[nextpage title=”Introduction”]

AMD/ATI has released two new high-end graphics chips recently, Radeon HD 4870 and Radeon HD 4850, known by the codenames RV770 XT and RV770 PRO, respectively. As you can guess by the numbers, HD 4850 provides a lower performance compared to HD 4870. In this review we will benchmark HD 4850 from Sapphire and compare it to its main competitors from NVIDIA. Is this video card a good buy? Check it out.

Sapphire HD 4850 follows AMD/ATI’s reference model both on its physical aspect as in specifications: the graphics chip runs at 625 MHz with its memory being accessed at 993 MHz (real clock) through a 256-bit memory interface. Since GDDR3 memories transfer two data per clock cycle, the memory achieves a performance as if it were working at 1,986 MHz, delivering a maximum transfer rate of 63.5 GB/s.

Both new chips are based on DirectX 10.1 (Shader 4.1) architecture, where all processing units inside the graphics chip are generic, being able to process any kind of vector (pixel shader, vertex shader, etc). What is really impressive about these two new chips is that they both have 800 processing units inside the chip. We will provide a short table comparing the specs from this new video card with the specs from its main competitors. Before that, let’s take an overall look at Sapphire HD 4850.

Sapphire HD 4850Figure 1: Sapphire HD 4850.

Sapphire HD 4850Figure 2: Sapphire HD 4850.

Sapphire HD 4850Figure 3: Sapphire HD 4850.

[nextpage title=”Introduction (Cont’d)”]

We removed the video card cooler to take a look. As you can see in Figure 4, the cooler is 100% made of copper, which is great, as this material dissipates heat better than aluminum.

Sapphire HD 4850Figure 4: Video card cooler.

In Figure 5, you can see the video card without its cooler. It uses eight 512-Mbit Qimonda HYB18H512321BF-10 GDDR3 chips, making its 512 MB memory (512 Mbits x 8 = 512 MB). These chips have a maximum real clock rate of 1 GHz (1 ns, 2 GHz performance), so on this video card they are practically running on top of their maximum spec.

Sapphire HD 4850Figure 5: Sapphire HD 4850 with its cooler removed.

In Figure 6, you can see all accessories and CDs/DVDs that come with this video card, which includes a 2 GB USB drive (limited time offer). With the accessories that come with this card you can convert the video output to VGA, HDMI, component video and composite video, plus the DVI and S-Video connectors already present on the product.

Sapphire HD 4850Figure 6: Accessories.

This video card comes with three games inside a DVD called “Ruby ROM 1.1”: Call of Juarez, Dungeon Runners and Stranglehold. Programs that come with this video card include CyberLink DVD Suite (PowerProducer 4, PowerDirector 5 Express, Power2GO 5.5, Mediashow 3 and trial versions of PowerBackup 2.5, PowerDVD Copy and LabelPrint 2), Cyberlink Power DVD 7,3DMark06 Full, EarthSim and GameShadow.

Now let’s compare the Radeon HD 4850 specifications to its main competitors.

[nextpage title=”More Details”]

To make the comparison between Sapphire HD 4850 and the other video cards we included in this review easier, we compiled the table below comparing the main specs from these cards. If you want to compare the specs of Radeon HD 4850 to any other video card not included in the table below, just take a look at our NVIDIA Chips Comparison Table and on our AMD ATI Chips Comparison Table.

GPU Core Clock Shader Clock Processors Memory Clock Memory Interface Memory Transfer Rate Memory Price
GeForce GTX 280 602 MHz 1,296 MHz 240 1,107 MHz 512-bit 141.7 GB/s 1 GB GDDR3 USD 620 – 660
GeForce 9800 GX2 600 MHz 1,500 MHz 128 1,000 MHz 256-bit 64 GB/s 1 GB GDDR3 USD 420 – 550
GeForce 9800 GTX 675 MHz 1,688 MHz 128 1,100 MHz 256-bit 70.4 GB/s 512 MB GDDR3 USD 200 – 300
Rsdeon HD 4850 625 MHz 625 MHz 800 993 MHz 256-bit 63.5 GB/s 512 MB GDDR3 USD 195
Sapphire Atomic HD 3870 X2 857 MHz 857 MHz 320 927 MHz 256-bit 59.33 GB/s 1 GB GDDR3
Radeon HD 3870 776 MHz 776 MHz 320 1,125 MHz 256-bit 72 GB/s 512 MB GDDR4 USD 135 – 240

As you can see the main difference between the new Radeon HD 4850 and its competitors is the amazing number of processing units, 800 against 128 on GeForce 9800 GTX, for example.

In order to compete with Radeon HD 4850, NVIDIA reduced the maximum suggested price (MRSP) of GeForce 9800 GTX to USD 199, so Radeon HD 4850 and GeForce 9800 GTX are, today, on the same price range. At least officially – think about it, if you are a reseller and has your stock loaded with USD 300 video cards, would you lower them to USD 200 just because NVIDIA told you so? One week ago GeForce 9800 GTX was quoted between USD 270 and USD 355. NVIDIA also has launched a “new” high-end video card, GeForce 9800 GTX+, which is basically an overclocked GeForce 9800 GTX.

Some important observations regarding this table:

  • All NVIDIA chips are DirectX 10 (Shader 4.0), while all AMD/ATI chips are DirectX 10.1 (Shader 4.1).
  • The memory clocks listed are the real memory clock. Memory clocks are often advertised as double the figures presented, numbers known as “DDR clock.”
  • GeForce 9800 GX2 and Radeon HD 3870 X2 have two GPU’s. The numbers on the table represent only one of the chips.
  • All video cards included on our review were running at the chip manufacturer default clock configuration (i.e., no over
    clocking), except Sapphire Atomic HD 3870 X2. The official core clock for Radeon HD 3870 X2 is 825 MHz, while the official memory clock is 900 MHz. So this card was a little bit overclocked. We couldn’t reduce these clocks to their reference values and since we hadn’t any other Radeon HD 3870 X2 available we included this video card anyway.
  • Prices were researched at Newegg.com on the day we published this review.
  • We couldn’t find Sapphire Atomic HD 3870 X2 for sale. This model will be more expensive than cards from other vendors based on the same GPU because it features water cooling. Just for you to have an idea, prices on the regular Radeon 3870 X2 are quoted between USD 300 and USD 380.

Before going to our tests let’s recap the main features from Sapphire HD 4850. [nextpage title=”Main Specifications”]

Sapphire Radeon HD 4850 main features are:

  • Graphics chip: Radeon HD 4850 (codename RV770 PRO), running at 625 MHz.
  • Memory: 512 MB GDDR3 memory (1 ns, 256-bit interface) from Qimonda (HYB18H512321BF-10), running at 993 MHz (“1,983 MHz”).
  • Bus type: PCI Express x16 2.0.
  • Connectors: Two DVI (with HDMI support) and one S-Video output (with component video support).
  • Video Capture (VIVO): No.
  • Cables and adapters that come with this board: S-Video to component video cable, S-video to composite video adapter, DVI-to-VGA adapter, DVI-to-HDMI adapter, standard 4-pin peripheral power plug to 6-pin PCI Express auxiliary power plug (PEG) adapter and Crossfire bridge.
  • Number of CDs/DVDs that come with this board: Five.
  • Games that come with this board: Call of Juarez, Dungeon Runners and Stranglehold (inside “Ruby ROM 1.1” DVD).
  • Programs that come with this board: CyberLink DVD Suite (PowerProducer 4, PowerDirector 5 Express, Power2GO 5.5, Mediashow 3 and trial versions of PowerBackup 2.5, PowerDVD Copy and LabelPrint 2), Cyberlink Power DVD 7, 3DMark06 Full, EarthSim and GameShadow.
  • Extra features: free 2 GB flash drive (limited time offer).
  • More information: https://www.sapphiretech.com
  • Average price in the US*: USD 195.00

* Researched at Shopping.com on the day we published this review.

[nextpage title=”How We Tested”]

During our benchmarking sessions, we used the configuration listed below. Between our benchmarking sessions the only variable was the video card being tested.

Hardware Configuration

Software Configuration

  • Windows Vista Ultimate 32-bit
  • Service Pack 1

Driver Versions

  • nForce driver version: 15.17
  • AMD/ATI video driver version: Catalyst 8.5
  • AMD/ATI video driver version: Catalyst 8.6 + hotfix (8.501.1.0, 6/21/2008) (Radeon HD 4850)
  • NVIDIA video driver version: 175.16
  • NVIDIA video driver version: 177.34 (GeForce GTX 280)

Software Used

Resolutions and Image Quality Settings

Since we were comparing very high-end video cards, we ran all our tests under three 16:10 widescreen high resolutions: 1680×1050, 1920×1200, and 2560×1600. We always tried to run the programs and games in two scenarios for each resolution, one with low image quality settings and then maxing out the image quality settings. The exact configuration we used will be described together with the results of each individual test.

Error Margin

We adopted a 3% error margin; thus, differences below 3% cannot be considered relevant. In other words, products with a performance difference below 3% should be considered as having similar performance.

[nextpage title=”3DMark06 Professional”]

3DMark06 measures Shader 3.0 (i.e., DirectX 9.0c) performance. We run this software under three 16:10 widescreen resolutions, 1680×1050, 1920×1200, and 2560×1600, first with no image quality enhancements enabled – results we call “low” on the charts and tables below –, then setting 4x anti-aliasing and 16x anisotropic filtering. See the results below.

Sapphire HD 4850
3DMark06 Professional 1.1.0 – 1680×1050 – Low Score Difference
Sapphire Atomic Radeon HD 3870 X2 16260 37.31%
GeForce 9800 GX2 15623 31.93%
GeForce GTX 280 14904 25.86%
GeForce 9800 GTX 12759 7.74%
Sapphire Radeon HD 4850 11842  
Radeon HD 3870 10694 10.73%

Sapphire HD 4850

 

3DMark06 Professional 1.1.0 – 1920×1200 – Low Score Difference
GeForce 9800 GX2 15547 45.42%
Sapphire Atomic Radeon HD 3870 X2 15489 44.88%
GeForce GTX 280 14215 32.96%
GeForce 9800 GTX 11631 8.79%
Sapphire Radeon HD 4850 10691  
Radeon HD 3870 9454 13.08%
Sapphire HD 4850
3DMark06 Professional 1.1.0 – 2560×1600 – Low Score Difference
GeForce 9800 GX2 13015 61.14%
Sapphire Atomic Radeon HD 3870 X2 12315 52.47%
GeForce GTX 280 11766 45.67%
GeForce 9800 GTX 8743 8.25%
Sapphire Radeon HD 4850 8077  
Radeon HD 3870 6823 18.38%
Sapphire HD 4850
3DMark06 Professional 1.1.0 – 1680×1050 – High Score Difference
Sapphire Atomic Radeon HD 3870 X2 16260 83.09%
GeForce 9800 GX2 13900 56.51%
GeForce GTX 280 12157 36.89%
GeForce 9800 GTX 8981 1.13%
Sapphire Radeon HD 4850 8881  
Radeon HD 3870 6915 28.43%
Sapphire HD 4850
3DMark06 Professional 1.1.0 – 1920×1200 – High Score Difference
Sapphire Atomic Radeon HD 3870 X2 15489 94.29%
GeForce 9800 GX2 12213 53.20%
GeForce GTX 280 10991 37.87%
Sapphire Radeon HD 4850 7972  
GeForce 9800 GTX 7811 2.06%
Radeon HD 3870 6114 30.39%
Sapphire HD 4850
3DMark06 Professional 1.1.0 – 2560×1600 – High Score Difference
Sapphire Atomic Radeon HD 3870 X2 12315 108.87%
GeForce 9800 GX2 9829 66.71%
GeForce GTX 280 8704 47.63%
Sapphire Radeon HD 4850 5896  
GeForce 9800 GTX 5774 2.11%
Radeon HD 3870 4319 36.51%

[nextpage title=”3DMark Vantage Professional”]

3DMark Vantage is the latest addition to the 3DMark series, measuring Shader 4.0 (i.e., DirectX 10) performance and supporting PhysX, a programming interface developed by Ageia (now part of NVIDIA) to transfer physics calculations from the system CPU to the video card GPU in order to increase performance. Mechanical physics is the basis for calculations about the interaction of objects. For example, if you shoot, what exactly will happen to the object when the bullet hits it? Will it break? Will it move? Will the bullet bounce back? Notice that we didn’t upgrade the PhysX to the latest version, which would make the physics calculations for CPU Test 2 to be made by the GPU instead of the CPU on NVIDIA video cards (since we aren’t considering CPU or 3DMark scores this change wouldn’t produce any increase in our results anyway).

We ran this program at three 16:10 widescreen resolutions, 1680×1050, 1920×1200, and 2560×1600. First we used the “Performance” profile, and then we used the “Extreme” profile (basically enabling anti-aliasing at 4x, anisotropic filtering at 16x, and putting all detail settings at their maximum or “extreme” value. The combination of 2560×1600 resolution with extreme settings didn’t produce reliable results according to the program, so we aren’t going to add them here. The results being compared are the “GPU Scores” achieved by each video card.

Sapphire HD 4850
3DMark Vantage Professional 1.0.1 – 1680×1050 – Performance Score Difference
GeForce GTX 280 7695 60.41%
GeForce 9800 GX2 6990 45.72%
Sapphire Atomic Radeon HD 3870 X2 5651 17.80%
Sapphire Radeon HD 4850 4797  
GeForce 9800 GTX 3805 26.07%
Radeon HD 3870 2977 61.14%
Sapphire HD 4850
3DMark Vantage Professional 1.0.1 – 1920×1200 – Performance Score Difference
GeForce GTX 280 6106 63.92%
GeForce 9800 GX2 5379 44.40%
Sapphire Atomic Radeon HD 3870 X2 4336 16.40%
Sapphire Radeon HD 4850 37
25
 
GeForce 9800 GTX 2891 28.85%
Radeon HD 3870 2269 64.17%
Sapphire HD 4850
3DMark Vantage Professional 1.0.1 – 2560×1600 – Performance Score Difference
GeForce GTX 280 3549 73.12%
GeForce 9800 GX2 2910 41.95%
Sapphire Atomic Radeon HD 3870 X2 2382 16.20%
Sapphire Radeon HD 4850 2050  
GeForce 9800 GTX 1557 31.66%
Radeon HD 3870 1244 64.79%
Sapphire HD 4850
3DMark Vantage Professional 1.0.1 – 1680×1050 – Extreme Score Difference
GeForce GTX 280 6005 74.31%
GeForce 9800 GX2 4858 41.02%
Sapphire Atomic Radeon HD 3870 X2 3567 3.54%
Sapphire Radeon HD 4850 3445  
GeForce 9800 GTX 2703 27.45%
Radeon HD 3870 1855 85.71%
Sapphire HD 4850
3DMark Vantage Professional 1.0.1 – 1920×1200 – Extreme Score Difference
GeForce GTX 280 4732 71.89%
GeForce 9800 GX2 3508 27.42%
Sapphire Radeon HD 4850 2753  
Sapphire Atomic Radeon HD 3870 X2 2669 3.15%
GeForce 9800 GTX 2038 35.08%
Radeon HD 3870 1439 91.31%

[nextpage title=”Call of Duty 4″]

Call of Duty 4 is a DirectX 9 game implementing high-dynamic range (HDR) and its own physics engine, which is used to calculate how objects interact. For example, if you shoot, what exactly will happen to the object when the bullet hits it? Will it break? Will it move? Will the bullet bounce back? It gives a more realistic experience to the user.

We ran this program at three 16:10 widescreen resolutions, 1680×1050, 1920×1200, and 2560×1600, maxing out all image quality controls (i.e., everything was put on the maximum values on the Graphics and Texture menus). We used the game internal benchmarking feature, running a demo provided by NVIDIA called “wetwork.” We are putting this demo for downloading here if you want to run your own benchmarks. The game was updated to version 1.6.

Sapphire HD 4850
Call of Duty 4 – 1680×1050 – Maximum Score Difference
GeForce 9800 GX2 106.2 46.69%
GeForce GTX 280 105.3 45.44%
Sapphire Atomic Radeon HD 3870 X2 75.7 4.56%
Sapphire Radeon HD 4850 72.4  
GeForce 9800 GTX 69.1 4.78%
Radeon HD 3870 43.0 68.37%
Sapphire HD 4850
Call of Duty 4 – 1920×1200 – Maximum Score Difference
GeForce 9800 GX2 94.5 59.90%
GeForce GTX 280 91.7 55.16%
Sapphire Atomic Radeon HD 3870 X2 61.3 3.72%
Sapphire Radeon HD 4850 59.1  
GeForce 9800 GTX 57.7 2.43%
Radeon HD 3870 35.4 66.95%
Sapphire HD 4850
Call of Duty 4 – 2560×1600 – Maximum Score Difference
GeForce 9800 GX2 64.8 76.57%
GeForce GTX 280 64.8 76.57%
Sapphire Atomic Radeon HD 3870 X2 40.6 10.63%
GeForce 9800 GTX 38.3 4.36%
Sapphire Radeon HD 4850 36.7  
Radeon HD 3870 22.4 63.84%

[nextpage title=”Crysis”]

Crysis is a very heavy DirectX 10 game. We updated this game to version 1.2.1 and used the HOC Crysis Benchmarking Utility to help us collecting data. Since we don’t think the default demo based on the island map stresses the video card the way we want, we used the HOC core demo available with the abovementioned utility. We ran this demo under three 16:10 widescreen resolutions, 1680×1050, 1920×1200, and 2560×1600, first with image quality set to “low” and then with image quality set to “high.” Since all video cards achieved a number of frames per second below 10 at 2560×1600 with image details set to “high,” we are not including this test as the results aren’t reliable. We ran each test twice and discarded the first result, as usually the first run achieves a lower score compared to the subsequent runs since the game loses time loading files. The results are below, in frames per second (FPS).

Sapphire HD 4850
Crysis 1.2.1 – 1680×1050 – Low Score Difference
Sapphire Atomic Radeon HD 3870 X2 125 48.81%
GeForce GTX 280 125 48.81%
GeForce 9800 GTX 84 0.00%
Sapphire Radeon HD 4850 84  
GeForce 9800 GX2 75 12.00%
Radeon HD 3870 71 18.31%
Sapphire HD 4850
Crysis 1.2.1 – 1920×1200 – Low Score Difference
GeForce GTX 280 115 71.64%
Sapphire Atomic Radeon HD 3870 X2 108 61.19%
GeForce 9800 GTX 69 2.99%
Sapphire Radeon HD 4850 67  
GeForce 9800 GX2 63 6.35%
Radeon HD 3870 58 15.52%
Sapphire HD 4850
Crysis 1.2.1 – 2560×1600 – Low Score Difference
GeForce GTX 280 95 120.93%
Sapphire Atomic Radeon HD 3870 X2 71 65.12%
GeForce 9800 GTX 44 2.33%
Sapphire Radeon HD 4850 43  
GeForce 9800 GX2 42 2.38%
Radeon HD 3870 35 22.86%
Sapphire HD 4850
Crysis 1.2.1 – 1680×1050 – High Score Difference
GeForce GTX 280 42 44.83%
GeForce 9800 GTX 29 0.00%
Sapphire Radeon HD 4850 29  
Sapphire Atomic Radeon HD 3870 X2 26 11.54%
GeForce 9800 GX2 25 16.00%
Radeon HD 3870 19 52.63%
Sapphire HD 4850
Crysis 1.2.1 – 1920×1200 – High Score Difference
GeForce GTX 280 34 47.83%
Sapphire Radeon HD 4850 23  
GeForce 9800 GTX 22 4.55%
GeForce 9800 GX2 21 9.52%
Sapphire Atomic Radeon HD 3870 X2 20 15.00%
Radeon HD 3870 16 43.75%

[nextpage title=”Unreal Tournament 3″]

Unreal Tournament 3 is the latest installment from this famous first person shooter franchise, supporting DirectX 10 graphics when installed on Windows Vista with a DX10 compatible card. We upgraded Unreal Tournament 3 to version 1.2 and benchmarked it with the help of HOC UT3 benchmarking utility using the “Containment” demo, maxing out image quality settings (image quality at “high” and anisotropic filtering at x16). It is important to note that we haven’t applied the PhysX mod to this game, which would transfer PhysX processing from the CPU to the GPU on NVIDIA cards. The results are below, in frames per second (FPS).

Sapphire HD 4850
Unreal Tournament 3 – 1680×1050 – Maximum Score Difference
GeForce 9800 GTX 112 16.67%
GeForce 9800 GX2 108 12.50%
GeForce GTX 280 104 8.33%
Sapphire Radeon HD 4850 96  
Sapphire Atomic Radeon HD 3870 X2 84 14.29%
Radeon HD 3870 83 15.66%
Sapphire HD 4850
Unreal Tournament 3 – 1920×1200 – Maximum Score Difference
GeForce 9800 GTX 108 21.35%
GeForce 9800 GX2 106 19.10%
GeForce GTX 280 91 2.25%
Sapphire Radeon HD 4850 89  
Sapphire Atomic Radeon HD 3870 X2 78 14.10%
Radeon HD 3870 75 18.67%
Sapphire HD 4850
Unreal Tournament 3 – 2560×1600 – Maximum Score Difference
GeForce 9800 GTX 92 53.33%
GeForce 9800 GX2 92 53.33%
GeForce GTX 280 62 3.33%
Sapphire Radeon HD 4850 60  
Sapphire Atomic Radeon HD 3870 X2 51 17.65%
Radeon HD 3870 47 27.66%

[nextpage title=”Half-Life 2: Episode Two”]

Half-Life 2 is a popular franchise and we benchmark the video cards using Episode Two with the aid of HOC Half-Life 2 Episode Two benchmarking utility using the “HOC Demo 1” provided by this program. We ran the game in three 16:10 widescreen resolutions, 1680×1050, 1920×1200, and 2560×1600, under two scenarios. First with quality set to maximum, bilinear filtering and anti-aliasing set to x0. This configuration we are calling “low” on the charts and tables below. Then we maxed out image quality settings, enabling x16 anisotropic filtering and 16xQCS anti-aliasing. Th
is configuration we are calling “high” on our charts and tables. We updated the game up to the June 9th 2008 patch.

Sapphire HD 4850
Half-Life 2: Episode Two -1680×1050 – Low Score Difference
Sapphire Radeon HD 4850 164.9  
Sapphire Atomic Radeon HD 3870 X2 160.4 2.81%
GeForce GTX 280 156.3 5.50%
GeForce 9800 GTX 153.8 7.22%
Radeon HD 3870 145.7 13.18%
GeForce 9800 GX2 136.8 20.54%
Sapphire HD 4850
Half-Life 2: Episode Two – 1920×1200 – Low Score Difference
Sapphire Atomic Radeon HD 3870 X2 156.7 4.61%
GeForce GTX 280 156.3 4.34%
Sapphire Radeon HD 4850 149.8  
GeForce 9800 GTX 146.9 1.97%
GeForce 9800 GX2 135.2 10.80%
Radeon HD 3870 120.1 24.73%
Sapphire HD 4850
Half-Life 2: Episode Two – 2560×1600 – Low Score Difference
GeForce GTX 280 145.1 54.53%
GeForce 9800 GX2 130.6 39.08%
Sapphire Atomic Radeon HD 3870 X2 129.7 38.13%
GeForce 9800 GTX 107.9 14.91%
Sapphire Radeon HD 4850 93.9  
Radeon HD 3870 72.8 28.98%
Sapphire HD 4850
Half-Life 2: Episode Two – 1680×1050 – High Score Difference
GeForce 9800 GTX 137.9 18.67%
Sapphire Atomic Radeon HD 3870 X2 126.1 8.52%
GeForce 9800 GX2 125.4 7.92%
Sapphire Radeon HD 4850 116.2  
GeForce GTX 280 89.3 30.12%
Radeon HD 3870 68.3 70.13%
Sapphire HD 4850
Half-Life 2: Episode Two – 1920×1200 – High Score Difference
GeForce 9800 GTX 116.3 19.65%
GeForce 9800 GX2 111.1 14.30%
Sapphire Atomic Radeon HD 3870 X2 106.5 9.57%
Sapphire Radeon HD 4850 97.2  
GeForce GTX 280 70.3 38.26%
Radeon HD 3870 56.8 71.13%
Sapphire HD 4850
Half-Life 2: Episode Two – 2560×1600 – High Score Difference
GeForce 9800 GTX 71.3 22.09%
Sapphire Radeon HD 4850 58.4  
Sapphire Atomic Radeon HD 3870 X2 50.6 15.42%
GeForce 9800 GX2 37.5 55.73%
GeForce GTX 280 35.5 64.51%
Radeon HD 3870 34.9 67.34%

[nextpage title=”Quake 4″]

We upgraded Quake 4 to version 1.4.2 and ran its multiplayer demo id_perftest with SMP option enabled (which allows Quake 4 to recognize and use more than one CPU), under the same three 16:10 widescreen resolutions, 1680×1050, 1920×1200, and 2560×1600, first with image quality settings configured at “low” and then with image quality settings configured at “ultra.” You can check the results below, given in frames per second.

Sapphire HD 4850

Quake 4 – 1680×1050 – Low Score Difference
Sapphire Atomic Radeon HD 3870 X2 285.30 18.20%
GeForce GTX 280 268.80 11.36%
Sapphire Radeon HD 4850 241.38  
Radeon HD 3870 227.75 5.98%
GeForce 9800 GTX 225.52 7.03%
GeForce 9800 GX2 220.48 9.48%
Sapphire HD 4850
Quake 4 – 1920×1200 – Low Score Difference
Sapphire Atomic Radeon HD 3870 X2 266.23 28.25%
GeForce GTX 280 235.92 13.65%
Sapphire Radeon HD 4850 207.58  
Radeon HD 3870 188.40 10.18%
GeForce 9800 GX2 174.06 19.26%
GeForce 9800 GTX 158.87 30.66%
Sapphire HD 4850
Quake 4 – 2560×1600 – Low Score Difference
Sapphire Atomic Radeon HD 3870 X2 197.82 54.55%
GeForce GTX 280 168.81 31.88%
Sapphire Radeon HD 4850 128.00  
Radeon HD 3870 116.01 10.34%
GeForce 9800 GTX 114.34 11.95%
GeForce 9800 GX2 100.07 27.91%
Sapphire HD 4850
Quake 4 – 1680×1050 – Ultra Score Difference
GeForce GTX 280 246.39 1.85%
Sapphire Radeon HD 4850 241.91  
Sapphire Atomic Radeon HD 3870 X2 237.98 1.65%
GeForce 9800 GX2 218.80 10.56%
GeForce 9800 GTX 194.65 24.28%
Radeon HD 3870 167.26 44.63%
Sapphire HD 4850
Quake 4 – 1920×1200 – Ultra Score Difference
GeForce GTX 280 224.44 8.13%
Sapphire Atomic Radeon HD 3870 X2 218.62 5.32%
Sapphire Radeon HD 4850 207.57  
GeForce 9800 GX2 158.35 31.08%
GeForce 9800 GTX 158.18 31.22%
Radeon HD 3870 144.80 43.35%
Sapphire HD 4850
Quake 4 – 2560×1600 – Ultra Score Difference
Sapphire Atomic Radeon HD 3870 X2 177.36 38.69%
GeForce GTX 280 168.43 31.71%
Sapphire Radeon HD 4850 127.88  
GeForce 9800 GTX 102.04 25.32%
GeForce 9800 GX2 94.68 35.07%
Radeon HD 3870 94.40 35.47%

[nextpage title=”Conclusions”]

With so much data presented in the previous pages, we think we should summarize our review answering two questions: how is Sapphire HD 4850 compared to Radeon HD 3870 – once the most high-end video card from AMD/ATI – and how is Sapphire HD 4850 compared to GeForce 9800 GTX, especially now that NVIDIA is promoting a price cut on this model in order to make it a competitor to HD 4850.

AMD raised the performance of the new Radeon HD 4850 to a completely new level. Sapphire HD 4850 was between 6% and 91% faster than Radeon HD 3870 during our tests, depending on the game, resolution and image quality settings. The highest differences were found when we increased image quality. You can see the complete breakdown in the previous page.

We saw a good battle between Sapphire HD 4850 and GeForce 9800 GTX. Here is the breakdown.

We saw Sapphire HD 4850 being faster than GeForce 9800 GTX on two programs: 3DMark Vantage (between 26% and 35% faster) and Quake 4 (between 7% and 31% faster).

We saw Sapphire HD 4850 and GeForce 9800 GTX achieving the same performance level on Call of Duty 4, Crysis and 3DMark06 with image quality settings enabled (with image quality settings disabled, GeForce 9800 GTX was between 8% and 11% faster).

And we saw GeForce 9800 GTX being faster than Radeon HD 4850 on Unreal Tournament 3 (between 17% and 53%) and on Half-Life 2: Episode Two (between 15% and 22%, however both cards achieved the same performance at 1920×1200 with no image quality settings enabled and Sapphire HD 4850 was 7% faster than GeForce 9800 GTX at 1680×1050 with no image quality settings enabled).

So we have to call it a technical tie, because which one is faster will depend on the game. On the two of the heaviest games around today – Crysis and Call of Duty 4 – both cards achieved similar performance.

Then comes pricing. On this aspect Sapphire HD 4850 has, today, the biggest advantage, even with NVIDIA pushing their partners to drop GeForce 9800 GTX’s price down to USD 200. This model from Sapphire can be found around USD 195 and you can buy it on NewEgg.com for USD 175 after a mail-in rebate. Even though the cheapest GeForce 9800 GTX was being sold by USD 200 on the day we published this review, you could find GTX models being sold as high as USD 300, with most models being sold by USD 250. As we mentioned before we doubt resellers will drop their prices just because NVIDIA wants them to.

So unless you only play Unreal Tournament 3 or Half Life 2 we honestly believe that Sapphire HD 4850 is the video card that brings the best cost/benefit ratio for the savvy user looking for a high-end video card: it costs at least half the price of GeForce 9800 GX2, GeForce GTX 260 and GeForce GTX 280 and still brings a high-end gaming performance to your PC.

A final trick: at this time Radeon HD 4850 will only run stable if you install Catalyst 8.6 drivers and a hotfix made available by AMD.