Resolution Change for Benchmarking Thread

What should the new benchmarking resolution be?


  • Total voters
    15
I've been in almost every bench since the start and SF4 has really been the first bench we've done where I'm not seeing full load or close to it on gpu. The argument for increasing size due to cpu bottleneck based on one bench just doesn't hold water IMO.
 
No. SuperPi is boring to everyone but you.
We get it. You have the fastest processor on this website (excluding a 980x, of course).
 
I guess my point is should we be benchmarking games at a resolution that is lower than half of the install base of steam itself?
 
I've been in almost every bench since the start and SF4 has really been the first bench we've done where I'm not seeing full load or close to it on gpu. The argument for increasing size due to cpu bottleneck based on one bench just doesn't hold water IMO.

That's true.
But I would like to increase the resolution to push my card a bit. Running it at a measly 720p isn't much to brag about (unless everyone else runs at 720p).

I eat 980x's for breakfast in superpi

You mean that thing that only you care about? :biggrin:
 
I'm trying to get bass to let us do SuperPi, so I can flex my 2600K.
poke.gif

Yea. SuperPi doesn't do well on AMD CPU's. Talk about a one sided benchmark.
 
Yea. SuperPi doesn't do well on AMD CPU's. Talk about a one sided benchmark.

....talk about crappy CPUs.

It isn't the bench's fault, it is amd and their crappy single core performance.

all hail Intel master race.
 
Check this out:

SF4.jpg


q6600 at 3Ghz and stock 5870.

And to further my point about the resolution:
sf4-1080.jpg



That's run at 1920x1080, so more than double the pixels and a drop of 10fps?
 
Maybe this benchmark just isn't very stressful on the GPU at all -- here is 6048x1080

SF6048x1080.jpg
 
Last edited:
It isn't very stressful, but it does scale almost perfectly.

This is why i wanted to base it off the score (notice it drops dramatically as your res goes up), but Bass is being a noob, and insists on FPS
 
It isn't very stressful, but it does scale almost perfectly.

This is why i wanted to base it off the score (notice it drops dramatically as your res goes up), but Bass is being a noob, and insists on FPS

What is the score measuring anyway? I noticed it dropped off, but the fps average seemed good and I didn't notice any wild fluctuations. They visually appeared identical.

I've also noticed people with slower CPU's and GPU's getting higher scores than me -- I'm wondering if it's memory bandwidth related?
 
Last edited:
you have the slowest CPU on the list, and a middle of the road GPU, and a slower chipset interface than most of us
 
you have the slowest CPU on the list, and a middle of the road GPU, and a slower chipset interface than most of us

I'm not talking about on the ego forums -- I did a bunch of googling. Also, a 5870 is at least an equal to a 6870, so going by the current leaderboard it's hardly middle of the road. I won't dispute the CPU being slow -- I'm still on socket 775 and and obviously not getting a sandy bridge any time soon.

see here: http://www.hardocp.com/article/2010/10/21/amd_radeon_hd_6870_6850_video_card_review/9

"The AMD Radeon HD 6870 is not as fast as the Radeon HD 5870, and it was never meant to be. The Radeon HD 6870 comes in-between the 5850 and 5870 in performance, but the kicker is that it is much cheaper, runs cooler, and uses less power doing it.
...

In regards to the 6870, if you currently own a 5870 then the 6870 is not an upgrade, but if you own a 5850 or below, then the 6870 is an upgrade."
 
Last edited:
Maybe this benchmark just isn't very stressful on the GPU at all -- here is 6048x1080

SF6048x1080.jpg

You have eyefinity? :D
How is it? reflex doesn't really talk about it or use it (since his GPU solution is broken half the time :lol: )
 
Back
Top