What do you think is the best driver to go with the amd? Should I stick to the original clevo or try the amd beta drivers?

I belive they can improve the driver so much so that it will come even closer to the GTX 780M performance!!

In light of the FCAT aka frame rating issue, what is the real FPS of the 8970m when runt frames are deducted? When it comes to any gpu from AMD, FPS is highly suspect.

Check out this video of Crysis 3 running at 60FPS on a 7970M! http://www.youtube.com/watch?v=kI3mzeM9rr4

Well said ivuzmek, I agree with all you've said.

I found the spec sheets for the 7970M & the 8970M on the AMD website, here are the links:

http://www.amd.com/us/products/notebook/graphics/7000m/7900m/Pages/radeon-7900m-series.aspx#2

http://www.amd.com/us/products/notebook/graphics/oem-solutions/Pages/8950-8970.aspx#2

The only difference is the 6% higher clock and and 'up to 4GB' for the Video RAM.  This is reflected in the increase in 'GFLOPS Single Precision compute power' of 6% seen on the spec sheets.

If 6% higher clock is the only improvement AMD managed in the past year, than it sounds like they've hit a brick wall. NVIDIA clearly showed that they can go further from the 680M, while AMD is still struggling to create something better than 7970M. Could this be a beggining of the end for the AMD (as far as matching NVIDIA for performance)? Their CPUs already suck. Thank God for the price tag on the 8970M, this will hurt NVIDIA more than tech specs.

If you knew, how the test went, you probably wouldn´t write some of the things  :)

1) It is very time-killing to build a big comparison-table for our website. So in case of doubt, i choose the notebook, which has more results. That has nothing to do with making a certain gpu look bad or good.

2) The Alienware M17x R4 is one of the best HD 7970M machines so far, because it has a non-enduro-mode (brings better results in some cases).

3) I had to retest ALL the benchmarks (including the 30 games), because the official Catalyst 13.4 / 13.6 Beta(2) sucked so much on our system. That took about eight to ten hours. In other words: a whole working day.

4) The GTX 780M and the GTX 770M dindn´t work at first. I had to search several hours for the problem (reinstalling the Intel-driver did the trick). In the end, there was so much lost time, i couldn´t find the nerve to create new CAD- oder gaming-benchmarks with the HD 7970M, the GTX 680M, the GTX 770M and the GTX 780M. I hope, you can understand that a liitle...

Quote from: 8970M is nothing new! on July 27, 2013, 18:53:05
Any performance difference you see between the 7970M and the 8970M on this website is due to the fact that the 7970M was tested with older drivers, so I guess the benchmark figures for the 7970M are a little dated on this site, that's why it looks like the 8970M performs significantly better than the 7970M.

Right why did Mr. Glaser include the Alienware M17xR4 with the old 8.934, ^{http://www.notebookcheck.com/Test-Alienware-M17x-R4-Notebook.74805.0.html}
and not the much newer results from this article? ^{http://www.notebookcheck.net/Review-Update-Radeon-HD-7970M-vs-GeForce-GTX-680M.87744.0.html}

The answer seems obvious: There are no later CAD benchmark data available than those from the Alienware tests.

Later on at the games benchmarks he uses newer results, but - unfortunately - not each and every time. To my untrained eye though it seems a little too far fetched to discredit his article and the whole site therefore. At least he's craftsman enough to cite the driver and models used for reference. He even points out that "newer" doesn't necessarily mean better - particularly when drivers are to be concerned.

I for my part find those kind of articles very readable and informative. The layout is better than that of most competitors and data is well presented and comprehensive, links are at the spots where you need want them.
The easy to grasp comprehensiveness is what I like most about notebookcheck articles. It's almost unparalleled to my knowledge, though I would like even more "scientific reference" sometimes - like a link to a raw database of all tests and results of the devices tested.

You're aware of a better site out there you think I should check out?

Back to the article. I agree with your anger about AMD basicaly just rebranding the 7970m to 8790m. Sometimes results and facts are better than thorough guessing though. I was waiting for such an article - 4GB GDDR5 not 2GB like with the 7970m - I was about to fall for that marketing trick. And who knows what else AMD changed, things you cannot tell from looking at shaders and frequencies - I'm no engineer ... .

If I had to chose between 7970m and 8970m for roughly the same price I would obviously go with the 8970m (no information on any overclocking attempts of 8970m yet). In reality I don't have to chose since AMD already took the 7970m from the market.

Nvidia's got more justification for their 780m price tag than they had for their 680m price tag. 

I can't believe AMD released the 8970M, it's exactly the same as the 7970M except that the core clock is 50Mhz higher or in other words a 6% higher core clock than the 7970M.  So, at best the 8970M will perform 6% faster than the 7970M it replaces.  That's almost not even worth talking about when it come to an increase in performance!

Any performance difference you see between the 7970M and the 8970M on this website is due to the fact that the 7970M was tested with older drivers, so I guess the benchmark figures for the 7970M are a little dated on this site, that's why it looks like the 8970M performs significantly better than the 7970M.  If you took a 7970M with the latest drivers & compared it to the 8970M with the latest drivers then I can't imagine the difference being any greater than 6%, which is due to the 6% increase in core clock that I mentioned in the first paragraph.  Quite some audacity to call this the 8970M, why not just get a 7970M and overclock it 6%!