Mobility HD5870 review is out!

Hello There,

I found test of mobility radeon HD5870 on Czech website.


For me, its actually little performance dissapointment. Its on par with GTX 280M in crysis high details, GTX 285M is 10% faster. Full HD medium details is HD5870 loosing a lot compared to GTX 28XM. Hopefully new drivers will boost it performance up (or not)!

Check it out here: http://www.ddworld.cz/notebook/note...cenove-dostupneho-herniho-nb-na-svete--5.html

 

Even despite performance that is on par to a GTX 285M, it has DirectX 11 support, which a GTX 285M does not. That's a fairly big selling point to most people.

 

Hmm I can't get the ddworld.cz site to load in Chrome or Firefox. Will check back later.

 

Go over to the Asus section. There are all kinds of benchmarks and reviews over there.

 

Not impressive at all.. and the comparison is stupid.. 1366x768 on the asus with 5870 vs 1920x1200 on the d900f with gtx280m.. is this supposed to be a joke or a marketing thing?
Not even slightly interested in upgrading my 280m in D900f with either 5870 since its no better (sure dx11 sound fine but with such weak fps already at dx9 and dx10 max resolution i dont think it can run dx11 at the highest settings so its worth crap to me) and not with 285m since its just a slightly higher clocked 280m and i can get higher clocks on my 280m with a little overclocking..
Both ATI and Nvidia havent come up with anything worth buying..

 

Hard to suggest someone needs to upgrade from a 280 at this point. If you like to benchmark then yes the ATI card will blow away the 280 in synthetics.

BTW, the Asus marketing info is total crap and full of lies.

 

Mandrake, could you please send me link with gaming benchmarks? I cant find real benchmarks except synthetics crap. Crysis, COD, Crysis Warhead, FarCry etc. Thank you in advance.

I would like to know if its really more powerfull compared to GTX 285M.

EDIT: I am also interested in temps. Thank you.

 

Sure HERE. All the info you need is in there. The 5870 is only more powerful in synthetic benchmarks. In gaming benchmarks it really depends on the game itself.

 

Lol this is like music to my ears ! So can we say that the 5870 isnt as fast as the 285m In real world performance I.E Gmaing ?

And as for the DX11 content who really gives a toss at this stage ? DX 11 is hardly a huge performance increase in games, and secondly as said above DX 10 at high or DX 11 at low ?

Ill let you be the judge.

 

DX11 is more for realism than performance.

Have you seen this yet? LINK

 

Wow that vids awesome,
DX11 will definately become mainstream if its offering changes like that, and triangle/poly count is off the roof :S.

And guys, where you getting the 285m information from?

 

WOW I haven't seen that although I have seen comparisons between dx10 and dx11 games and it didn't look that impressive.

 

DIRT is the only game that supports DX11 that I know of, it would be cool to see that game in DX11 and compare.

 

http://www.youtube.com/watch?v=fEC0WAF2NIg

http://www.youtube.com/watch?v=riOGnywckiA

http://www.youtube.com/watch?v=NlyKmFLi-e4

 

There's also battleforge and Stalker:Call of Pripyat, both already out.
Aliens Vs Predator is due soon, and more DX11 titles on the horizon.


Oh, also official ATI mobility drivers as of 10.3 due next month, so the future is bright.

Edit: Just read the site linked with benches. Jesus christ those benchmarks are AWFUL! They're not even worth comparing, they've used a smattering of machines with different games at largely different settings, utterly utterly pointless.
The news from the Asus thread is positive on heat, minimum fps etc so I'm going to stick with the opinions over there until we see something more reputable.

 

What's so real about 10fps? It is going to take a while for notebook hardware to catch up enough to make DX11 truly realistic. Unless you have that psychological disorder in which your vision is constantly blacking out- then it would seem real. Even on that desktop system the guy was demoing on, DX11 was a bit less smooth than 10. No thanks. Performance IS realism in my book.

 

Looks like the difference between using displacement maps and not. Other than that I don't see any major dif. Kinda sad to learn that my brand new D900f with the 285 will never support Direct X 11..

 

wow that dragon is amazing. but how much more horsepower would you need to animate it?
also with consoles still ruling the gaming industry (and which doesnt support dx11), do you guys think crossplatform developers would have the motivation to implement these features for pc's? i dont know how difficult it would be, but it looks like it will take more effort from developers to make all those triangles and such.

 

I guess my Asus is better than that desktop. I was getting 20 at full res and 30 at the default settings.

 

Those vids do look great
Maybe DX11 is something to look forward to.

 

If Nvidia drops the process manufacture to 40nm and grafts on DX11, then I don't see why they can't compete. DX11 may "seem" far off now but it's not really. Both next gen consoles will have DX11, end of 2010 or beginning of 2011, so 9-12 months from now DX11 will be the only thing out there.

Oh and DX11 is also about splitting up the workload better on multiple cores, so it goes hand in hand with quad cores.

 

Bench mark comparison from notebookcheck.net:

 

You have to laugh though when you consider the single 5870 is beating the 3870 crossfire setup by a reasonable amount, and isn't far behind the other SLI setups either

 

Yeah!, as long as it's used with GDDR5 memory it seems quite promising.

 

Yep, its basically a 5750 with a 5770s shader count, in a mobile package. Overclock these to roughly 5770 clocks and you're laughing, 2 of these overclocked well together would be roughly equivalent to a 4870x2 or somewhere in the 5850/5870 desktop region

 

Pretty impressive! Crossfire is the way to go then. Just hope the notebook OEM's can implement it within the thermal budget.

 

8991 on a 4870 x 2 setup? I don't think so. Try 11,300 by default and >14k overclocked. Those numbers are unreliable. In gaming benchmarks published in the Asus thread, the results are pitiful compared to the 4870 Crossfire setup I have. The BF BC2 numbers were especially low. The 5870 seems about 10-20% (20% is rare) faster than the 4870 at best. I'll know soon enough when I buy a couple from RJTech. The 5870s may excel at benchmarking but real world gaming results are showing it's nothing special. An upgrade for sure but nothing to get overly excited over.

 

I thought Axelrose1uk meant the equivalent to the desktop 4870x2 score not the mobile version.

 

I was referring to the notebookcheck numbers.

 

Actually I was referring to the desktop model. 2 overclocked 5870m's would be in the same ballpark (as in a crossfire setup) as a 4870x2 desktop card. An overclocked 5870m is very similar power wise to a desktop 5770, ergo 2 of these in crossfire are similar to a 4870x2/5870 (at stock speeds they'd be closer crossfired to a 5850)
A 4870x2 gets around the 15-17k GPU score, we've seen a single 5870m hit 8600, so dependant on crossfire utilisation, they'd be quite similar and results would tally as expected.
I have nothing against the 4870ms, I would have quite liked some, and negotiated for an M17x with them before Alienware pulled out of offering me an M17x as an apology for thier messups with my M17.
I wasn't discussing the 4870x2m setup however.

That is what I was on about. There seems to have been some wires crossed above, having had 4850ms and a 4870 desktop card as well as access to a 5750, Im well aware of the rough performance regions of these cards

 

Alex, you're referring to vantage numbers which are nice for benchmark chasers but real world results show the 5870m's only being marginally faster than a single 4870. Again, I refer everyone to the Asus threads where benchmarks with the 5870m are plentiful.

 

I know the 5870m is about 10% ahead of a 4870m, however I'm not talking about the 4870*M*s, and I'm talking about a crossfire setup, not a single card.

I'm talking about specifically in relation to the desktop cards. I was also speaking about an overclocked setup. A well clocked 5870*m*, is roughly on par with a DESKTOP 4870, ergo a crossfire setup would be roughly in line with a desktop 5870 or 4870x2.

Make a bit more sense now? I was comparing them in terms of how they'd stand against a gaming desktop.
Sorry if this wasnt well explained and caused confusion.