Tuesday, February 14th 2012
Radeon HD 7770 Put Through 3DMark 11
AMD's Radeon HD 7770 mid-range graphics card, which is slated for a little later this month, got its first public shot at 3DMark 11. The card was put through the Performance preset of the benchmark, where it scored P3535 points. The bench was driven by an Intel Core i7-3960X processor. The reviewer also took GPU-Z screenshots of this card, revealing low core temperature. Based on the 28 nm Cape Verde GPU, the HD 7700 is said to have 640 Graphics CoreNext stream processors, and 1 GB of memory over a 128-bit wide GDDR5 memory interface. It is designed for sub-$200 price points.
Source:
VR-Zone
59 Comments on Radeon HD 7770 Put Through 3DMark 11
Thing is it might be on par with a 460, it's newer tech and as such is lower powered will probably demolish a 460 when overclocked if it's bigger brothers are anything to go by and has DX 11.1 as long as it's priced right it then it should sell well.
Essentially you can look at it as either improving performance in a given power envelope (TDP envelope) or providing similar performance at a much improved power envelope*.
I would have preffered to see something a little more potent though for a x7xx model. Considering the HD78xx are predicted to have 1280 GCN units I would have expected along the lines of 960 or so for the HD7770.
For mid range gaming that would be perfect.
*Considering the HD6850 hardly sucks down power how much of an improvement will we see in the real world?
Furthermore under 180$ you can buy GTX560, 460 even a similar performing card (6850) as low as 130$ and a bunch of second hand better performing cards.
I really don't see anything to be happy about. There is absolutely no reason whatsoever to pay the same price for the same performance for a new generation card (although the naming scheme is different)
on an i2500k at stock look for the overall score to be 200-400pts lower.
www.legitreviews.com/article/1558/11/
The graphics test still has to use the cpu.
So yeah this is likely slightly worse than a 6850.
The combined test is meant to tax both the CPU and GPU and the Physics test is meant to tax the CPU.
However,we will soon find out how good or how bad the card performs tomorrow.
TBH,it would be a shame if the HD7770 cannot at least match an HD6850 1GB.
www.techpowerup.com/157715/More-HD-7770-Leaks-Pictures-Plus-3DMark-Benchmarks.html
It used a Core i5-3550 supposedly.
The graphics score looks very similar to the newest leak. The physics score only seems slightly higher than a Core i5 2500K.
It seems the combined scores are similar and the latest leaked is only skewed by the greatly increased physics score.
However,it would be nice to see the graphics score for an HD6850 1GB with a Core i5 2500 running at stock clockspeeds.
7970 - new level
7950 - 580
7870 - 570
7850 - 560ti
7770 - 550
Everything we've seen coincides with this thinking. You can see why Pitcairn is important.
THE 7000 SERIES PURPOSE IS TO COMPETE WITH THE 500 SERIES AT A SIMILAR PRICE AND PERFORMANCE BECAUSE THEY KNEW THEY WOULD BE FIRST.
They will adjust prices as necessary compared to absolute performance of kepler designs as they arrive. I imagine Sea Islands will have some-what similar designs to Kepler as 28nm will be mature and in super mass (fab 15) production by then. I digress.
I would feel a very large sad if 7870 is 20 @ 1ghz instead of 22 @ 950.
Not only would they continue the trend of way too much bandwidth for a design, it would seem to me 22 CU is a down-right perfect match for 24 ROPs...more perfect than any other possible design from either camp's arch...why blow that opportunity with anything less in this highly-contentious space?
On top of it all, one can assume 660 (salvage gk104) may be 14 SMs. If that were the case it would be VERY close to the same PPC as 22 CUs (a small amount faster...think 7950 to 7970; a few percent). I would hope they don't balk that market to nVIDIA when the full gk104 is already probably going to push prices down in that segment.
I really want to see 7870 make a mockery of 660's die size and the inefficiency of fermi/kepler because of the simple reality that it is not a very precise modular design and 660 will show why that sucks, especially since it would be running slower clockspeeds as a salvage part; the same medicine 660ti looks to use on 7950 (the later being a 90% salvage part on a similar size die with still greater ppc but the former running 1/6th greater clock) just to break even.
If AMD doesn't give me that I will have a disappoint.
That's right bro............long gone are those generations where you could see a 50 - 80% increase.
HD7770, sorry, you're not the one.
I think MSRP for reference 7770 cards will be $180 and those will perform right above a 6850, while OC's will approach 6870’s and there's nothing wrong with that. I've been saying based on the TSMC 28Nm price bump mostly, and against past MSRP of 5770 and 6850/6870 it will all be what it is. There's no use crying especially given Nvidia will also be passing along the TSMC pricing when they show.
While we have no idea what GK106 parts have in store; figure by end of summer nicely appointed OC'd 7770 could be $140 –AR, and that’s the way it goes...
had the 6850 been benched with the SB-E, the score would be higher. No matter how much you try to focus a benchmark for a particular component, the others will affect the result.
you don't have to take my word for it. Run 3dmark 11 yourself, then underclock your cpu by 500MHZ and run it again. The graphics scores will drop.
Solid card at the sub $200 price point.