Most important number in benchmarking a video card?

B

BradF1979

Limp Gawd
Joined
Nov 27, 2009
Messages
181
Shouldn't it be a minimum FPS? We are getting into a time when there are multiple cards performing well above where they need to be for adequate gameplay. That said, to me, the most important thing is the minimum frames per second (assuming that the average is close to 60 FPS).

What does everyone else think?
 
Well minimum numbers also tend to be the most fishy. There are too many cases where a videogame will briefly hiccup or pause that have nothing to do with the videocard (usually the hard drive) but that will show up as your minimum FPS even though it may have only happened one time during the whole run. If you have the chance to look at an actual FPS graph then looking at the minimums can be useful.

I agree that low minimum FPS is bad but you have to be able to look at that figure with enough context to determine if it is actually going to be an issue or not, and in the absence of that context, I find the average FPS to be a better indicator.
 
I agree completely, you definitely need a graph over time of the minimum FPS values. Unfortunately, I've seen few reviews that offer this view.
 
I've seen sites start doing gameplay evaluations, and frameate over time, only to back off and stop doing it because it takes so much work.

Luckily for you guys, we aren't lazy with our evaluations.
 
Brent_Justice said:
I've seen sites start doing gameplay evaluations, and frameate over time, only to back off and stop doing it because it takes so much work.

Luckily for you guys, we aren't lazy with our evaluations.
Click to expand...
^ :)
Plus you guys state if it's smooth gameplay vs. numbers claiming otherwise.
 
When I had 3gb of memory my team fortress 2 would hit a minimum of 2fps (never saw it, but it showed it in the cl_showfps 2 log). But my SLI gtx 280's were still averaging roughly 120fps. Then I moved to 9gb of memory and the same log shows about 24 as my low. Again, I never actually notice any slow down, I figure its just a small hickup.

So like others say, a graph would be ideal. Or at least something where you can throw out the outliers.
 
You must log in or register to reply here.
Back
Top