• Competitor rules

    Please remember that any mention of competitors, hinting at competitors or offering to provide details of competitors will result in an account suspension. The full rules can be found under the 'Terms and Rules' link in the bottom right corner of your screen. Just don't mention competitors in any way, shape or form and you'll be OK.

ATI cuts 6950 allocation

The improvement in Metro 2033 framerates is huge.

benchq.jpg
 
Maybe the charts are right and they have concentrated in improving dx11 games? I mean look at the metro score!?! It destroys anything, probably other cards on crossfire/SLI
 
buttkinz said:
Maybe the charts are right and they have concentrated in improving dx11 games? I mean look at the metro score!?! It destroys anything, probably other cards on crossfire/SLI
Click to expand...

Well tessellation performance has definitely improved massively, and that is a big part of the DX11 spec (not that any games really put it to usde properly yet, or will for a couple of years, but anyhow...)

The 4xAA / tessellation / DOF in Metro setting has always caused problems for cypress, so it's not quite so surprising that it's seeing massive improvements here. As DM pointed outl, it's still not playable.

For reference, I found a user benchmark with a GTX480 on these settings. It scores 24.64fps average, which is slightly above the score of the 6970. (see here) The next page has another user showing 43.34fps for GTX480 SLI, so the first result seems realistic. The result for Cayman will certainly "destroy" dual-card cypress setups, but it's still lagging behind the GTX480 in this particular configuration. I'll have a search round for a better source though - I know they're out there somewhere as I've come across them before.
 
Last edited:
CAT-THE-FIFTH said:
Metro2033 is a Nvidia sponsored game IIRC. I suspect with the 10.12 drivers performance should be more consistent IMHO.
Click to expand...

The AMD cards do a hell of a lot better at more realistic and playable settings (no tessellation, no DOF, AAA instead of 4xAA), and get a lot closer to the Fermi cards. Those are the settings that should show the real-world performance improvements. The improvements in the table simply show that a severe bottleneck in Cypress at these settings has been removed.
 
seena said:
AMD_HD6970_Slide_02.jpg

Two Graphics Engines with fifteen 64-core Clusters each for a grand total of 1920 cores
is it??
Click to expand...

Does it show 2x 12 SIMD engines for 24x64=1536 shaders? No, its a NDA covered secret, the structure would be the same 6 simds per engine or 66, gives the idea, not the specifics. Think after ALL this secretive stuff a slide leaked a month ago would have confirmed the shader count?

Also the SIMD collections aren't really "engines", last gen had the SIMD's split in two groups but weren't called engines, AFAIK the engines are the ROP's and, front end, it used to be one part, now its split into two engines.
 
seena said:
AMD_HD6970_Slide_02.jpg

Two Graphics Engines with fifteen 64-core Clusters each for a grand total of 1920 cores
is it??
Click to expand...

There is no indication of the number of SIMD cores on that slide...

Also, the "dual graphics engine" is something that has been in place since r800 (two parallel banks of stream processors). The change to Cayman above seems relatively minor; the dispatch processor has been split to communicate with each separate bank of shaders.
 
Skyrocket said:
Looking at those chart it does really appear that 6970 is nothing but a overclocked 5870 with a bolted on tesselators.
Click to expand...
I tend to fully agree. The only way I can justify this for AMD is that future games will use higher and higher levels of tesellation. That being the case, we may eventually see 6900 pipping Fermi. Don't hold your breath.
 
555BUK said:
I tend to fully agree. The only way I can justify this for AMD is that future games will use higher and higher levels of tesellation. That being the case, we may eventually see 6900 pipping Fermi. Don't hold your breath.
Click to expand...

Just seem's real bizarre that they would just make the same improvements they made with 68xx (efficiency changes for shaders), improve the tessellator (the buffer increased?) and not actually actively try to improve performance levels above the 5870 significantly.

I don't think AMD relying on a dual card solution to solve this issue is a smart move, in fact I think it's bloody foolish IF they have gone down this path.

I'm not 'pro' AMD, but it is disappointing, I really hope for them that in two days time they actually have a decent single card on the table. I'm struggling to believe they haven't frankly.

One last thought. Are those claiming to have the 69xx's cards not actually using the driver that came on the CD in the packaging? It would make more sense to use that then the latest beta if AMD are keeping things close to their chest?
 
You must log in or register to reply here.
Back
Top Bottom