Ati HD 3670 better then Nvidia 9600m GT !

so tomorrow my friend will put his laptop together, (he will get the ati 3670 card, he could get the 4670 (if he paid some extra money) but he is not some super-gamer and said he could live with 3670 without any problem.

i knew that ati 3670 512mb and nvidia 9600M gt 512mb was pretty similar, but ati 3670 performs actaully better, (quite much better).

im actually impressed of the ati 3670........... nice work AMD !

Benchmark is brought from Notebookreview.com (of course )
--------------------------------------------------------------------------------------

 

There's 2 versions of the 9600m GT, one using DDR2 and the other using GDDR3. The GDDR3 version performs better than the Radeon 3670.

 

I'd take that benchmark list with a pinch of salt.

For starters 3dmark06 standard res is 1280*1024. Running at a lower res skews the results higher than they should be.

Secondly there's no way a 9600 gt and a 2.8Ghz is only getting ~ 4000 points.

 

@Lithus, you wrong. that HP HDX uses GDDR3 in the test. (you really think HP would use GDDR2 inside their HDX-Flagship ??? no man).

like i said, Ati 3670 get +700 points more then 9600M GT in 3dmark06, which is pretty impressive !

here you see, same HP as in the 3dmark06 test. (also from Notebookreview.com)

 

I have compared the 3670 with a 9600gt with ddr3 and I can tell you in most games the 9600gt performs better (sometimes much better). The 3670 can be faster in 3dmark (depending on the drivers) but not in actual games. A reasonable score for a 9600gt with ddr3 and T9600 at standard res (1280x1024) is around 5100-5200 3dmarks (i had around 5150 on my hp). The 9600gt with ddr2 is much slower and is indeed slower than the 3670.
All in all the 3670 is a very good card but is not faster than a 9600gt with ddr3, except maybe in valve engine games which are optimized for amd hardware.

 

No the HDX series used the DDR2 version of 9600M GT.

 

The 1GB 4670 is better than the 9600GT rite?

 

Yes, much better. No doubt about that.

 

Even the 8600M GT 256MB GDDR3 in the XPS M1530 outperforms the 9600M GT 512MB DDR2.

 

i agree

and when comparing ATI with Nvidia, you can have very different results, because they perform different with each game or application, so some times it will perform better, but some times it wont.

 

Exactly, some games just run better on ATI and some games run better on Nvidia. Same goes for benchmarks such as 3DMark. Sometimes it helps to look up benchmarks for the particular games you plan on playing. Take two equally priced graphics chips, one from Nvidia and one from ATI, and just because one chip is better for someone who plays these games it may not be the best option for someone who plays those games.

 

Pranalien....are you blind ???

first of all, that HP HDX 18" who is in test use GDDR3, (i even post the spec, how hard can it be ???)
http://www.notebookreview.com/default.asp?newsID=4610
look at the HP HDX 18 review, its clearly says it got 4100 points in 3dmark06. , Ati 3670 got 4850, that 700 points better.
http://www.notebookreview.com/default.asp?newsID=4610&p=2

its nothing hard why Ati 3670 is better. Ati 3670 use 378 million transistors, (9600M gt use 314 million), that one of the answers why Ati 3670 is better.

both 3dmark06 tests gave same scores on 9600M gt DDR3

 

Dude sometimes reviews are wrong. I don't know what they tested, I don't know what 9600GT the hdx has but i'm telling you, I tried it myself! 9600GT with ddr3 and T9600 gets 5100 3dmarks, not 4000.
3670 is not faster than the ddr3 version of the 9600gt on average. It is about 10-15% slower. I tested many games (crysis, world of warcraft, half life2 etc). It may be BETTER depending on what you want in a video card (drivers, video decoding, power consumption etc) but it is NOT faster (on average). I'm an ATI fan myself, but facts are facts.
Also the transistor count argument is irrelevant. The GTX200 series chips from nvidia have around 1.4 billion transistors and the ati 4870/4890 chips have ~950 million and yet they perform very close to one another. Transistor counts by themselves mean little, it's also about chip architecture, frequency, memory subsystem, drivers, etc.

 

Remember, sometimes graphics chips are underclocked by notebook makers to be able to run cooler. As well, it also depends on the driver version used, and there are many other factors as well. The chipset in the laptop, the CPU, the bloatware on the laptop, etc. Also the power settings in Windows can make a large difference as well.

 

well, Ati 3670 have more transistors.....way more pipelines......higher Watt.

im not chocked 3670 beats 9600M GT in 3dmark06

 

You have no clue what you're talking about. I don't know why you're all riled up for the 3670, but it easily loses to the GDDR3 version of the 9600m GT.

http://www.notebookreview.com/default.asp?newsID=4628&p=2


http://forum.notebookreview.com/showthread.php?t=343349


Now stop talking about transistors. You have no clue what they are.

 

I ran 3dmark06demo on my 9600m GT GDDR3 in bootcamp at 1280x1024 and i got a score of 5533 points.

 

I was going to post scores from the MBP (they're much easier to find than MSI reviews), but I was worried about the effect of the 9400 on the 3DMark06 scores. Does the MBP use both cards in Windows, or is it just limited to using the 9600?

 

Just the 9600m GT in Windows at the moment , but we have hopes of using the 9600m GT and 9400m in hybrid SLI might be possible, see post below.

http://forum.notebookreview.com/showthread.php?t=410230&page=3

 

Technically, most ATi cards are using more processors with a greater collective output. Ever do the FLOPs calculation for some of these ATi cards? The DDR3 4870 crushes the 280m, in theory, but when it comes to actual games - not the case - the 280m exhibits a clear-cut advantage.

ATi cards have a well-forged reputation for doing mildly-to-significantly better than their nVidia counterparts in Vantage benchmarks. It's always been the case with ATi cards scoring higher than the nVidia cards but with in-game benchmarks, the tables are turned. Maybe it's an issue of driver support, or maybe it's the programming behind these benchmark utilities.

However, synthetic benchmarks and static tests are not the full picture, period. Too many people are caught up in this hype.

This whole competition is meaningless, and ultimately, futile.

 

well, looks like all people get different scores with 9600M GT (it depends how good CPU you use of course)...............
anyway, with a bit OC on 3670 you get 5820 points in 3dmark06 you need a good cooler though.

 

LOL congrats, an OC'd card performs better than a stock card. What a revelation.

GDDR3 9600m GT > ATI Radeon 3670 > DDR2 9600m GT.

Which is exactly what I posted on my first reply. I'm done. Thanks.