, set syncrate to 0 or run in windowed mode
and measure FPS or time to update display... anything will work"


Running in windowed mode, have the FPS set to 40.

Quote: "2 - determine a way to measure how many textures are in use at any one time... Maybe count the number of individual models types in use and so forth."


There are no textures, there are however 40 images, each one 40x40 pixels

Quote: "3 - count polygons actually rendered. I think DB has a method for that."


If you're talking 3d polygons here, then there are none.

-----

~ J ~
Back to top
Profile PM Website
Jaze
22
Years of Service
User Offline
Joined: 20th Nov 2003
Location: Connecticut, USA
Posted: 5th Dec 2003 09:59
Link
Um...Guys - You all sound on the mark to me. Window OS not seamless multitask - its not a realtime system by default - its a priority based preemptive scheduling system..if I'm not mistaken.

I just wanted to share something: I had a DECENT (so I thought) 3d card - 32megs of ram - I had to put detail low or medium on many games.

I got a new NVidia Card, 64megs of ram - and I can have all the detail on and there seems to be speed to spare. The CPU has remained the same. Night and Day - Both "quality cards" but I'm shocked at just how much load these new fangled vid cards with their own processors can do to lighten the CPU load! In otherwords - it wouldn't surprise me if flight sim worked on a 486 (if it had the AGP port for the card) with this card I got.

The best thing I can think of to add to your suggestions is to make a benchmark program (or perhaps DarkBasic could release a generic one) that runs a series of tests to benchmark various poly counts, res, depth etc...and maybe dump it to a text file so we can use that to configure optimal performance settings for PC in use

-=/Jaze/=-
Back to top
Profile PM Email Website
Back to top

Login to post a reply

Server time is: 2026-07-26 20:16:46
Your offset time is: 2026-07-26 20:16:46