Hercules 3D Prophet 9800 PRO 128MB 

I must say that the RADEON 9800 PRO has actually two coolers. The second one is located on the back of the card - this is a heatsink for cooling down the processor underneath. Both devices are reliably joined with bolts. And each pair of the memory chips have copper heatsinks which are attached with thermal grease. 

Now let's have a look at the accessory packs: 
 

Note that 

• Overclocking requires additional card cooling (for its memory, in particular): 




• Overclocking depends on a certain sample, and you shouldn't extend single-card results to the entire series or trade mark. Overclocking results are not obligatory characteristics of video cards. 

Testbed and drivers

Testbed: 

• Pentium 4 3066 MHz based computer:
• Intel Pentium 4 3066 MHz processor;
• ASUS P4G8X (iE7205) mainboard; 
• 1024 MB DDR SDRAM; 
• Seagate Barracuda IV 40GB hard drive; 
• Windows XP SP1;
• ViewSonic P810 (21") and ViewSonic P817 (21") monitors;
• ATI drivers v6.343 (CATALYST 3.4).

VSync off in drivers, texture compression off in applications. Texture detail set to High Quality. 

Test results

Before we start examining 2D quality, I should say there are no complete techniques for objective 2D quality estimation because: 

1. 2D quality much depends on certain samples for almost all modern 3D accelerators; 
2. Besides videocards, 2D quality depends on monitors and cables; 
3. Moreover, certain monitors might not work properly with certain video cards. 

As for the samples tested, together with the ViewSonic P817 monitor and BNC Bargo cable they showed excellent quality at the following resolutions and clock speeds:

The RADEON 7500 was purposely redesigned. This product has excellent graphics quality for a 2D product. It's a rear picture at $50. Add the capabilities of two integrated RAMDACs of 350 MHz, and the HydraVision technology with the multi-monitor and multi-desktop support. But we have written about all that in the previous reviews of cards based on this GPU. 

Scandal around 3DMark03: What happened?

The scandal flared up recently around 3DMark03 involves NVIDIA, the developer of FutureMark and ATI. 

It started when FutureMark published the results of the audit of the new NVIDIA Detonator FX driver in 3DMark03 on the NVIDIA GeForce FX 5900 Ultra card. 

Her is an extract: 

Futuremark's audit revealed cheats in NVIDIA Detonator FX 44.03 and 43.51 WHQL drivers.

Earlier GeForceFX drivers include only some of the cheats listed below:

1. The loading screen of the 3DMark03 test is detected by the driver. This is used by the driver to disregard the back buffer clear command that 3DMark03 gives...
2. A vertex shader used in game test 2 (P_Pointsprite.vsh) is detected by the driver. In this case the driver uses instructions contained in the driver to determine when to obey the back buffer clear command and when not to...
3. A vertex shader used in game test 4 (M_HDRsky.vsh) is detected. In this case the driver adds two static clipping planes to reduce the workload... This cheat was introduced in the 43.51 drivers as far as we know.
4. In game test 4, the water pixel shader (M_Water.psh) is detected. The driver uses this detection to artificially achieve a large performance boost - more than doubling the early frame rate on some systems... 
5. In game test 4 there is detection of a pixel shader (m_HDRSky.psh). Again it appears the shader is being totally discarded and replaced with an alternative more efficient shader in a similar fashion to the water pixel shader above...
6. A vertex shader (G_MetalCubeLit.vsh) is detected in game test 1. Preventing this detection proved to reduce the frame rate with these drivers, but we have not yet determined the cause.
7. A vertex shader in game test 3 (G_PaintBaked.vsh) is detected, and preventing this detection drops the scores with these drivers. This cheat causes the back buffer clearing to be disregarded...
8. The vertex and pixel shaders used in the 3DMark03 feature tests are also detected by the driver. When we prevented this detection, the performance dropped by more than a factor of two in the 2.0 pixel shader test.

ATI was also accused of some cheats but they were not so strong. 

All details of the optimizations and the methods of their prevention are given in the document above. We decided to check whether the quality drop of the NVIDIA GeForce FX 5900 and RADEON 9800 on 3DMark03 v3.20 is really so bad. At that same time we will clear up how much the speed falls down on the version 3.30, where these optimizations are traced. Finally, we will try to find out whether the 3DMark03 really lost its authority. Its objectivity does look doubtful because it's clear that the test can be manipulated, because it was originally adjusted for ATI's cards supporting pixel shaders 1.4 (RADEON 8500/9000/9100/9200/9500/9700/9800 and GeForce FX5200/5600/5800/5900 can fulfill many tests in two passes, while all the rest cards do it in 4 passes), because 3DMark03 was originally meant to play up to 10-12% of accelerators, and the other part of them was in not so promising conditions. 

But we keep on using this test as the 3DMark is very popular. And even if we find out that the benchmark is not fully objective, we will keep on using this packet but without analysis of the scores. 

It should be explained that it's not because FutureMark wanted a lot of money for the beta testing and NVIDIA left the rows of beta-testers damaging relations with the test developers. And it's not FutureMark's revenge for such behavior of the leading GPU maker (if it takes place at all!). The most important fact is that NVIDIA will keep on making optimizations and the test developers will have to patch it time and again. What results are trustworthy? Can you be sure that with a given patch and given drivers the NVIDIA cards' results will be objective? I can't. It also concerns ATI. Although the company admitted that the drivers had optimizations for the 3DMark03 and promised to remove them in the next version, no one guarantees that it won't happen again. 

The reputation of the 3DMark03 is spoiled by the scandal and by the fact that the test developers preferred to exhibit arrogance to normal collaboration with all chip makers. 

Let's get back on track. I have carefully examined operation of two cards: RADEON 9800 PRO and GeForce FX 5900 Ultra, as competitors, in 3DMark03 v3.20 and in v3.30 after the last scandalous patch. Instead of performance diagrams below are the tables demonstrating the percentage ratios. The suffix "OLD" marks the previous version of 3DMark03 - 3.20. 

Scandal around 3DMark03: Game 1

Compare the quality:

Scandal around 3DMark03: Game 2

Compare the quality:

The test developers affirm that the NV35 shows snow instead of the dark sky on the beta version of this test. Once again, if users have a perfect sky on their screens on the retail version of the test, then they do not care how some or other programmers make their optimizations. We can see no difference in quality between the versions 3.20 and 3.30 in case of the NV35 and RADEON 9800 PRO. But the NV35 has lost 10-13%, while the R350 has spun up. It looks strange, doesn't it? By the way, have a look at the artifacts when the RADEON 9800 PRO renders the space ship. 

Scandal around 3DMark03: Game 3

Compare the quality:

There is no difference in quality between the patches of both cards again. And again FutureMark finds something invisible for an ordinary user, which makes the NV35 speed down and the R350 speed up. By the way, the latter has some artifacts again. 

Scandal around 3DMark03: Game 4

Compare the quality:

Like in Game1, it's well seen that NVIDIA tries to raise the speed when rendering the water surface. Well, such tricks deserve punishment. Is the penalty of -50% fair in this test for NV35? I don't think so. Taking into account that water is displayed only 1/5 of the overall test time, , -50% is too much. The RADEON 9800 PRO doesn't suffer or gain from this patch. 

Scandal around 3DMark03: Vertex Shader

Compare the quality:

Scandal around 3DMark03: Pixel Shader 2.0

Compare the quality: