ATI 2600HD Review

ATI 2600HD Review

By: Crimsonman

I purchased an HP 8510p with

T7300
3 gigs
ATI HD2600 GDDR3
Vista Business
160 gig 5400 RPM harddrive

PEOPLES: I will add some stuff, so it will be more detailed as time goes on.
This is a review for the ATI HD2600 (or 2600HD, same thing). ATI 7.9 catalyst drivers. I won't be reviewing many flashy games. Just Medieval II: Total War, Medieval II: Total War Kingdoms, Two Worlds, and Age of Empires 3. As I get more games (Empire Earth 3, Assassins Creed) I will put them up. Also, I don't think 3Dmarks show jack$h!t, so I won't be using those. In my experience using this, the card doesn't get hot at all, it just spits out warm air. After 8 hours of me taking screen shots, okay it got tired, but it really stayed cool.

GPU-Z Thingy Magig

Too unclear to post, so its attached at the bottom, this card is insanely underclocked, according to it.
3Dmark06

Battlefield 2142:







The only time there was frame stuttering was when a helicopter kept on blasting me with its trillion missles.
Medieval 2 Total War:


Recommended settings, all high, except textures, with 4 AF. pretty damn good in my opinion.

Medieval 2 total war kingdoms





Same for Kingdoms, except it has some fight screens.

Two Worlds


Just a nice view.

I hit that cyclops once, then it swung its arm and I flew about 10 feet to my death.


This is with 3 gigs of RAM. I tried 4 AA and I had the same framerates. But there seemed to be no lag at all, even with 15 FPS. I also tried 1440x900 res with no AA and I think it was in the low 20s

Age of Empires III:



I know in this one bloom and water reflections are off, but in the pictures, its on.


Okay, in this one I actually gave up because I was getting attacked so much, but this is me, Mr. Orange (I'm always orange), killing the English.


Yep, those are my ships getting beaten.
This set is of textures on normal and with the others on high, 1440x900 res, and shader on very high. At first, when I began I noticed lag, but then I realized I was hitting the edge of the map I noticed minimal lag around the sea, which went away after playing, even with the titanic frigate shooting. This was at Carribean. The frames never went below 20, except when I took screenshots, it went to about 18 then back up to the 25 area

A note about this game: I hate Napolean and Frederick, evil bastards, they sank my only ship.
Supreme Commander
1280x800 res, all medium except for shadow fidelity which was low and fidelity, which was high. No AA and Vsync was on


Frame rates kinda tanked when I went all the way out.



F.E.A.R.


****Seraph0000****, not mine
max settings with no AA for now, 1280x900 res

Conclusion:
I have to say, this card is essentially the same as the 8600 GT DDR2. I think you could claim this card as an 8600 GT DDR2 1/2. It plays games exceptionally well. All people who underestimated ATI, I'm laughing at you, hahaha. I can't believe people said that it was comparable to the 8600 GS. Anyway, this card rocks. If you want something non-Nvidia for an 8600 GT, you've found your call. Also, comparing desktop cards and then saying they'll perform proportionally as well as the laptops, is wrong.

I have to give ATI mass amounts of credit for this thing. It is a powerhouse.

If there is anything you want (I won't do demos, it takes forever to load, 10 hours for World in Conflict), tell me.

 

I cant belive you didnt put a 3dmark05 and 3dmark06 score in there, you know its the most static thing that people will use to compare it to other cards.

Also you need to overclock it, if my theory is right about the 2600, it will beat a 8600gt because its on a smaller diameter processing and thus will produce less heat and use less energy and these things always lead to a higher overclock. So this card should overclock like a monster.

 

I dont think they mean anything in the real world, just like high school. No, I wont overclock it either. (I know, I'm a pain)

 

I'm also thinking of buying that notebook. But to be fair on Nvidia..........the 8600GT DDR3 does perform better than the HD2600 DDR3. One might deduce that a HD2600 DDR2 will be a little slower than the 8600GT DDR2.

Although for some reason the DDR3 8600GTs are quire rare so that makes the point somewhat moot.

 

I gave rep points, although I was hoping for a little more. I.e. you didn't mention clock speeds, include 3dmark, or the F.E.A.R demo (a well used benchmark).

 

Well, yeah, GDDR3 outperforms it no doubt, but it seems as if its like a midsegment for the cards

I tried GPU Z and I got lke 252 mem clock... so I was like, no thanks

I think its 400/500

 

GDDR3 at 500 isn't really taking full advantage of it, did you try any overclocking results per chance?

 

3dmark doesnt show true life performance always but its a good view because its static, just going into a game and playing with fraps is not static as in no two test will be the same.

As the poster above mentioned the fear benchmark is popular because of the reason that it is static. Its not going to hurt to give your own little personal review of the card, I was just giveing you pointers on the things you would need to achieve a perfessional review.

Then from there you want to mention your video card drivers, the tempatures and things of that nature, the things people will need to compare there card to yours.

 

Okay, Ill load fear, although I dont like those kinds of games.

anyway, how do I measure temperature?

 

Software, chances are you ATI control panel has it. I have not used ATI before so I am not sure. I know Nvidia has it also there program called nmonitor can even shot a graph of tempatures over time.

 

How about instead of FEAR, I use CS:S or some battlefield game?

 

If you have time, could you check the Battlefield 2 demo?

 

surely. I'll do BF2142 as well.

 

I have an Asus A7k with a 2600... CPU is prolly a bit slower, but I'll post some results too if anyone cares... Kinda seems like his is running a bit slower than mine, but I could be wrong.

 

it might also be DDR2 instead of GDDR3, mine is the latter

 

mine's gddr3

 

Look at my review for 3dmark06 scores...they're all about the same for each card.

Anyway I firmly believe that as our HD 2600 is underclocked 200mhz core/mem it's taking performance hits and making people go UGH. If we could get ATI Tool or Tray Tools to work properly and OC it to stock clocks...Then I think we'd see some really interesting numbers vs the 8600 GT GDDR3.

Also nice review

 

So, as it stands, is it on par, or better than a 8600M GT 256MB GDDR2? That's on the shipped clock (be it stock or underclocked).

And also, how much did that HP Lappy cost you?

 

I used the driverheaven modder thing to mod the newest Catalyst driver off the ATI site. My vista gaming score went up, but I can't access the CCC. Is it supposed to do that?

 

It cost me 1479. deal ends the 31st of october

adding bf2142 screens in about 5 min

 

Nice nice .

How about answer my first question, in your own opinion?

Is the 2600HD better than the 8600M GT GDDR3? Or is it more or less on par?

And how easy is it to overclock ATi cards (if you have any experience with OC'ing, that is)?

 

does this help anyone any? i was told 1280x1024 is the standard resolution... but 1280x900 is whatcha got lol... so... deal with it

*edit* lemme through the gpu-z stuff on here too...

 

ITs less, but it is still very good. well worth the money. OCing is something I am clueless in. But, since its so underclocked, Im guessing you can OC the crud out of it

 

Wow... That's' some crazy underclocking!

Why the heck would they do that? Even my GDDR2 has the same bandwidth as your GDDR3 with the 2600HD's memory only at 400Mhz! And the 8600M GT has 3 times the pixel and texture fillrate, too... That's odd, seeing as how you have 120 unified shader pipes, while the 8600M GT only has 32 !

EDIT: It does seem to score higher on 3DMark, though. The general score for 8600M GT is around 2900-3100 at 1280 x 800... I guess it's true what people say: 3DMark scores aren't all that meaningful in real-life performance.

 

It's been running all my games really well. Let me grab some F.E.A.R. benchies... I'll run max settings with no AA for now... I'll do more later if you guys want. One minute...

*edit* okay here we go... it runs way faster at slightly lower resolutions, but just to keep it uniform I ran it at 1280x900

 

Are you talking about me?

I guess, the possible explanation for HD2600 mobility's relative performance would be that HD2600's less raw-power oriented design tend to be less handicapped by slower VRAMs on many laptops. Many 8600GS/GT are really crippled by their DDR2. Still, I think it is reported that AA will kill HD2600's performance (both desktop and mobility) more so than 8600GT, but I guess that is not very important for mid-range cards.

Although you say you don't do demo - can you bench recently released UT2K7 demo? Another thing I want to see are Stalker and NWN2. Thanks.

BTW, I think SPU-Z is showing wrong core speeds.

 

AMDGPU Clock Tool shows the same speeds

 

Can you run the FEAR benchy at 1024 x 768? That's the highest resolution FEAR Combat seems to support natively without any config file tweaks (also the resolution I play at since I'm too lazy for config file tweaks ).

 

yeah sure... I'm still up trying to find a good overclocking/temp monitoring tool that works

here you go... max all cept AA(none) at 1024x768

 

okay, I own NWN2, so I'll do it. And I guess Ill do a UT2K7 demo

 

just wondering did you guys run tests under vista or xp?
also do you have soft shadow turned on in fear Seraph0000?

 

vista... and yes soft shadows was on... i had everything on cept AA

 

i use vista too (it says it in the review)

 

got this with clock speeds of 587 MHz / 814 MHz... I'm a big noob with overclocking... but I got this without any artifacts or anything, and the card runs pretty cool(and I'm using a zalman laptop cooler)

I haven't seen temps get over 61 C under load so far. I used AMD GPU clock tool to get this... Overclocking my 2600 has been kinda finicky. These settings seem to run pretty stable. Not sure what all I can do with drivers and more OC'ing software. There aren't many programs out there that work for this card atm. So anyways... I thought this was a pretty good score, but I'm sure someone else could do better.

*edit* by finicky I mean that the screen will black out and come back with the clocks reset. This happens when I underclock it(or bump the settings back to what they were from an overclock) too though so I think it's just the tool that's giving me the problem not the card. I'll experiment more later and see if I can push it some more. I'll post FEAR benchmarks later too (they went up to an average of 32fps with max settings no AA @1280x900)

 

the default res u test ur gpu in 3dmark is 1280*1024...i don't know where u came up with those scores...those should be a little more on that resolution as most gt's score in that range on 1280*1024. good review, by the way, crimsonman

 

I ran 3dmark at 1280x900... That's the default for me... I think I'd have to plug it into an external monitor to get the "standard". And I'm much too lazy to do that right now

 

if my internet wasn't so ****ing slow then I would download 3dmarks, but it takes like 4 hours.

NWN2 should be up today or wednesday (busy tuesday)

 

That's a great score. Compares to my 256MB Quadro FX 570m on stock clocks with gaming drivers. Very nice OC and score. Amazing temp as well.

 

I'm pretty sure I could get it higher... I'll play with it some more... I just got the COD4 demo(which ran really smooth btw not OC'd) and the UT3 demo... I'm gonna run fraps in both games with normal and OC'ed settings. I'll let you guys know later what happens.

 

GPU Z doesn't show the right clocks (it's not 300/400MHz). Stock clocks should be 500/700 (I think), but HP underclocks it. As far as I know, nvidias midrange cards take a larger performance hit with AA (the gap between ati and nvidia lessens).

 

well mines clocked at 300/400 according to the AMD clock tool as well... but when I overclock it GPU-Z still reads 300/400. stock clock is 500/600... at least that's what the AMD tool tries to set it to as "default", but they come set 300/400, and when I reboot it sets them to the underclocked speed.

 

Try setting them to 300/400 in atitool and run a 3dmark... Do you get lower results than stock?

 

ati tool wasn't working for me earlier... I'll try it tonight... I gotta go for now though

 

It might not seem that it sets the clocks... But it does.

 

@Seraph000: Wow, I think that's quite a bit better than my 8600M GT DDR2 !

I'll have to run another fresh benchy after a restart just to be sure--but it does seem like the 2600HD is better thus far.

EDIT: Oh, that's 'cause you overclocked it XD. Heh.

 

dude, yes:
1) he overclocked it
2) he keeps on benching it at 1280*900!(the default one that msot people bench it on is 1280*1024, hence the lower scores...)

 

I am curious bout this since i was (hm) considering the HP HDX packing a HD 2600 XT. But Hp (here in sweden) sayes it runs ddr2-memory (256mb).. Would that make a big difference compared to the the GDR3-version?
Or are they wrong? I thought i read somewhere that the XT version uses GDR3..?

//Fabian

 

no, they're probably right, the Xt is different than the non XT. Its performs kind of like a 8700 GT. GDDR3 and DDR2, from what we've seen, is like 600 points difference

@ Big V, it makes the scores higher, not lower

 

k... i finally hooked it up to my 19" and ran these two at 1280x1024

1280x1024 @ stock:


1280x1024 @ 575/800(tried to get similar clocks...):


1280x900 @ 594/819:

 

bloody hell, those scores are good.