It looks like the HD2600XT is rubbish *(moderator note: unreliable source)

Okay, up till this point, I was seriously torn between a Toshiba X205 and a HP HDX. The Toshiba is £1399 and the HP is £1699. They both have their graphics manufacturers top cards (i knowthe 7950 is faster but thats not what nvidia say).

Anyhoo, I want a good solid gamer. Not bleeding-edge quick so dont want SLI. I've had a clevo based gaming notebook in the past, but I want more of an entertainment all rounder now that can do pretty much everything.

Anyways, while the 8700mgt seems to be a solid enough card with it doing well in games (loads of X205 owners it seems), there are hardly any laptops with the HD2600XT. I've found reviews of 2, that are quite damning of it.

1st, HP HDX. In Cnet's review of it, they tested Quake4 at max settings at 1024x768, and got average of 39fps. Now Quake4 is an old game and it was tested on low res, thats seriously disappointing. I saw another review of the HDX that tested FEAR at 1024x768 with 25fps. Again, outrageously slow for a top-end card.

The new Imac uses an underclocked version of the HD2600XT (they call it HD2600 Pro, who knows why?). Again, reviews state its graphics performance is average and less than that of 8600MGT which is similar in price apparently.

This is all very disappointing. Seems like Ati are dropping the ball. I love the HP HDX (fell in love after seeing HDDVD on it). Was seriously considering buying it, but I am not paying £1700 for a laptop that cant game properly. Damn HP for not having nvidia cards.

finally, I have a suspicion that the low performance is down to driver issues (so do the one of the reviewers). Having said that, when you paying this much money for a laptop, you dont want to have to sit and wait (prey) for a good driver.

I remember when AMD and ATI were both at the top of their game. Looks like it is downhill for both

(edit. sorry, i didnt realise cnet were such an unreliable source)

 

Their desktop HD 2900 XT is actually pretty good card. I'm not sure if they have it for mobile notebooks, but it really is pretty good.

It's better than the 8800 GTS, and better in SOME games against the 8800 GTX. (Note: SOME, NOT ALL)

 

I would still prefer a user based review rather from Cnet. Or it could be the driver isn't fully optimized yet.

 

CNET is not very reliable...

 

I'm pretty sure the iMacs use desktop GPUs this time, and there is a HD 2600Pro variant. I believe the desktop HD 2600XT is supposed to compete against the desktop 8600GT and the HD 2600Pro against the 8500GT.

ATI's HD 2xxx series uses a VLIW architecture like the GeForceFX and Itanium. Admittedly, not the best company to keep. VLIW is very dependent on a good compiler in the drivers that is needed to extract maximum parallelism for the GPU to work properly. This is different from how previous ATI GPUs operated, so the drivers are comparatively immature. There is potential there, but by the time that it's revealed, the HD 2xxx series probably would have been replaced by the next iteration.

 

Yes I will second that also, but to the original poster please do some actual research before creating this sort of thread. This why I wish negative approval ratings were still around....because this is more of a rant than a legitimate post.

If you looked at the second most popular thread on the forums (HD2600 Performance Thread)( http://forum.notebookreview.com/showthread.php?t=150512) you would see that many of the benchmarks you posted for the HD2600XT were already bested by the HD2600 GDDR2....therefore they were way off/inaccurate.

I'll give you the benefit of the doubt this time but please do not go around quoting CNET (notorious for being wrong) and/or comparing desktop GPUs to mobile GPUs (which are always substantially different). I.e. the HD2600pro isn't a notebook card and will confuse people especially on a notebook forum.

 

CNet is in my opinion one of the worst tech websites.. I don't trust them one bit.

 

Actually, this isn't entirely accurate. It's faster in many benchmarks than the 320MB version of the 8800GTS, yes, when the resolution is high enough to bring that factor into play. But it is faster than the 640MB 8800GTS less than half the time, and yet costs almost as much as the 8800GTX (which it almost never outperforms - in fact, only in Source engine games does it hold a candle to the 8800GTX).

It's not surprising, then, that the HD 2600 is also slower than the 8600M. As for the 8700M - well, the two cards aren't even in the same class.

 

Quake 4 is a new game, is it not? I'm surprised a super recent gamer like yourself hasn't purchased the only decent GPUs - a 8800 or 2900.

 

Here are the benchmarks, and note; the XTX is the current XT, this was before the card were released.
http://www.dailytech.com/ATI+Radeon+HD+2900+XTX+Doomed+from+the+Start/article7052.htm

 

I don't think that's true. The standard retail HD 2900XT is still the same HD 2900XT that it always was using GDDR3. The HD 2900XTX was uses GDDR4, which is available now, but only as special overclocked editions of the HD 2900XT. The HD 2900XTX was never launched.

ATI's official retail variant is here with 512MB of GDDR3:

http://ati.amd.com/products/radeonhd2900/radeonhd2900xt/index.html

 

this is nonesense...there is no proof at all the 2600xt is rubbish, i would like to hear why

 

I read it from some magazine, comparing it with both cards at different resolutions testing games, it was a UK magazine, when i was in England.

 

im pretty sure the XTX isnt even out yet because of this.... the XT running DDR4 is our which took the place of the XTX here is the like of where to buy it:
http://www.newegg.com/Product/Product.aspx?Item=N82E16814102099

 

http://www.anandtech.com/video/showdoc.aspx?i=2988

Here's the review from Anandtech (which, unlike CNet, is a reliable source) showing the rather lackluster performance of the HD 2900XT compared to the G80 cards.

 

oh, sorry guys. i didnt know that cnet was an unreliable source. ive been looking all over for reviews, but seems that the hp HDX is the only review that has this card in it. anyways, the other review i read is here:

http://computershopper.com/reviews/hp_pavilion_hdx_entertainment_notebook_pc_review

in that one, they say that they get average fps on FEAR on autdetect settings at 1024x768. again, i dont know if computershopper is reliable or not.

anyways, sorry i didnt realise that cnet was not a great source. but alot of reviews of the HDX have said that the graphics were average.

 

http://www.bit-tech.net/hardware/2007/08/14/radeon_hd_2600_xt_vs_geforce_8600_gt/1

Take a look to see what you think of this indept review....

 

Are you serious?
The lowest 8800GTX cost about $500 while lowest price on the 2900XT is $360..I'm getting those figures from newegg. Nice FUD. That's quite far removed from "cost as much". More like apples and oranges.

 

I was merely going with the last figures I'd seen, I wasn't aware of the large price drops on the 2900XT. Of course, it still costs more than the 640MB 8800GTS. Which - at least according to Anandtech, Gamespot, and several others - is also a vastly superior card.

 

The only thing that held me back from buying a HP HDX was the graphic card. I mean the HD2600XT is utter crap. Right?!?! I've read it here and on other forums. Well i finally decided to buy the HP knowing that i am probably not going to be able to play games on it. But i just loved everything else about it.

So i bought it! And the HD2600XT has exceeded all my expectations. I am running Half Life 2 EP1 native resolution (1650 x 1050) smoth as silk. Same with Quake 4. IT LOOKS WONDERFUL.

Later this evening i am going to install FEAR.

I just love this machine.

 

Don't believe everything you read. Investigate and find out which source(s) those people were referring to. Probably the same place as in this thread (ComputerShopper/Cnet). Mainstream sources do not know how to benchmark right.

 

Stop comparing desktop cards on a notebook forum, I just said that five posts above yours read the whole thread!

 

benchmarks and/or fraps pics if you could.

 

yeah mate, if you could post some benchmarks, that would be great. still deciding to get one or not

 

I have been testing Half Life 2 Loast coast demo on the HDX.

Resolution 1680 x 1050
Every graphic option set to "Full" including HDR Lighting except for:
Antialiasing OFF and Trilinear filter.

58.26 FPS

Going to test Quake 4 and FEAR tomorrow.

Bye!

 

thats a good score no? how did you test? is that an average fps or maximum. from that, it seems that the hd2600xt is just as good as the 8700mgt

 

They do same thing to Aspire 5920G.All review sites say it's cool and good,but cnet said it's not worth buying

 

I am very happy with the performance of the 2600 so far.
There is a test in Lost Coast that measures the average frame rate. I have to work this evening so i will do some more testing tomorrow.

 

HD2600XT seems to have 128 stream processor vs only 32 on the 8600M GT.
What is the performance not too difference??

 

And you're going to be?
Nice Avitar by the way.

Got some RECENT benchmarks to back that up?
Like something with a 7.8 or 7.9 driver?
Even going back to early bnechmarks by places like Xbit Labs the HD2900XT-512 was already outperforming the GF8800GTS-640 more often than not, but you are right about the price being more than the GTS-640, but nowhere near the GTX.
I think you need to look at some benchies that are a little more recent than those at launch. Just like the GF8800 series the HD2900 benefited alot from driver updates.

That's a pretty open ended statement considering that the GF8600MGS is such a cripple. I doubt that the plain Mobile Radeon HD2600 is outperformed by the GS on average. I wouldn't be surpised that the GT might outperform, but the 8600 mobile line is far different than it's desktop cousins.

I think you need to do a little more research, just like the OP.

Right now the biggest drawback for the mobile HD2600 is the slow release of drivers by the various OEMs. I hate that AMD/ATi and nV don't release the mobile drivers as quickly or easily as the desktop drivers.

 

I assume that question is.... WHY is the performance not too different, and that would be because of a few things;

A) the HD2600 is extremely held back by it's low ROP count with 1 ROP cluster (of 4).

B) the GF8600GT's 32 stream processors run faster than the HD2600's processors, and also run dual issue.

C) The HD2600 doesn't have hardware AA resolve in the ROPs and uses the shaders to apply AA, while the GF8600 has dedicated hardware in it's 8 ROPs.

Those two factors make the architectures perform very closely to each other despite each having this or that benefit on paper.

Performance depends alot on the type of game and situation they all play to different strengths.

 

I realize that the topic is about the MR HD2600XT, but just for all you ATI haters, I have an HD2600XT in my third desktop, just bought it 3 days ago. I've been runing tons of stuff on it, HL2, FEAR, COD 2, AOE3 etc. I'm getting 110fps in HL2 lost coast at max w/HDR, 16xAF. But when I make AA even 4x, it KILLS the performance.. like down to 25fps. It's the same way with FEAR, at max it runs smooth, with no hiccups under 60fps, yet when I turn on AA to any setting, the FPS drops significantly.

Seeing how the notebook varients are heavily based on the desktop cards, I'm assuming this is the issue with the notebooks as well. The HD2xxx series aren't bad, infact they're great, you just can NOT use AA! Which really isn't all the big an issue, most people keep it off anyhow.

 

The card is good for gaming, but the notebook variant needs more time to prove itself against similar nvidia cards for a good objective comparison. That isn't to say that people with laptops using this card have a cripple, they should be very satisfied with the quality.

 

i think AA will be needed less with directx 10 which is WHAT THESE CARDS ARE MEANT FOR. i recall that it looked as if AA was being used in Bioshock in directx 10 even though there is no option for it.

 

I thought AA was a must in DX10 (that's y there was no option)...
benchies with the HD2600 in Bioshock is very welcomed.

 

my general point was that it was forced without causing a performance drop which is the point of directx 10

 

I am having a hard time getting bioshock to run in DX10 mode. I have installed the latest Catalyst drivers 7.9 but it still does'nt work. A lot of graphic is missing and a lot of areas are for example completely black or brown.

Anyway, the first fraps fps recording on the HDX gives the following results:

2007-09-13 23:00:42 - Bioshock
Frames: 2488 - Time: 60000ms - Avg: 41.467 - Min: 29 - Max: 56

1024 x 768 with the all three sliders on High on DX9.

Honestly i was a bit surprised because the game feels slow and sluggish and not very snappy compared to Half Life 2.

Can Fraps give wrong numbers?

 

ati's drivers just suck i think, try on direct x 10 mode it should run better. why was it not working in direct x 10 mode?

 

Well, something is wrong with Fraps, here is the figures when running Bioshock at 1024 x 768 all sliders set to Medium. Now the game feels quick and playable. But the FPS is almost the same as on high settings.

2007-09-13 23:29:40 - Bioshock
Frames: 2771 - Time: 60000ms - Avg: 46.183 - Min: 34 - Max: 61

Strange?!?!

 

the HD2600 does have problem with dx10.. have u contacted ati? (i doubt hp would reply you.. but try it oso, since it's their top end)

 

It's drivers related, since all DX10 games had problems with the 2900, a driver released by ATI 3 days after game's released ficxed them

 

a new version of fraps released recently. (v2.9.2).