ATI RADEON X800SE 256MB AGP

Video card's features
Testbed configurations, benchmarks, 2D quality
Test results: Serious Sam: The Second Encounter
Test results: Code Creatures
Test results: Unreal Tournament 2003
Test results: Unreal II: The Awakening
Test results: RightMark 3D
Test results: TRAOD
Test results: FarCry
Test results: Call Of Duty
Test results: HALO: Combat Evolved
Test results: Half-Life2(beta)
Test results: Splinter Cell
Test results: Quake3 Arena
Test results: Return to Castle Wolfenstein
Test results: DOOM III
Test results: 3DMark05 Game1
Test results: 3DMark05 Game2
Test results: 3DMark05 Game3
Test results: 3DMark05 MARKS
Conclusions

We have been looking forward to the third element in the R420 (X800) series for a long time. The fact is that from the very beginning NVIDIA announced three video cards based on NV40 (GeForce 6800): 6800 Ultra, 6800GT, and just 6800 to occupy the main price niches in High-End sector (499, 399 and 299 USD respectively). But canadian ATI contented itself with two items: RADEON X800 XT PE and X800 PRO with the recommended prices at 499 and 399 USD.

Having appeared later, RADEON X800 XT did not push down X800 PRO from 399 to 299 USD. It nicely settled down in the sector belonging to X800 XT PE, because the latter video card had been rarely seen on the shelves by common consumers.

Taking into account GeForce 6800LE appearing on sale after GeForce 6800, the american manufacturer has covered a larger market with its new products than the canadian one. ATI would not agree to launch a cheaper modification of X800 PRO, in order to fight back the increasingly popular GeForce 6800.

That's why it was crystal clear that X800SE would appear sooner or later, but what would it look like? Interestingly, there appeared no video card named RADEON X800 without any suffixes, and all ATI and NVIDIA suffixes do not match in any card.

So, what do we have now (late October, 2004)?

Price sector, USD	ATI	NVIDIA
600-700	RADEON X800 XT PE	GeForce 6800 Ultra
500-550	RADEON X800 XT	Several GeForce 6800GT modifications
450-500	RADEON X800 PRO	GeForce 6800GT
400-450	-	GeForce 6800 256MB GDDR3
350-400	RADEON X800SE (initial inflated prices)	GeForce 6800 GT 128MB DDR
300-350	RADEON X800SE (final prices)	GeForce 6800 128MB DDR
250-300	-	GeForce 6800LE

As we can see, video cards from NVIDIA have more distinctive and better-marked positions. Besides, the market also has video cards from ASUS (hybrids between GeForce 6800 and 6800GT), because the californian manufacturer does not interfere with engineers working for its partners. It means that there are more interesting modifications.

Somewhat flat attitude of the canadian manufacturer prohibiting to manufacture X800 independently and actually selling card assemblies instead of chips seems to me too oppressive for the development of this sector. If the manufacturing companies can produce their own PCBs for GeForce 6800/6800GT, which are by far more complex than PCBs for X800, why this stubbornness in not letting your partners manufacture what cards they like if they don't cheat with wrong card names.

There would have been no blanks in the table above. Bus alas, more mandative ATI policies lead to the stagnation in the X800 popularity and even to some signs of decline, to say nothing of keeping it above its competitors. So, what can X800SE offer in this regard? So far we can note that it does not offer much. Firstly, this video card is based on the same PCB and it's equipped with the same GDDR3 memory as X800 PRO. Thus its net cost is lower than that for X800 PRO only by the cost difference between 12-pipline and 8-pipeline R420. Taking into account that the latter is a production residue from the former, this difference can be substantial, but still it will hardly be appropriate to the price drop for RADEON X800SE video cards on the whole. Such video cards must cost 100-150 USD less. Rate of profit of such cards will certainly be lower than that of X800 PRO.

And the second aspect: X800SE mission. Many readers remember that GeForce 6800LE cards appeared on the market quite unexpectedly and were intended for integrators. Avalanche-like expansion of retail sales of this product forced NVIDIA to change its plans and to allow its partners to sell these cards officially, even in retail. By the way, according to the latest information, NVIDIA stopped selling GeForce 6800 cards to partners to increase the sales volume, and now manufacturers produce video cards based only on GeForce 6800LE (and senior 6800GT/Ultra). Statistics of unlocking pipelines in these cards show that the lion's share yields to this operation and thus some part of the pipelines is deliberately and systematically locked. We shall not inquire the reasons – that's none of our business. I'm just establishing facts.

So, X800SE cards are certainly not intended for us the potential retail customers but for integrators. And these cards are manufactured on the same plants as all X800PRO/XT cards. That's why ATI knows very well every card, especially as all of them are marked with numbers of the canadian company. No matter which partner sells these cards, ATI will still be in the know how many cards and to whom they have sold. Their control seems efficient. But nevertheless, these cards appeared on retail. Not as a jettison from some integrator as it had happened with 6800LE, but from partners. Sapphire turned out the first such "disobedient slave". The card under review is just from this vendor. To be more exact, it WAS PUT ON SALE as if by this company. But in reality you'll see that there is even no sticker from this "jewel company", only the bundled CD has the Sapphire logo on it.

But still it's true, this company decided to sell such cards by retail. X800SE retail sales from Sapphire have been already placed on record in Japan.

So far ATI is just smolderingly discontent and threatens with divine retribution. We cannot possibly know what will happen next. I remember that all attempts of the canadian company to suppress product jettisons for OEM had been a failure (remember RADEON LE). It seems to me that X800SE cards will enter the retail market anyway and that ATI will have to swallow that in the long run. Yes, I understand that the cards are delivered to integrators at completely different prices than for retail sales, that's why the canadian company does not like the idea of losing its profits for the cards put on the retail market.

So, this mysterious product should not have appeared in our lab at all. It should have quietly gone to PC integrators, who would have surely called it X800. We have already seen that with GeForce4 MX, when almost no integrator added the MX suffix in the specifications, just GeForce4. And many buyers of such computers were dead sure about the GeForce4 Ti cards inside supporting all shaders and demonstrating performance stated in reviews they read (surely not about GeForce4 MX).

We have come close to what X800SE really is. I'll tell you right away that it's based on the common R420, only cut down in the number of pipelines, so all its functions have been retained. And we have written plenty of articles about them.

Theoretical materials and reviews of video cards, which concern functional properties of the GPU ATI RADEON X800 (R420)/X700 (RV410) and NVIDIA GeForce 6800 (NV40)/6600 (NV43)

What do we know about this card before our tests? - Frequencies and a number of pixel pipelines. These are 425/400 (800) MHz and 8 pipelines, each with 1 texture unit. That is twice as little as in X800XT. The number of vertex pipelines should also be reduced twofold (3 pipelines instead of 6), but this fact remains to be checked out.

Let's compare it with RADEON 9800XT. It has a similar design: 8 pixel pipelines as well, but 4 vertex pipelines. Their frequencies are also similar: 412/365 (730) MHz. It seems that X800SE ought to be close in performance, as both cards have 256-bit memory bus and 256 MB on-board memory. However the tests demonstrated that the difference exists. It'll be described below in this review.

And another important issue for overclockers: 9800XT is based on the 0.15-micron manufacturing process, while X800SE (aka R420) – on 0.13-micron using LOW-K. Thus the overclocking potential of the latter video card is much higher. Besides, it's equipped with 2.0ns GDDR3 memory operating at 500 (1000) MHz, instead of 2.5ns DDR as in 9800XT. Getting ahead of our review, this video card easily operated at X800 XT PE frequencies: 520/560 (1120) MHz, which made it possible to compare its performance with the original X800XT PE, which has twice as many pipelines.

So, a brief description of the video card itself.

Video card

ATI RADEON X800SE 256MB AGP

ATI RADEON X800SE 256MB AGP
The card has the same AGP 8x/4x interface, 256 MB GDDR3 SDRAM allocated in 8 chips on the front and the back sides of the PCB. Samsung memory chips. 2.0ns memory access time, which corresponds to 500 (1000) MHz. Memory operates at 400 (800) MHz. GPU frequency — 425 MHz. 256bit memory bus. 8 pixel pipelines; and according to our tests, 6 vertex pipelines (exactly as in X800PRO/XT).

Comparison with the reference design, front view
ATI RADEON X800SE 256MB AGP	Reference card ATI RADEON X800 XT

Comparison with the reference design, back view
ATI RADEON X800SE 256MB AGP	Reference card ATI RADEON X800 XT

As I have already mentioned above, this video card is obviously a reference card and it has no differences from X800XT/PRO. It just has a different cooler.

ATI RADEON X800SE 256MB AGP

The cooling system is fragrantly primitive :). A closed heatsink with a fan in the middle. However the fan speed depends on the temperature, so the cooler may be noisy. The new version of RivaTuner will be able to control the fan speed and so you will be able to set its rotational speed to a compromise level (between the noise and cooling efficiency).

And here is the graphics processor.

You can see that the chip is marked as X800SE, that is the die was tested before it had been marked. Or, as I have already assumed above, it's marked like this on purpose and there is a laser-burned track between the contacts on the right side (you can see a dot). Pay attention that all the tracks in X800 XT (16 pipelines) are not damaged:

Winding up the description of the card, I want to note the OEM package. Thus there is no point in describing the bundle (it only contains a CD with drivers and an external power cable) and the box.

This video card is very good at overclocking, so we easily managed to test the card at X800 XT PE frequencies: 520/560 (1120) MHz.

Installation and Drivers

Testbed configurations:

Athlon 64 3200+ based computer
- AMD Athlon 64 3200+ (L2=1024K) CPU
- ASUS K8V SE Deluxe mainboard based on VIA K8T800
- 1 GB DDR SDRAM PC3200
- Seagate Barracuda 7200.7 80GB SATA HDD
Operating system – Windows XP SP2; DirectX 9.0c
Monitors: ViewSonic P810 (21") and Mitsubishi Diamond Pro 2070sb (21").
ATI drivers v6.490; NVIDIA drivers v66.81.

VSync is disabled.

It should be noted that both companies have enabled trilinear filtering optimizations in their drivers by default.

The following cards were used as contenders in this review:

NVIDIA GeForce 6800LE (Palit, 325/350 (700) MHz, 128MB);
NVIDIA GeForce 6800 (Palit, 325/350 (700) MHz, 128MB);
NVIDIA GeForce 6800 (ASUS V9999GE, 350/500 (1000) MHz, 256MB GDDR3);
NVIDIA GeForce 6800GT (ASUS V9999GT, 350/350 (700) MHz, 128MB).

Test results

Before giving a brief evaluation of 2D, I will repeat that at present there is NO valid method for objective evaluation of this parameter due to the following reasons:

2D quality in most modern 3D accelerators dramatically depends on a specific sample, and it's impossible to evaluate all the cards.
2D quality depends not only on the video card, but also on the monitor and a cable.
A great impact on this parameter has been recently demonstrated by monitor-card combos, that is there are monitors, which just won't "work" with specific video cards.

What concerns the combo of our sample under review and Mitsubishi Diamond Pro 2070sb, this card demonstrated the excellent quality in the following resolutions and frequencies:

ATI RADEON X800SE 256MB AGP

1600x1200x85Hz, 1280x1024x120Hz, 1024x768x120Hz

Test results: performance comparison

We used the following test applications:

Serious Sam: The Second Encounter v.1.05 (Croteam/GodGames) – OpenGL, multitexturing, ixbt0703-demo, test settings: quality, S3TC OFF
Quake3 Arena v.1.17 (id Software/Activision) – OpenGL, multitexturing, ixbt0703-demo, all test settings to maximum: detailing level – High, texture details – #4, S3TC OFF, curved surfaces are strongly smoothed using variables r_subdivisions «1» and r_lodCurveError «30000» (note that by default r_lodCurveError «250» !), the configurations can be downloaded here
Code Creatures Benchmark Pro (CodeCult) – gaming test demonstrating how the video card works with DirectX 8.1, Shaders, HW T&L.
Unreal Tournament 2003 v.2225 (Digital Extreme/Epic Games) – Direct3D, Vertex Shaders, Hardware T&L, Dot3, cube texturing, default quality
Unreal II: The Awakening (Legend Ent./Epic Games) – Direct3D, Vertex Shaders, Hardware T&L, Dot3, cube texturing, default quality
RightMark 3D v.0.4 (one of game scenes) – DirectX 8.1, Dot3, cube texturing, shadow buffers, vertex and pixel shaders (1.1, 1.4).
Tomb Raider: Angel of Darkness v.49 (Core Design/Eldos Software) – DirectX 9.0, Paris5_4 demo. The tests were conducted with the quality set to maximum, only Depth of Fields PS20 was disabled.
HALO: Combat Evolved (Microsoft) – Direct3D, Vertex/Pixel Shaders 1.1/2.0, Hardware T&L, maximum quality
Half-Life2 (Valve/Sierra) – DirectX 9.0, demo (ixbt07. The tests were carried out with enabled anisotropic filtering as well as in heavy mode with AA and anisotropy.
Tom Clancy's Splinter Cell v.1.2b (UbiSoft) – Direct3D, Vertex/Pixel Shaders 1.1/2.0, Hardware T&L, Very High settings; demo 1_1_2_Tbilisi
Call of Duty (MultiPlayer) (Infinity Ward/Activision) – OpenGL, multitexturing, ixbt0104demo, test settings – maximum, S3TC ON
FarCry 1.2 (Crytek/UbiSoft), DirectX 9.0, multitexturing, demo01 (research) (the game was started with the -DEVMODE option), test settings – Very High.
Return to Castle Wolfenstein (MultiPlayer) (id Software/Activision) – OpenGL, multitexturing, ixbt0703-demo, test settings – all to maximum, S3TC OFF,
DOOM III (id Software/Activision) – OpenGL, multitexturing, test settings – High Quality (ANIS8x),
3DMark05 (FutureMark) – DirectX 9.0, multitexturing, test settings – trilinear,

If you want to get the demo-benchmarks, which we use, contact me at my e-mail.

Before we proceed to our performance analysis, I want to tell you that we tested the cards in 3DMark05 Vertex Shaders and in RightMark3D to check the number of vertex pipelines in X800SE. And we found out that at 520/1120 MHz the X800SE card demonstrated the same vertex unit performance as in X800XT PE. So we proved in action that RADEON X800SE has the whole set of 6 vertex pipelines.

Serious Sam: The Second Encounter

RADEON X800SE 520/1120 MHz versus RADEON X800XT PE – certainly, the difference does not reach the appropriate 50%, but we should take into account that X800SE is not cut down in pipelines.
RADEON X800SE versus RADEON 9800XT – there is a small performance gain, especially with AF. But on the whole it's insignificantly higher.
RADEON X800SE versus GeForce FX5950 Ultra – almost parity with a small lag from 5950U
RADEON X800SE versus GeForce 6800 – heavy defeat, almost twofold! 12 pipelines in 6800 play a more important role than operating frequencies in X800SE.
RADEON X800SE versus GeForce 6800LE – though the defeat is not that heavy, but still, it's a serious blow to the ATI product
RADEON X800SE versus GeForce 6800GT 128MB (ASUS V9999GT) – it's quite clear that if the card didn't hold against 6800, then it's much worse here.

I remind you that while the X800SE prices are on the level of ASUS V9999GT (380 USD), this card will be almost unclaimed.

Code Creatures

RADEON X800SE 520/1120 MHz versus RADEON X800XT PE – note that the difference again does not reach 50%, it does not depend only on pixel pipelines.
RADEON X800SE versus RADEON 9800XT – excellent advantage! And AF provides fantastic gains to X800SE relative to 9800XT! What is it? - It's obviously optimizations inside R420 plus drivers. Pay attention that this test is oriented to Shaders 1.1.
RADEON X800SE versus GeForce FX5950 Ultra – victory as well, which is not surprising considering how slow the entire NV3x series is with shaders.
RADEON X800SE versus GeForce 6800 – overall victory, which is interesting. Due to the AA mode in high resolutions. Though in 1024x768 the competitor turned out a tad stronger.
RADEON X800SE versus GeForce 6800LE – unconditional victory. However it would have been strange, if the product with the same number of pipelines and higher frequencies were outscored.
RADEON X800SE versus GeForce 6800GT 128MB (ASUS V9999GT) – defeat. Strong emphasis on shader power in this test played its role, and despite the heavy memory limitation in ASUS V9999GT, the core performance prevailed.

Unreal Tournament 2003

RADEON X800SE 520/1120 MHz versus RADEON X800XT PE – in this test X800XT PE turns on CPU capacity even with AA + AF, that's why the difference between the contenders is not that big.
RADEON X800SE versus RADEON 9800XT – excellent result, we can see again how heavily 9800XT is outscored with almost equal frequencies (only the number of vertex pipelines in the latter is a tad less than in X800SE).
RADEON X800SE versus GeForce FX5950 Ultra – the same picture, victory!
RADEON X800SE versus GeForce 6800 – Victory again! It's amazing, but the reason is probably in drivers: the most profitable is the AA+AF mode, where the advantage of X800SE reached almost 50%. But with separate AA and AF the performance gains are not that large.
RADEON X800SE versus GeForce 6800LE – if the previous test brought victory, this one is so much the more! Even double!
RADEON X800SE versus GeForce 6800GT 128MB (ASUS V9999GT) – overall victory (with AA+AF) despite some percentage fluctuations and various results.

Unreal II: The Awakening

RADEON X800SE 520/1120 MHz versus RADEON X800XT PE – identical to the previous test
RADEON X800SE versus RADEON 9800XT – the same picture
RADEON X800SE versus GeForce FX5950 Ultra – the same picture
RADEON X800SE versus GeForce 6800 – it's not particularly different from UT2003.
RADEON X800SE versus GeForce 6800LE – the same picture.
RADEON X800SE versus GeForce 6800GT 128MB (ASUS V9999GT) – this 6800GT modification won this battle.

RightMark 3D

RADEON X800SE 520/1120 MHz versus RADEON X800XT PE – again the test turns on the CPU capacity in X800XT, and thus the difference between this card and X800SE is not big
RADEON X800SE versus RADEON 9800XT – +20% advantage, which is not bad at all (optimizations!)
RADEON X800SE versus GeForce FX5950 Ultra – the same picture, but the main role here is played by the shader speed in the NVIDIA product.
RADEON X800SE versus GeForce 6800 – small advantage of X800SE
RADEON X800SE versus GeForce 6800LE – impressive victory.
RADEON X800SE versus GeForce 6800GT 128MB (ASUS V9999GT) – defeat, because 6800GT is obviously outclassing its competitor in this shader test, where the performance mostly depends on the core capacity.

TR:AoD, Paris5_4 DEMO

RADEON X800SE 520/1120 MHz versus RADEON X800XT PE – the difference between the competitors grows in the test where much depends on the core capacity in pixel operations. And though this difference does not reach 50%, it's still close to this figure.
RADEON X800SE versus RADEON 9800XT – almost parity, and thus the competitors are almost on par in pure shader operations, the difference tells only due to frequency differences and the number of vertex pipelines.
RADEON X800SE versus GeForce FX5950 Ultra – considering how weak the latter video card is in such operations, the result was quite expected.
RADEON X800SE versus GeForce 6800 – there it goes! As soon as the pixel capacity of the core as well as the fillrate snap into action, X800SE immediately turns out defeated with its 8 pipelines against 12 in its competitor despite the difference in frequencies.
RADEON X800SE versus GeForce 6800LE – But with the equal number of pixel pipelines and higher frequency in X800SE, the latter is victorious!
RADEON X800SE versus GeForce 6800GT 128MB (ASUS V9999GT) – quite an expected result considering all the above! X800SE is heavily defeated!

FarCry, demo01

RADEON X800SE 520/1120 MHz versus RADEON X800XT PE – identical to the previous test.
RADEON X800SE versus RADEON 9800XT – impressive victory! Those optimizations!
RADEON X800SE versus GeForce FX5950 Ultra – the same picture.
RADEON X800SE versus GeForce 6800 – overall defeat.
RADEON X800SE versus GeForce 6800LE – Overall parity. But if not take into account AA, only AF, then X800SE demonstrates excellent results!
RADEON X800SE versus GeForce 6800GT 128MB (ASUS V9999GT) – defeat.

Call of Duty, ixbt04

RADEON X800SE 520/1120 MHz versus RADEON X800XT PE – the difference is not so big, because X800XT turns on the CPU capacity.
RADEON X800SE versus RADEON 9800XT – quite good advantage over the competitor again.
RADEON X800SE versus GeForce FX5950 Ultra – the same picture.
RADEON X800SE versus GeForce 6800 – heavy defeat!
RADEON X800SE versus GeForce 6800LE – the same picture
RADEON X800SE versus GeForce 6800GT 128MB (ASUS V9999GT) – all the more defeated!

HALO: Combat Evolved

RADEON X800SE 520/1120 MHz versus RADEON X800XT PE – the game considerably rests on the pixel potential of the chips, that's why the difference between these competitors has almost reached 50%, which ideally must be the case.
RADEON X800SE versus RADEON 9800XT – a little advantage.
RADEON X800SE versus GeForce FX5950 Ultra – considering shader capacities of 5950U, it's not surprising that X800SE is victorious
RADEON X800SE versus GeForce 6800 – defeat (X800SE has a lower fillrate, thus the consequences).
RADEON X800SE versus GeForce 6800LE – almost parity, though X800SE must have had more advantage.
RADEON X800SE versus GeForce 6800GT 128MB (ASUS V9999GT) – heavy defeat.

Half-Life2 (beta): ixbt07 demo

RADEON X800SE 520/1120 MHz versus RADEON X800XT PE – the difference is at 26%, which is not so strong considering requirements of this game to pixel calculations. Though there is enough geometry and lighting and the competitors are equal in vertex pipelines.
RADEON X800SE versus RADEON 9800XT – super advantage! It seems strange to me! Why these results? Only because of 6 vertex pipelines against 4 in 9800XT plus optimizations? To raise the performance gain almost by 50% is worth all the efforts!
RADEON X800SE versus GeForce FX5950 Ultra – no need in comments, 5950U is written off and crawls away to a museum.
RADEON X800SE versus GeForce 6800 – overall little defeat without AA, but very close to parity.
RADEON X800SE versus GeForce 6800LE – serious advantage of X800SE!
RADEON X800SE versus GeForce 6800GT 128MB (ASUS V9999GT) – defeat!

Splinter Cell

RADEON X800SE 520/1120 MHz versus RADEON X800XT PE – strangely little difference: the game turns on CPU.
RADEON X800SE versus RADEON 9800XT – victory again! Impressive victory.
RADEON X800SE versus GeForce FX5950 Ultra – the latter must hide itself and never appear in public again.
RADEON X800SE versus GeForce 6800 – defeat
RADEON X800SE – GeForce 6800LE – even here it is defeated!!! This is unpardonable for X800SE!
RADEON X800SE versus GeForce 6800GT 128MB (ASUS V9999GT) – no need in comments.

Quake III Arena

RADEON X800SE versus RADEON 9800XT – victory
RADEON X800SE versus GeForce FX5950 Ultra – the same picture
RADEON X800SE versus GeForce 6800 – very ambiguous, approximate parity.
RADEON X800SE versus GeForce 6800LE – victory
RADEON X800SE versus GeForce 6800GT 128MB (ASUS V9999GT) – defeat.

Return to Castle Wolfenstein (Multiplayer)

RADEON X800SE 520/1120 MHz versus RADEON X800XT PE – the difference reaches 30%, considering that even with AA+AF the game still rests on the system capacities.
RADEON X800SE versus RADEON 9800XT – approximate parity
RADEON X800SE versus GeForce FX5950 Ultra – the same picture
RADEON X800SE versus GeForce 6800 – defeat
RADEON X800SE versus GeForce 6800LE – the same picture
RADEON X800SE versus GeForce 6800GT 128MB (ASUS V9999GT) – heavy defeat.

DOOM III

RADEON X800SE 520/1120 MHz versus RADEON X800XT PE – as soon as the game heavily loads the GPU, the difference between these improvised competitors increases
RADEON X800SE versus RADEON 9800XT – weighty advantage!
RADEON X800SE versus GeForce FX5950 Ultra – overall advantage as well, but it's particularly noticeable with AF.
RADEON X800SE versus GeForce 6800 – heavy defeat (only in 1600x1200 with AA+AF the limitation of 128 MB in 6800 has its effect resulting in the advantage of X800SE)
RADEON X800SE versus GeForce 6800LE – the same picture.
RADEON X800SE versus GeForce 6800GT 128MB (ASUS V9999GT) – still larger failure of X800SE.

All NVIDIA cards will obviously look better, probably except for the previous generations.

3DMark05: Game1

RADEON X800SE 520/1120 MHz versus RADEON X800XT PE – weighty difference, which corresponds to reality
RADEON X800SE versus RADEON 9800XT – a slight advantage of the former
RADEON X800SE versus GeForce FX5950 Ultra – there is even no point in commenting
RADEON X800SE versus GeForce 6800 – victory
RADEON X800SE versus GeForce 6800LE – all the more...
RADEON X800SE versus GeForce 6800GT 128MB (ASUS V9999GT) – overall defeat (350*16 is still faster than 425*8).

3DMark05: Game2

Everything is similar to the previous test, except 6800 turned out a little faster than X800SE.

3DMark05: Game3

Identical to Game2.

3DMark05: MARKS

RADEON X800SE 520/1120 MHz versus RADEON X800XT PE – quite a fair difference corresponding to theoretical calculations of the differences between the cards, which confirms again that X800SE is just X800XT with a reduced number of pixel pipelines.
RADEON X800SE versus RADEON 9800XT – a little advantage: optimizations, plus more vertex shaders, and higher frequencies.
RADEON X800SE versus GeForce FX5950 Ultra – vivid advantage, but this is the fault of 5950U, which has a very weak shader unit.
RADEON X800SE versus GeForce 6800 – almost parity
RADEON X800SE versus GeForce 6800LE – victory
RADEON X800SE versus GeForce 6800GT 128MB (ASUS V9999GT) – defeat.

Conclusions

Thus, we can see that in some up-to-date complex games RADEON X800SE is outscored by GeForce 6800, which has an average price of 320-330 USD in late October. That's why the retail price for X800SE must not exceed this figure, otherwise there would be no point in buying this product.

Our tests once again demonstrated that RADEON 9800XT and GeForce FX 5950 Ultra are written off and conceptually dead without any chances to revive.

For those 380 USD (the price of RADEON X800SE) there is a point in buying ASUS V9999GT, which is obviously stronger than X800SE. X800SE will become more attractive only when its prices drop to the acceptable level of $300.

But will it be profitable for ATI to sell these cards? Though if GeForce 6600GT with the same 2.0ns GDDR3 costs 200 USD, then why not sell X800SE for $300, considering that the 128 MB difference is well compensated by those 100 dollars. Plus – they are recycling the chips that do not qualify for X800 PRO. So, time will show. And now we can only watch how the canadian company rages about the leakage of X800SE and that regular customer must not have seen such cards at all. But the reality is against the dreamers from North America again. X800SE has entered the retail market. I suppose it's just a beginning.

In our 3Digest you can find more detailed comparisons of various video cards.

Andrey Vorobiev (anvakams@ixbt.com)

October 28, 2004

Write a comment below. No registration needed!