View Full Version : The problem with benchmarks
omiknight52
Oct 2, 2008, 02:52 PM
Why is it that testers always use high end components to get results from benchmarks.
Why not use or compare all three; low-end, high-end and mid-range computers so people don't get discourage or put off because they can't afford or run a game without spending thousands of dollars. I mean seriously u can run crysis with a mid-range gpu on high and get decent fps without dropping $1000 on a pc. It would be nice to not see all test use core 2 extreme or quad, maybe a amd athlon 4400 or even an intel core 2 duo e4600.
Edito
Oct 2, 2008, 03:19 PM
They use high end components to do a benchs of high end components you want to see a benchmark of a 280GTX with a E2160 512MB of ram DDR2 533 or something like this??? Don't get me wrong but i think the benchmarks are just fine they it is...
Brightmorn
Oct 3, 2008, 12:25 AM
The reason I believe that they use high end components is to make the limiter of the benchmark is the card and not the CPU or whatever other part of the system. If you used a lower end CPU to compare the cards directly you may or may not see a difference due to the CPU not being able to keep up.
omiknight52
Oct 3, 2008, 01:57 AM
My point is to test so ppl can get an idea of what they need to run new games for budget gaming with mid-range components. You may be right about the cpu not playing a vital part since games use gpu heavily, but still some get the picture that u need expensive cpu or components.
ghost101
Oct 3, 2008, 02:13 AM
My point is to test so ppl can get an idea of what they need to run new games for budget gaming with mid-range components. You may be right about the cpu not playing a vital part since games use gpu heavily, but still some get the picture that u need expensive cpu or components.
The thing is, these are graphics card review. Not system reviews. Therefore, all that matters is the relative positions when determinging which graphics cards are better.
KainXS
Oct 3, 2008, 02:14 AM
thats why you heavily research everything you buy beforehand,
PennySavre
Oct 3, 2008, 02:17 AM
I think you can find pretty much any review on something that has something you are looking for when a video card or whatever has been out for awhile... the real task is searching for such reviews.
KainXS
Oct 3, 2008, 02:30 AM
I think you can find pretty much any review on something that has something you are looking for when a video card or whatever has been out for awhile... the real task is searching for such reviews.
the problem is that there are many false, and biased reviews on the web from respectable sites.
Kwod
Oct 3, 2008, 03:12 AM
One thing to remember is that most benchmarks attempt to max the game, ie, crank up the aa and af, so the framerates between COD4 at 1920x1200+4xaa+16af and 1280x800 without aa and af maxed will be significant.
I now have a large LCD, so things are a bit different, but when I had my old 19in CRT, even 1024x768 with zero aa and af+ medium to high in game settings looked good to very good.
Moral of the story is that a 9600GT will play all games well at modest settings, and because the latest games are hi-detail, they look a lot better than older games do at lower res.
TonyStark
Oct 3, 2008, 03:33 AM
There are plenty of benchmarks and reviews of low/midrange hardware. Its not just the enthusiast stuff that gets reviewed. :confused:
omiknight52
Oct 3, 2008, 03:35 AM
not lately and show some links :)
TonyStark
Oct 3, 2008, 03:41 AM
not lately and show some links :)
I just went to firingsquad.com and on the first page there is a review of the HD4550.
http://www.firingsquad.com/hardware/ati_radeon_4550_performance/
:confused:
omiknight52
Oct 4, 2008, 10:23 PM
I guess I assumed wrong about other components playing a huge part in gpu performance. It would be good if benchmarks reflect a certain computer build for example an intel c2d 4400 pared with a any 3800's or 4670. I think its called bottlenecking or something like that were you don't get best performance or so because one component is to much for the other; correct me on this one not sure
Kwod
Oct 8, 2008, 03:20 AM
I guess I assumed wrong about other components playing a huge part in gpu performance. It would be good if benchmarks reflect a certain computer build for example an intel c2d 4400 pared with a any 3800's or 4670. I think its called bottlenecking or something like that were you don't get best performance or so because one component is to much for the other; correct me on this one not sure
Yep....you've got to match your components, for eg, a good gaming rig these days would be e7200+4850+2gig ram, but a much better one would be a e8400+4870 1 gig+ 4 gig ram.
vBulletin® v3.8.6, Copyright ©2000-2013, Jelsoft Enterprises Ltd.