Is the new 3DMark useful for your buying decisions?

[W1zzard]

Futuremark just released the latest 2013 version of their famous benchmark 3DMark.

What do you think about it? Is it useful for you?

 

[Kaynar]

Cool looking and good way to APPROXIMATELY compare hardware, but every game and every application is different so just one benchmarking suite doesn't tell you everything you need to know.

A 3dMark11 number doesn't tell you about the microstutter nor fratetime (hot topic) issues, for example. Not to mention that I am pretty sure both AMD and nVidia do extra tweaking on their drivers dedicated to achieve better benchmark suit results (which does not benefit other apps)...

 

[Frick]

The last 3dmark I kinda cared about was 05.

Mostly because I have never gotten good numbers from it.

 

[Initialised]

Will be putting it through it's paces, if it does what I want I'll be using it.

 

[Phusius]

well my 7950 at 1200 core is beating out gtx 680's in fire mark score.

$270 free ship no tax. ./shrug

 

[BigMack70]

Loving the new version of 3dmark compared to 3dmark11. Much nicer feature set, more interesting tests, seems more optimized (Fire Strike both looks and performs better than 3dmark11 Extreme preset). The Fire Strike Extreme combined test is absolutely BRUTAL, though... it makes my computer cry.

It is kind of interesting that it favors AMD's hardware right now, which is unusual for 3dmark. I wonder how things will look after a driver update from both sides...

 

[brandonwh64]

Still have no base to go off of or a standard but once that is established then yes it would base our decisions on future GPU purchases.

 

[WhiteLotus]

Futuremark just released the latest 2013 version of their famous benchmark 3DMark.

What do you think about it? Is it useful for you?

If you are asking this because you are interested in including it in your benchmarks, I say get rid of all "non-game" benchmarks. Work more on different AA settings and such in different games at resolutions.

 

[brandonwh64]

There should be a maybe added to this thread.

 

[Steevo]

Im not sure, after running it through the paces on my systems and comparing others scores it seems to be biased towards more vmem and system RAM than what games actually require. Also changing the CCC AA and AF from forced on to application control seemed to have very little effect on the score, I noticed that framerates were 10% lower with 4X AA Box, 8X AF, and mipmap set to quality and forced. That is how I play most every game, and the performance drop in most games is much more substantial.

 

[Fluffmeister]

I prefer Catzilla.

 

[Disparia]

No, I won't be using it more than once.

Don't find it as entertaining as the earlier versions and haven't really cared about marks since I stopped building computers commercially many years ago. For me, the drive space will be better used by a game and/or some small and simple benchmarks.

 

[TRWOV]

well my 7950 at 1200 core is beating out gtx 680's in fire mark score.

$270 free ship no tax. ./shrug

Probably due to the VRAM. Fire is rendered at 1440p AFAIK.

 

[Batou1986]

Honestly i don't give much thought to 3dmark, all game engines are so very different now its not a good indication of a systems systems performance.
Two systems may be about the same in 3dmark but vastly different in specific games.

 

[MxPhenom 216]

I like benchmarks like 3D just to see where your system lines up compared to others. I won't use synthetic benchmarks like these to determine buying decision. For that ill look at real FPS numbers in the games I play.

 

[Deleted member 74752]

Appears that a good many are running it with unapproved drivers still.

 

[erocker]

Appears that a good many are running it with unapproved drivers still.

I don't know why any 7 series AMD user would want to use anything other than the 13.2 beta's.

 

[InnocentCriminal]

I used to really like benchmarking and it can be a good tool when trying to gage increases from overclocking, but it definitely doesn't affect why I would purchase a GFX card or a component. Games that I play, at the resolutions and settings that I would play them are far more appealing and beneficial to my decision.

 

[MxPhenom 216]

Appears that a good many are running it with unapproved drivers still.

EVERYONE is using "unapproved" drivers! Nvidia and AMD. Its something to do with 3Dmark, not the driver everyone is using.

 

[Deleted member 74752]

I don't know why any 7 series AMD user would want to use anything other than the 13.2 beta's.

Perhaps to make a valid HWBot submission? I see that the AMD 13.2 is a approved driver now.

EVERYONE is using "unapproved" drivers! Nvidia and AMD. Its something to do with 3Dmark, not the driver everyone is using.

Then why are there "approved" submissions out there? http://www.3dmark.com/fs/23169

 

[Frick]

Perhaps to make a valid HWBot submission? I see that the AMD 13.2 is a approved driver now.



Then why are there "approved" submissions out there? http://www.3dmark.com/fs/23169

There is something weird there. I dowloaded the layest drivers from nvidia and they are unapproved.

 

[Deleted member 74752]

There is something weird there. I dowloaded the layest drivers from nvidia and they are unapproved.

Scroll down on the Nvidia site and select beta/archived drivers...313.96 is the one you want. AMD's 13.2 seem to be working fine as well.

 

[jihadjoe]

310.90 WHQL is unapproved lol.

 

[Black Panther]

I couldn't vote because my answer is yes and no.

I read reviews, and take into consideration a multitude of benches (both dedicated benchmarks and in-game benchmarks) before I decide to buy hardware.

So yes, the new 3DMark is useful. But at the same time, no, because it's not alone in determining my choice.

 

[Delta6326]

I always skip over any non real world test. I see no point in a number that means nothing about how my game will play and how fast apps will load.

I only see a use for these benchmarks for OC'ers and nothing more.

(non of these # are real just a example)

This is how i see these benchmarks.
Say a 680get's 10,000 and a 7970 get's 9,000 I would go ahead and say the 7970 is 10% slower.
So by basing on this you then go out and buy a 680 you go to play a game you run that games benchmark compare to a 7970 and the numbers have changed to 680 90 fps and 7970 100 I would than be angry and be confused as it was 10% faster in 3Dmark, but in this game it was 10% slower!

Now this example probably was wrong, but I'm trying to say is that you can't base your purchase off of 1 benchmark.

I also only pay attention in reviews to the games that " I " play on, because I don't care if a 7970 was to get 120fps and a 680 10000 fps in a game that I don't play. That's why I like W1zz's reviews because they do have lots of Game's and many of the ones I do play.