Sign in with
Sign up | Sign in

Comparing The GPUs And Test Setup

The Fastest 3D Cards Go Head-To-Head
By

The test includes the 3800- and 4800-series models from AMD, as well as the GeForce 8-, 9-, and GTX 200-series boards from Nvidia. The lower limits are bound by the Radeon HD 3650 and GeForce 8600 GT.

Nvidia graphics cards
Chip Code name Memory GPU rate Shader Memory rate SPs
GeForce GTX 280 SLI GT200 1024 GDDR3 602 MHz 4.0, 1296 MHz 2214 MHz 240
GeForce GTX 280 GT200 1024 GDDR3 602 MHz 4.0, 1296 MHz 2214 MHz 240
GeForce GTX 260 SLI GT200 896 GDDR3 576 MHz 4.0, 1242 MHz 1998 MHz 192
GeForce GTX 260GT200 896 GDDR3 576 MHz 4.0, 1242 MHz 1998 MHz 192
GeForce 9800 GX2 2xG92 2x512 MB GDDR3 600 MHz 4.0, 1500 MHz 2000 MHz 2x128
GeForce 9800 GTX SLI G92 512 MB GDDR3 675 MHz 4.0, 1688 MHz 2200 MHz 128
GeForce 9800 GTX G92 512 MB GDDR3 675 MHz 4.0, 1688 MHz 2200 MHz 128
GeForce 9600 GT SLI G94 1024 MB GDDR3 650 MHz 4.0, 1625 MHz 1800 MHz 64
GeForce 9600 GT G94 1024 MB GDDR3 650 MHz 4.0, 1625 MHz 1800 MHz 64
GeForce 8800 GTS OC G92 512 MB GDDR3 730 MHz 4.0, 1825 MHz 1944 MHz 128
GeForce 8800 GTS SLI G92 512 MB GDDR3 650 MHz 4.0, 1625 MHz 1944 MHz 128
GeForce 8800 GTS G92 512 MB GDDR3 650 MHz 4.0, 1625 MHz 1944 MHz 128
GeForce 8800 GT SLI G92 1024 MB GDDR3 600 MHz 4.0, 1500 MHz 1800 MHz 112
GeForce 8800 GT G92 1024 MB GDDR3 600 MHz 4.0, 1500 MHz 1800 MHz 112
GeForce 8800 GT SLI G92 512 MB GDDR3 600 MHz 4.0, 1500 MHz 1800 MHz 112
GeForce 8800 GT G92 512 MB GDDR3 600 MHz 4.0, 1500 MHz 1800 MHz 112
GeForce 8800 Ultra SLI G80 768 MB GDDR3 612 MHz 4.0, 1512 MHz 2160 MHz 128
GeForce 8800 Ultra G80 768 MB GDDR3 612 MHz 4.0, 1512 MHz 2160 MHz 128
GeForce 8800 GTX G80 768 MB GDDR3 576 MHz 4.0, 1350 MHz 1800 MHz 128
GeForce 8800 GTS G80 640 MB GDDR3 500 MHz 4.0, 1188 MHz 1600 MHz 96
GeForce 8800 GTS SLI G80 320 MB GDDR3 500 MHz 4.0, 1188 MHz 1600 MHz 96
GeForce 8800 GTS G80 320 MB GDDR3 500 MHz 4.0, 1188 MHz 1600 MHz 96
GeForce 8600 GTS G84 512 MB GDDR3 675 MHz 4.0, 1450 MHz 2016 MHz 32
GeForce 8600 GTS SLI G84 256 MB GDDR3 675 MHz 4.0, 1450 MHz 2016 MHz 32
GeForce 8600 GTS G84 256 MB GDDR3 675 MHz 4.0, 1450 MHz 2016 MHz 32
GeForce 8600 GT SLI G84 256 MB GDDR3 540 MHz 4.0, 1180 MHz 1400 MHz 32
GeForce 8600 GT G84 256 MB GDDR3 540 MHz 4.0, 1180 MHz 1400 MHz 32

Chip Memory bus Manufacturing process Transistors Interface
GeForce GTX 280 SLI 512 Bit 65 nm 1400 MB PCIe 2.0
GeForce GTX 280 512 Bit 65 nm 1400 MB PCIe 2.0
GeForce GTX 260 SLI 448 Bit 65 nm 1400 MB PCIe 2.0
GeForce GTX 260 448 Bit 65 nm 1400 MB PCIe 2.0
GeForce 9800 GX2 2x256 Bit 65 nm 2x754 MB PCIe 2.0
GeForce 9800 GTX SLI 256 Bit 65 nm 754 MB PCIe 2.0
GeForce 9800 GTX 256 Bit 65 nm 754 MB PCIe 2.0
GeForce 9600 GT SLI 256 Bit 65 nm 505 MB PCIe 2.0
GeForce 9600 GT 256 Bit 65 nm 505 MB PCIe 2.0
GeForce 8800 GTS OC 256 Bit 65 nm 754 MB PCIe 2.0
GeForce 8800 GTS SLI 256 Bit 65 nm 754 MB PCIe 2.0
GeForce 8800 GTS 256 Bit 65 nm 754 MB PCIe 2.0
GeForce 8800 GT SLI 256 Bit 65 nm 754 MB PCIe 2.0
GeForce 8800 GT 256 Bit 65 nm 754 MB PCIe 2.0
GeForce 8800 GT SLI 256 Bit 65 nm 754 MB PCIe 2.0
GeForce 8800 GT 256 Bit 65 nm 754 MB PCIe 2.0
GeForce 8800 Ultra SLI 384 Bit 90 nm 681 MB PCIe 1
GeForce 8800 Ultra 384 Bit 90 nm 681 MB PCIe 1
GeForce 8800 GTX 384 Bit 90 nm 681 MB PCIe 1
GeForce 8800 GTS 320 Bit 90 nm 681 MB PCIe 1
GeForce 8800 GTS SLI 320 Bit 90 nm 681 MB PCIe 1
GeForce 8800 GTS 320 Bit 90 nm 681 MB PCIe 1
GeForce 8600 GTS 128 Bit 80 nm 289 MB PCIe 1
GeForce 8600 GTS SLI 128 Bit 80 nm 289 MB PCIe 1
GeForce 8600 GTS 128 Bit 80 nm 289 MB PCIe 1
GeForce 8600 GT SLI 128 Bit 80 nm 289 MB PCIe 1
GeForce 8600 GT 128 Bit 80 nm 289 MB PCIe 1

AMD graphics cards
Chip Code name Memory GPU rate Shader Memory rate SPs
Radeon HD 4870 CF RV770 512 MB GDDR5 750 MHz 4.1 3600 MHz 800
Radeon HD 4870 RV770 512 MB GDDR5 750 MHz 4.1 3600 MHz 800
Radeon HD 4850 CF RV770 512 MB GDDR3 625 MHz 4.1 1986 MHz 800
Radeon HD 4850 RV770 512 MB GDDR3 625 MHz 4.1 1986 MHz 800
Radeon HD 3870 X2 R680 2x512 MB GDDR3 825 MHz 4.1 1802 MHz 2x320
Radeon HD 3870 CF RV670 512 MB GDDR4 775 MHz 4.1 2252 MHz 320
Radeon HD 3870 RV670 512 MB GDDR4 775 MHz 4.1 2252 MHz 320
Radeon HD 3850 CF RV670 256 MB GDDR3 670 MHz 4.1 1658 MHz 320
Radeon HD 3850 RV670 256 MB GDDR3 670 MHz 4.1 1658 MHz 320
Radeon HD 3650 CF RV635 512 MB GDDR3 725 MHz 4.1 1602 MHz 120
Radeon HD 3650 RV635 512 MB GDDR3 725 MHz 4.1 1602 MHz 120

Chip Memory bus Manufacturing process Transistors Interface
Radeon HD 4870 CF 256 Bit 55 nm 965 MB PCIe 2.0
Radeon HD 4870 256 Bit 55 nm 965 MB PCIe 2.0
Radeon HD 4850 CF 256 Bit 55 nm 965 MB PCIe 2.0
Radeon HD 4850 256 Bit 55 nm 965 MB PCIe 2.0
Radeon HD 3870 X2 2x256 Bit 55 nm 2x666 MB PCIe 2.0
Radeon HD 3870 CF 256 Bit 55 nm 666 MB PCIe 2.0
Radeon HD 3870 256 Bit 55 nm 666 MB PCIe 2.0
Radeon HD 3850 CF 256 Bit 55 nm 666 MB PCIe 2.0
Radeon HD 3850 256 Bit 55 nm 666 MB PCIe 2.0
Radeon HD 3650 CF 128 Bit 55 nm 378 MB PCIe 2.0
Radeon HD 3650 128 Bit 55 nm 378 MB PCIe 2.0

Memory rate = DDR clock rate doubled; physical rate is half
DDR5 clock rate quadrupled, physical rate is one quarter
SPs = stream processors; P and V = Pixel shader and Vertex shader
TC = Turbo Cache
HM = Hyper Memory
OC = Overclocked (clock rate higher than default)
SLI = Parallel operation with 2 Nvidia cards
3 SLI = Parallel operation with 3 Nvidia graphics chips
4 SLI = Parallel operation with 4 Nvidia graphics chips
CF = CrossFire parallel operation with 2 AMD cards
3 CF = CrossFire parallel operation with 3 ATI graphics chips
4 CF = CrossFire parallel operation with 4 ATI graphics chips
R680 = 2x RV670
Shader 2.0 = DirectX 9.0; 3.0 = DirectX 9.0c; 4.0 = DirectX 10; Shader 4.1 = DirectX 10.1

Nvidia single graphics cards, AMD single and CrossFire graphics cards
CPU Intel Core 2 Extreme X6800 @ 2.93 GHz (11x 266 MHz),
Socket 775, 1.28 V, 65 nm, L2 cache 4096 KB
FSB 1066 MHz (4x 266 MHz)
Motherboard Asus P5E3 Deluxe, PCIe 2.0 2x16, ICH9R
Chipset Intel X38
Memory 2x 1 GB, Ballistix (Crucial Technology) 1.5 V,
DDR3 1066 7-7-7-20 (2x 533 MHz)
Audio Intel High Definition Audio
LAN Intel 1000 Pro
Drives Western Digital WD5000AAKS 500 GB, SATA, 16 MB Cache,
Hitachi 120 GB, SATA, 8 MB Cache
DVD Gigabyte GO-D1600C
Power Supply CoolerMaster RS-850-EMBA 850 W

Nvidia SLI graphics cards
Motherboard Asus P5N-T Deluxe, PCIe 2.0 2x16
Chip set Nvidia nForce 780i SLI
Memory 2x 1 GB, A-Data Technology 1.8 V, DDR2 800 5-5-5-18 (2x 400 MHz)
Audio ADI 1988B SoundMax
LAN Marvell 88E1116 Gigabit

Drivers & Configuration
Graphics AMD Catalyst 8.6
Nvidia ForceWare 175.16, GTX 260 and GTX 280 Forceware 177.39
Operating system Windows Vista Enterprise SP1
DirectX 10 and 10.1
Chipset driver X38 Intel 8.3.1.1009
780i Nvidia nForce 9.64

Ask a Category Expert

Create a new thread in the Reviews comments forum about this subject

Example: Notebook, Android, SSD hard drive

Display all 146 comments.
This thread is closed for comments
Top Comments
  • 20 Hide
    elbert , August 29, 2008 10:55 AM
    Version AMD Catalyst 8.6? Why not just say i'm using ATI drivers with little to no optimizations for the 4800's. This is why the CF benchmarks tanked.
  • 19 Hide
    wahdangun , August 29, 2008 11:07 AM
    WTF, hd4850 SHOULD be a lot faster than 9600 GT and 8800 GT even tough they have 1Gig of ram
  • 16 Hide
    mjam , August 29, 2008 11:09 AM
    No 4870X2 and 1920 X 1200 max resolution tested. How about finishing the good start of an article with the rest of it...
Other Comments
  • -4 Hide
    San Pedro , August 29, 2008 10:14 AM
    Looks like the results for SLI and Crossfire were switched with the single card results. . .
  • 14 Hide
    Duncan NZ , August 29, 2008 10:40 AM
    Not a bad article, really comprehensive.
    My one complaint? Why use that CPU when you know that the test cards are going to max it out? Why not a quad core OC'ed to 4GHz? It'd give far more meaning to the SLI results. We don't want results that we can duplicate at home, we want results that show what these cards can do. Its a GPU card comparason, not a complain about not having a powerful enough CPU story.

    Oh? And please get a native english speaker to give it the once over for spelling and grammar errors, although this one had far less then many articles posted lately.
  • 14 Hide
    elbert , August 29, 2008 10:50 AM
    No 4870x2 in CF so its the worlds top end Nvidia vs ATI mid to low end.
  • 15 Hide
    Lightnix , August 29, 2008 10:51 AM
    It'd be a good article if you'd used a powerful enough CPU and up to date Radeon drivers (considering we're now up to 8.8 now), I mean are those even the 'hotfix' 8.6's or just the vanilla drivers?
  • 20 Hide
    elbert , August 29, 2008 10:55 AM
    Version AMD Catalyst 8.6? Why not just say i'm using ATI drivers with little to no optimizations for the 4800's. This is why the CF benchmarks tanked.
  • 9 Hide
    Anonymous , August 29, 2008 10:57 AM
    at 1280, all of the highend cards were CPU limited. at that resolution, you need a 3.2-3.4 c2d to feed a 3870... this article had so much potential, and yet... so much work, so much testing, fast for nothing, because most of the results are very cpu limited (except 1920@AA).
  • 19 Hide
    wahdangun , August 29, 2008 11:07 AM
    WTF, hd4850 SHOULD be a lot faster than 9600 GT and 8800 GT even tough they have 1Gig of ram
  • 16 Hide
    mjam , August 29, 2008 11:09 AM
    No 4870X2 and 1920 X 1200 max resolution tested. How about finishing the good start of an article with the rest of it...
  • 15 Hide
    Anonymous , August 29, 2008 11:50 AM
    I agree, the 4870 X2 should have been in there and should have used the updated drivers. Good article but I think you fell short on finishing it.
  • -8 Hide
    Anonymous , August 29, 2008 11:59 AM
    @pulasky - Rage much? It's called driver issues you dumbass. Some games are more optimised for multicard setups than others, and even then some favour SLi to Crossfire. And if you actually READ the article rather than let your shrinken libido get the better of you, you'll find that Crossfire does indeed work in CoD4.

    Remember, the more you know.
  • 7 Hide
    buzzlightbeer , August 29, 2008 12:03 PM
    isnt forceware 177.41 out for gt200 series? so they are using a recent driver for the nvidia cards yet not for the ATI cards...plus yes would have to agree with wahdangun the 4850 is alot faster then the 9600gt and the 8800gt i have 2 friends with both cards with q6600s one at 3.2 (9600gt) and the other at 3.0 (4850) and the 4850 machine destroys the other one even with a lower clocked cpu
    but yes the article was off to a great start, maybe throw some vantage in there as well?
  • 15 Hide
    chesterman , August 29, 2008 12:06 PM
    agree with the others. u guys should use a more recent driver for ati/amd cards, use a more game-effective cpu and REALLY should have put the 4870x2 on the fight
  • 11 Hide
    masterwhitman , August 29, 2008 12:09 PM
    elbertVersion AMD Catalyst 8.6? Why not just say i'm using ATI drivers with little to no optimizations for the 4800's. This is why the CF benchmarks tanked.


    Precisely; several other websites tested with 8.7 and 8.8 long before this article was published. Why couldn't you? Look at the 8.6 release notes; it doesn't even mention the HD4000 series cards as supported devices.

    Brilliant guys.
  • 0 Hide
    roynaldi , August 29, 2008 12:27 PM
    NVISION comes around and IRONicallY, a 36 page article is produced that is magically in favor of, whats that, NVIDIA!!!

    After having the Mythbusters appear, you would think this would be the most comprehensive, "scientific," factual, and update article meeting Tom's usual standards.... I didn't finish reading this.
  • 10 Hide
    xrodney , August 29, 2008 12:47 PM
    Using old drivers with no optimalisation at all fo newest card whitch was released months ago seems too strange to me. Also temperature results for 48xx are quite oposite reality, at least when compare to 8.8 catalyst.
    (82 temperature in 2D 69 in 3D with no fanfix)
  • 2 Hide
    jitpublisher , August 29, 2008 1:00 PM
    Pretty good, finally. Wish you would have have used an overclocked Quad so the newer GPU's could show their full potentianl, and you really should have used the latest drivers, but I give this article 2 thumbs up. Lot of good information in here.
  • 3 Hide
    Haiku214 , August 29, 2008 1:13 PM
    Well the main reason why they don't have the 4870x2 and the latest drivers is simply because they made this article a couple of weeks ago. If you could just imagine how long and tedious it is to produce all these data and results. It's just sad that after finally finishing the article, a lot of new stuff has already happened(new drivers and the x2).
  • 6 Hide
    jameskangster , August 29, 2008 1:19 PM
    First I want to say that the article itself is not bad at all.
    Also, I can understand why TH didn't have time to use 8.8 since it was released publicly on August 20, 2008 (Although ATI would have gladly released a beta version to TH for testing purposes).

    However, AMD publicly released stable Catalyst 8.7(internal version 8.512) on July 21, 2008. That's more than a month ago. It has numerous improvements (for example, CF performance increase, improved stability and performance under Vista). To be honest, most of the improvements range from 4% to 15%. (In CF case, up to 1.7 X scaling)

    TH has rarely been unfair and/or inaccurate and they always owned up to their mistakes before, and I trust them to re-test ATI products with at least 8.7 if not 8.8 to continue to uphold their values and integrity.
  • 7 Hide
    outlw6669 , August 29, 2008 1:20 PM
    So, to start off with, this article is much better than many of the other recent reviews. I feel you put some thought into it and for the most part it is good. I found the comparative performance charts at the end interesting. Have you thought of changing the GPU charts in a similar fashion?

    Now on to my criticism.

    I can understand how you want to keep the results homogeneous with previous results but if you already know that a stock QX6800 will bottleneck the system, be proactive in fixing it. At the very least you should have done a small segment of the review showing the newer cards with a quad core overclocked to 4.0Ghz.

    Also, if you have ever read any of the older Toms articles, you would know that you can still minimise the bottleneck from a slow GPU bye raising the resolution. Perhaps you should test the fastest cards at the highest resolutions?

    I can also understand why you did not use the latest nVidia drivers. It takes time to create a review of this scale and the GF8/9 series drivers have been stable for some time. As the GT 200 series brings no new features to the table, they would needed little optimisation for their newer cards allowing the slightly dated drivers to perform nicely.

    What I can not understand is why you would use ATI's 8.6 drivers??
    The 8.7 drivers have been out for more than a month bringing quite a few fixes/optimisations with it. I understand it probably took more than 9 days to complete all of these benchmarks (today is the 29th, the 8.8 drivers were officially released on the 20th) but you should have called ATI and asked for their latest drivers. The 8.8 drivers were leaked at least a week before the official release which means, if you could nurture a relationship with the people you review, they could/probably would have provided them to you. There is still no excuse I can see for testing with the old 8.6 drivers. Seriously, it does not even have official support for the 48X0 cards...

    From the title of the article,"The Fastest 3D Cards Go Head-To-Head", I would have assumed that you would have been testing the Fastest 3D cards? What happened to your 4870x2? As you have already attempted to review it, we know you have your hands on one. How can you claim to review the "Fastest 3D Cards" and still leave out the fastest card?

    In summation, I liked many things from this article. The layout was nice and a little more technical than we have been seeing as of late. I enjoyed the comparison charts at the end and I think you should adopt a similar method for the CPU and GPU charts. I would have thought this was an excellent and well thought out article if it had not been for the glaring and obvious deficiencies in reason. I give you credit for stepping Toms in the right direction. With a little more unbiased comparison, critical thinking and common sense I could come to see reviews such as this in a very positive light.
Display more comments