AMD 7970m vs GTX 680m

They all believe 3DMark11 score reflect on gaming performance. Double 3DMark score means double FPS. Therefor the gaming benchmarks notebookcheck have is wrong...

Anyways, we should just wait and see for some review from notebookcheck or from any other site to see if these numbers are valid

 

No, I understand this very well. The problem is that way too many times these speculations are presented here as it's a done deal, which is far from the truth. I'm just balancing all the excitement about new 680m with a little bit of reality.

 

I think we're really, really starting to grasp at straws here. There's no new info and topics are being rehashed ad infinitum making it very tedious finding any useful info. What say we put a moratorium on this thread until new, substantiated (key word here) becomes available.

If you get the urge to bring up a question or argument in the meantime, look back through the thread first: I guarantee it's been discussed five times already. Now lets all go outside and get some fresh air - it's a beautiful day out there.

 

NO, lol. You must be new.

I have a theory that it's to do with the importance and addiction to being "right".

*the dopamine rush

 

delete....

 

Regarding benchmark comparissons...

Unless the benchmarks are conducted with prescripted sequences they are about as useful as a horny 7970M owner running in a circle in a small room in Metro and announcing "HOLY S**T 90 FPS IN METRO NOTEBOOKCHECK SAYS 675M ONLY GETS 17 BRB CHANGING DRAWERS"...

Don;t get too excited until someone can benchmark it in a highly controlled environment (test wise )

 

There are already a bunch of people getting their M17x R4 with the 7970M. I hope they can chime in and do some game benchmarks. I'll definitely be doing that once I get mine if this discussion is still going on.

 

BTW I would rather have a single card that performs 60% faster than 2 cards at 100% faster.

 

Is this some sad attempt at a personal attack? You can see exactly how long I've been an NBR member, and I've been working on computers since I was 15.

Try addressing my point and the logic behind it. I don't even have to be knowledgeable in computers to see the fallacies in your argument.



Sent from my Kindle Fire using Tapatalk 2

 

What we NEED is someone who is able to test a product out objectively, without any fanboyism that may affects the test results or the conclusion.

We need someone who have both GTX 675M and 7970M, who owns many games, who can swap out their GPU and test everything on the exact system with the exact same system settings. Usually its only the big reviewer sites that is able to do this, and which is why I trust them without listening to the usual "I`ve heard Homer did 50FPS in Metro on very high settings".

First of all, you heard about it. Second of all, high settings, what does that mean? 4xAA? 8xAA? AF? High on texture? Third, stock settings? Which CPU? Modified drivers? Drivers provided by OEM? Is it average FPS? Is it the highest FPS? What are you comparing it against? System with a slower CPU? etc etc etc.

Most people in this forum are WAY to much fan of either team, and is not able to discuss or comprehend results without pitching in some of their own biased opinions

 

I've posted on several threads several times, but unfortunately its hard to find with the hundreds of pages now

Skyrim...1080p ULTRA 4xMSAA 16xAF...killing dragons 80+ fps in the countryside, walking around White Run 70-80fps

Crysis 2 1080p DX11 ULTRA with high rez textures installed...2 test levels, never goes lower than 40fps...so i'm gonna estimate high 40's fps average

Metro 2033 1080p DX10 VERY HIGH AAA 45+fps, usually in the 50's...much higher in the cities

BF3...HUGE improvements over the last gen...looking at 50+ at 1080p Ultra.

Witcher 2 1080p Everything enabled except UBER Sampling...45+

These numbers are not unique to me. They are being reported by several other trusted members of this forum

My laptop feels like a gaming desktop now. if this isn't enough, Slight tweaks with in game settings (especially in Witcher 2) raise up the fps significantly, but i just haven't had the need to

 

Sheesh 580M vs 7970M is still being argued? Look at it with some sense. The 580M/675M as they stand are a little over 30% slower than the 560 Ti, so how on earth could the 7970M only be 30-something percent faster than those cards? You'll have to be saying that the 7970M is just barely faster than the 560 Ti. We know that's not true, so recheck your numbers.

 

And I've been using them since I was 11. That doesn't matter. I even have a forum account at EOCF registered 2003. That also doesn't matter.

No it's not a personal attack, you guys are making something so simple into the most complicated and futile argument.

The 7970M is a 7870 @ 800Mhz and a lower memory clock.

There's no need for any kind of unique methodology.

Look at the 7870 benchmarks and take off 15-20% to give a guesstimate of the performance.

The drivers matter as well, plain and simple. Compared to what there is now, the 3DM11 scores won't change that much but the games fps change drastically. No one uses 12.3 Beta 1 except Dell.

 

you have Geo, Widezu and me. Frankily I didnt know we were not allowed to get excited for upgrading to a better performance card (compared to 580/675M)

 

Look at the GTX560ti vs 7870 benchmarks.

Take 10-15% (same mem clocks but 15% lower core) off the 7870 and 25% off the GTX560ti (both mem and core are hit equally).

They are identical chips to these.

 

No one has suggested "unique methodology. Simple look at the benchmarks and gaming tests provided by notebookcheck and NBR members. That would seem the most logical.

BTW, you stating that I must be "new" implied that age/time registered played a part in who is right. It doesn't.

Sent from my Kindle Fire using Tapatalk 2

 

No no, I didn`t mean it like that. Of course you should be happy about your new GPU. It beats everything notebooks have so far I am not trusting Geo for a single bit. You and Widezu however, 100%.
Are you able to do a fully thoroughly and fair test of both? That would be so awesome if you did so we can lay this whole performance speculation dead.

Do a fully stock performance of both with genuine drivers, then do a overclock of both to see overclocking performance. And then compare everything after

 

Cloud, me/long/slick/widezu, we completely trust Geo, his numbers are right, trust me, also we are hardware enthusiasists, so it doesn't matter what company whatever it is, pretty much better performance is the important part, all of us except for Slick upgraded from 580m, so we were nvidia fans too, nothing wrong in it, and I think everybody is very very objective about the performance of the card.. anyway, forum will be swarming with 7970m soon, so we will see a ton of similar results

 

It implies what ever the interpretor assumes if it's obscure, in which it wasn't even that ambiguous. I wasn't taking about registration dates. That's arbitrary, I was talking about the subject matter which is drivers and game performance as well as the mobile card being based off desktop PitCairn XT.

 

When someone gets angry and result in personal attacks over some GPU discussions and says "you offended my baby" that is where I draw the line graz.

But I want exact specifications both with 580M and 7970M systems, what exactly the game settings they play are, exact FPS, atleast the average FPS from both GPUs. I won`t settle with 40+ FPS, and no mention of 580M with the same CPU. I would have easily provided with this if I owned any of them since its a fun comparison and should lay this discussion dead.

I know I am picky, but that is what reviewers do, and that is what a legit GPU review should contain.

 

Cloudfire. Can you please change your name to cloudfire.

Heaven was compared
Bf3 was
Skyrim
Crysis
3dmark
Many more

?its 2x?

What is the difference?

2x

What is the difference?

2x

What is the difference?

Nbr does not support brail!

 

Huh? Whats wrong with my nickname?

I haven`t seen a single review of 7970M yet x32993x. Nothing have been tested and compared with a valid test bench. No graphs or indepth comparison have been made. I am satisfied with the 3DMark11 and Vantage scores, I give you that. I made the Heaven benchmark comparison in this forum thank you, so I`ve seen it. But the game benchmarks with these +30FPS is not good enough for me. I want it tested in a controlled enviroment, and frankly I think its weird that no one have stepped up to the job since people own both 580s and 7970s now and not a single one of the reviewer sites have done any yet. But perhaps its just me who want to see one...

But Im not trying to force people. They can do whatever they want by me. Reviews is coming anyways soon from the big sites

 

Well there is one thing we do know.

It wont perform worse than 15% below the 7870, simple as.

 

I was going to try both (which I did for 3dmark11) but the darned card died the next day...

 

Are you getting a replacement btw? Any progress in your case?

 

I would be happy to supply some benchmarks of the 580M (675M) with my P170HM system (see sig) for the purpose of a stringent comparisson. Just send me a PM with the games oyu would like me to test and the specific levels you would like me to benchmark and i'll post them asap.

 

I just wanted to say Slick's benches are not 2X the performance of my 580m.

I get 40 average in BF3 ultra.

I get 35 average in Witcher 2 ultra.

Skyrim normally sits at 60 FPS locked vsync, never lower than 45 FPS and I play with forced ambient occlusion (which puts soft shadows on EVERYTHING and really kills performance).

I have a screenshot posted earlier in the thread, running Metro with ultra settings DX11 with tessellation and I was very close to what Slick said he got in DX10 without tess, I always was above 30 FPS and normally in the 45 range.

Bottom line, in real world games the 7970 is not 2X a 580m. There was a more professional comparison posted pages ago that showed that the improvement ranges from 10-80%, with the higher percent being on games that are already running well above 60 fps anyway. It is a great card for sure, I will be excited to see what Nvidia can do with the 680m and 780m using these new lower watt chips.

 

those numbers look like the 580 is overclocked?

Metro ultra? you mean Very high (which is the highest)...you're getting 45 fps VERY HIGH in DX11 1080p? Seems a tad high. Always over 30? don't belive you

Witcher 2 ultra? default ultra settings ? what have you enabled to get 35fps average?

My numbers are stock and i'm being very conservative with the estimates.

 

Thank you, Slickdude. Now since he proved that Crysis 2 maxed out can run 40 + fps on ultra settings with high texture pack on his 7970m, it seems that card can be 2x as powerful as 6990m cf.

 

Actually the highest gains were in those titles that were below 60fps.

 

speaking of which, would you mind sharing with us again your 'stock' benchmark results with 3DMark Vantage and 3dMark 11?

 

I just want to point out that saying things like "I get xx FPS in y game" is meaningless for comparing if there's no repeatable benchmark. Some of you guys are taking anecdotal fps estimates from people, comparing it to some other anecdotal fps, and trying to pass the difference off as fact. This is bad and people who do this should feel bad.

As many others have mentioned, just use the well documented 560ti and 7870 desktop numbers and subtract the clock differences... It's a much more accurate way to do a comparison.

The 7970m should come out to be like ~50% faster than the 580m/675m on average.

 

So here 7870 BF3, this is basically 7970m overclock.

HD 7870 Battlefield 3 Ultra settings - YouTube

 

wow wow wow bro.. a little slow there..

no 580m can achieve metro 40fps dx11 forget about ultra, cannot do high, at high I am dipping to 20 fps more than often.. you must have overvolted+overclocked, just overclocked you cannot go that far, and Slick's 7970m is stock..

also can you post a screenie of your witcher 2 ultra please? I am yet to see a none-sli/crossfire setup to achieve witcher 2 ultra (including DESKTOP, ehem..)...

true, 50% is a good estimate

 

Why are they not using max settings?

 

to make both cards look decent enough on paper?

 

But they should use max settings to show the power of the card.

 

Slickdude has the 7970m and he tested the games. Are you trying to say that his statements are not credible enough?

 

anyway that's the worst chart in the history of gpus I guess, which 580m at stock can score 15500 vantage to begin with? (I hope they are p gpu scores?)

 

He did benchmarks? I haven't seen any, but then again I haven't gone through the whole thread. Link?

 

http://forum.notebookreview.com/gaming-software-graphics-cards/658185-amd-7970m-vs-gtx-680m-102.html

It's on page 102. He played some levels with those settings and he got those FPS. The 7970m is really a strong card.

 

Those aren't benchmarks. That's what I mean. How are you supposed to compare "Witcher 2 1080p Everything enabled except UBER Sampling...45+"
with a 580m? Every playthrough will be different and will have different FPS.

Gameplay impressions are good for getting a general sense of performance, not for judging the relative performance of cards. To get a good comparison you need a repeatable benchmark so that the performance isn't affected by things other than the GPU.

 

^^ I am sorry, what??

The only thing missing from Slick's test is a chart showing average/min/max fps over time (and he is a busy man, doesn't have time for professional gaming bench), else it is perfect enough as he is not lying and giving the game settings..

also about repeatable bench, we provided a ton of 3dmark11's with P5600+, I think it is repeatable enough..

 

What really is the purpose of the 7970m card? Isn't not to increase performance on demanding games?

As long as the card in practice performs great on games that are demanding, that's fine with me. I am not into benchmarks on 3D mark or something. Make Crysis or BF 3 play with 40 to 50 fps on max settings with max res and I will be super happy

 

When released 680m?
News?

 

The only rumor we have is from Dell forums. September launch.

 

680m is suppose to be launched june 1st, then again when dell or any other company who would use the card probably half a month after its release.

tbh im just wondering if there will be 685m, and if there is going to be one, will 7990m beats it =)

 

If you want more tests you can read this topic. P4G / [REVIEW] ATI HD7970M + Medion X6813

Also on benchmarktest you cand find test in game http://www.notebookcheck.net/AMD-Radeon-HD-7970M.72675.0.html

but for the moment, the drivers for this card aren't optimised.

 

Always link your source please. Dell source is linked some pages ago.

 

I still think its coming out in June at Computex