News Headlines
- Wed, Jun 19
- The War Z becomes Infestation: Survivor Stories citing trademark conflicts, game otherwise unaffected
- Microsoft's One Mistake: Pressing reset on the Xbox One's aspirations of a digital future
- Surprise! Company of Heroes 2 beta extended through June 23
- Microsoft officially canceling Xbox One online and used game policies across the board
- Nintendo wins appeal over Wii Fit Balance Board patent litigation, probably not over
New Articles
Related Articles
Power Usage
To measure power usage, we used a Kill A Watt P4400 power meter. Note that the above numbers represent the power drain for the entire benchmarking system, not just the video cards themselves. For the 'idle' readings we measured the power drain from the desktop, with no applications running; for the 'load' situation, we ran a demanding part of 3DMark06.
The 9800GTX uses appreciably less power than the 8800 GTX. The optimization of the unified shader architecture has come a far way in the the last two years. Compared to the 8800GT, the XFX 9800GTX requires a reasonable amount of power -- about where you'd expect to be. NVIDIA recommends a 450W power supply or greater for the 9800GTX, while XFX recommends a 620W power supply or greater. 620W seems to a bit conservative -- a 500W or greater should do the trick, I would think. However be wary if you have a cheaper 500W PSU, you might be starving your system of a steady supply of power.
Regardless of what PSU you have, you'll need two six-pin PCIe power connectors to use your card. The bundle does come with one molex-to-PCIe adapter though, so if your PSU only has one PCIe connector, you should be fine.
Overclocking
After numerous experiments, we had trouble overclocking this video card in Vista. This will probably not be a factor with later driver revisions, but for now, it seems that overclockers are best suited to experiment using Windows XP.
We had high hopes for overclocking the XFX 9800 GTX, thanks primarily to the effective cooler it utilizes. These high hopes were met in our overclocking expeditions. While the stock clocks for this card are 675 / 1688 / 2200 (core / shader / memory), after setting the fan run at a 100% using RivaTuner, we were very happy with how far we could push the XFX 9800 GTX. We had no stability problems running the card at a 819 / 2050 / 2520 setting, which is very impressive, most especially in regards to that shader clock overclock.
It is possible that this card can be pushed even further, and we will be investigating this.
XFX definitely has ample opportunity to release a 'XXX' version of this card. With a volt mod or extravagant cooling measures, it'll be interesting to see how this video card can be pushed.
Article Index
|
|
next
1 2
9800GTX or a Palit GeForce 8800 GTS 1GB Sonic
http://www.legitreviews.com/article/683/1/
http://www.palit.biz/main/vgapro.php?id=631
__________________Core___ Shader__Memory___Memory_______Memory
__________________Clock___Clock___ Clock___________________Interface
__________________(MHz)__ (MHz)___(MHz)_________________________
8800 GTS Sonic____730___ 1625?____1050_____1GB GDDR3____256-bit
9800 GTX__________675___ 1688_____1100_____512MB GDDR3__256-bit
This review here from neo mirrors what the one from hardocp says, practically. And they both are saying that the 8800GTX and GTS are practically the same performance as the 9800GTX, sometimes better, sometimes worse, but always within a few fps.
Gay if you ask me.
Come on, tell me it's a joke. TELL ME!!!
Whoa whoa whoa whoa hold up.
What's wrong the graphs? They appear fine to me? What info is missing from the axis? Have you guys ever had that problem with older graphs?
Please let me know. Really concerned. Thanks.
Hey !
My next article is a look at the Palit 9800 GTX, with a focus on SLI performance.
Going to compare 8800 GT sLI, 9600 GT SLI, 9800GTX SLI, and 8800 GTS 512MB SLI.
Should be going up on Monday.
But that doesn't mean the 9800GTX is a bad card.
Many reviews, such as the one linked, trashed this card because it isn't much of a performance jump from the 8800GTX. But in my opinion, that is sort of throwing out the baby with the bath water.
Ya it is confusing that NVIDIA totally mixes stuff up often. If naming conventions were written in stone than the 9800GTX should be $600 and offer twice the level of performance it does.
But if you totally don't look at the names at all, and compare this card to what is available, and for what it offers for $300 - $350, it is a good card. Of course this is all my own opinion -- I can see your point of view, and many reviewers share your point of view, but I don't feel this way.
As for the testing methods, you might be happy to hear I'm going to be switching Crysis to a FRAPS run instead of the built-in bench, which was shown by OCP to be not the best bench. I'd be open to more suggestions on improving testing methods, but as I said before, I can't FRAPS run 6 games and 8 cards at different settings for every review, that's just not reasonable. Although my current system isn't perfect (no system is) the results are not totally unreliable either.
NVIDIA is already hyping up there next generation of cards, so expect that gen to offer the huge jumps.
The 9800 GTX can be considered a 8800 GTX II. In a perfect world it would have been sold as a 8900 GTX.
Nonetheless it isn't a bad card, as I've been saying.
One thing to keep in mind is that when the 8800 GTX came out, it did offer a huge jump in performance, but it also came on the market around the $600 mark, whereas this GTX will be selling for about half that.
next
1 2