IIRC the 7300GT has a 128 but interface, so screw your 128 v 256 v 320 memory interface comparison, it's pointless.  You could run a DX9 benchmark (3dmark) on the cards to get an apples to apples comparison, The 8800 wouldn't get to use SM4 or DX10.  The 7 series will suck when compared to the 8800 series card in a real world comparison, because the 8800 is superior in every friggin' way.

well id have to agree with your friend on this due to the fact that it is in a different class due to the fact that a dx9 card cant run dx10 and the hardware itself is built to do so.  So comparing a 8800 to a 7 series card is a lot like comparing a ps2 to a ps3 yea the ps3 is backwards compatible but a ps2 cant run a ps3 game thus making it a different class system and a video card of a pc is much like a console.  You can buy a new consol but in a pc u just buy a new video card.  Wile yes u can compare the older cards in the same way bc they do the same thing but with different hardware so a 6 series and a 7 series do the same thing wile the fact that a 8 series may at the present time do the same thing it has hardware that was made for the speed needed to run dx10 and will still utilize that speed in a dx9 environment.  no matter wat shader its using its still in a dif class.

well id have to agree with your friend on this due to the fact that it is in a different class due to the fact that a dx9 card cant run dx10 and the hardware itself is built to do so.  So comparing a 8800 to a 7 series card is a lot like comparing a ps2 to a ps3 yea the ps3 is backwards compatible but a ps2 cant run a ps3 game thus making it a different class system and a video card of a pc is much like a console.  You can buy a new consol but in a pc u just buy a new video card.  Wile yes u can compare the older cards in the same way bc they do the same thing but with different hardware so a 6 series and a 7 series do the same thing wile the fact that a 8 series may at the present time do the same thing it has hardware that was made for the speed needed to run dx10 and will still utilize that speed in a dx9 environment.  no matter wat shader its using its still in a dif class.

HAHA Nitro..Ur a one hit wonder  (This is the guy who I'm arguing this on  )

Anyways, the way the card is engineered/made is for direct x 10, but since it's not using it...its not running it, its doing the same as any other DX9 card. What gives the 8800 the performance is how it was made and the chipset. The G80 chip used for the 8800s are so much more advanced compaired to 7 series chips. Granit this chip was indeed made for the main purpose of being able to handle higher better physics and shaders along with it's normal duty. So the G80 itself is significantly faster compaired to other cards. But just because it's a good bit faster, doesn't mean you can exclude it from the conversation. He stated that you can indeed compare the 6 series to the 7. For this reason, you can compare the n40 of the 6 series vs the n70 of the 7 series. In short, you can compare chipsets between cards. The chip itself needed to be faster but it thats the only major gain. The hardware/firmware is responsable for the special rendering capibilites, not the chipset. So in short...The chip is faster than the N70, but then again the N70 was faster than the N40. So chipsets between the 3 are comparable. The N80 chip does nothing special on it's on to make it different than the the N70 or even the N40. The difference between all 3 is how they were made and how they were able to get significantly increased performance between them. The firmware is what effects DX10 rendering (which again isn't even being used) not the chip.

Memory is the same between 7 series and 8 series. There is no difference in memory besides the fact that the 8800's are running there gddr3 significantly faster than the 7 series, hence the 51GB/S you get with the 7900 GTX vs the 8800's GTS running at 68 GB/S.

Therefore with this post, I have proven the facts in which the 7 series and the 8 series can be compared because in the end, it's a significantly faster more efficient core clock and memory, vs a 7 series running a good bit slower N70 chipset and slower memory.   I believe people fail to relize the increased performance between the two chipsets. Even though the 7900 GTX (running the N70 chip) is running at faster speeds than the the 8800 series (650mhz vs 500mhz on a stock 320 GTS) the 8800 outperforms the 7900 by a good 15 FPS. It is running a slower clock, but the overall design of the N80 vs the N70 makes up for lower clock speed. Overall the N80 was designed with the thought in mind of running DX10, but even if DX10 wasn't even out, Nvidia would of released the new series with a new faster core, because thats how they continue to gain there perfomance (N40 vs N70..and now N70 vs N80). They knew they needed a significantly faster core clock to handle the new physics and such that the card will be forced to handle in DX10, but they would of came out with a faster core no matter what anyways.

at no point did i argue that the 8800 would win in an aqua test im simply arguing that the 8 series is in a different class.  thats all im saying that its a completely different card now the same can not b said for the 7 and 6 series due to the fact that they cant run dx10.

DX10 is 90% software base. Not hardware. Just like DX9 and DX7 before that. The only things that separates the DX10 cards from DX9 is the firmware supporting DX10 and the shader models DX10 uses (not DX9). It does not run the shader model or DX10 rendering firmware until it is called upon in DX10 applications, then it kicks in and performs. If you could get the supported firmware (which would never happen cause 7 series is dead now and not going to DX10 I think any time soon) for a 7900 or a 7800..it would be able to run. Though it wouldn't be able to run as efficiently and would fail to be able to render proper shader models and effects (since it does not have a separate shader clock and is running pipeline instead, that and SM4 is more graphically intense and requires a faster GPU than SM3)

With this said, you can run a TNT 2, a MX440 or anything in between in DX9 format. The card would be able to run in DX9, but when running, it would have to leave out SM2 or SM3 technology it lacks having. I can run a tnt2 card (I have run running..came out in like 99 no less too) and put in Unreal running DX9 and be able to run it fine. The card would be lagging like crazy but it can be done/ran none the less. Same concept if you could put DX10 firmware into a DX9 card..it could run it, just not as efficient...but non the less, running it  is running it. The shader clock is the only difference between these two cards in retrospect. You can compare the core and the memory all you want because essentially its similar just running faster. The cores have been designed faster, but not made specifically to run DX10, again the firmware is what helps that. So in truth, it's a 7 series running 20 pipelines vs a 8 series running a individual shader clock instead of a pipeline setup (giving more gpu speed since shaders are not running off the same as the GPU..Again this is for the improved physics and graphics rendering DX10 cards will be forced to handle) The shader clock however does not offer a significant improvement over DX9 cards.

There is no way you can say that the 8800 is running 17-20 FPS faster just because it has a separate shader clock to help render all 3% of the gpu performance running SM3. If it was running a more intensive GPU base SM4, it'd be different..But again, it's only running SM3 in DX9. So is 3% less stress off your GPU equivalent to 18-20 FPS gains in just the base 8800 GTS models. I hardly doubt it. Again, the faster Core and Memory is the main result in the increased difference. One last comparison to prove my point on this topic then I am done with this, because with this post I truly feel that I have more than made my point and supported it (if you still think you can't compare apples to apples, then your never gonna change your mind about it, but I would hope you would read this with a open minded). The base line 8800 runs 500 mhz core and 1700 mem clock with 63GB/S bandwidth. A Evga Superclock edition runs a faster 576 mhz core and 1800 mhz mem clock with 68 GB/S. The difference in FPS between a base line and a superclock is 15 fps on average. So it proves my point the drastic FPS increase you get with just a slight increase in speed in memory and core. As always, your thoughts are welcomed on this matter

HAHA Nitro..Ur a one hit wonder  (This is the guy who I'm arguing this on  )

Actually he is a On Hit Wonder.  Someone forgot the E. 

And I have to agree with F3ar0n and Nitro on this topic.  Of course the 8800 is in a different class of video cards than the 7 series.  They were designed from the beginning to outclass the 7 series.  That is the whole point.  But you can compare the two, and the 8 series blows the 7 series away in an apples to apples comparison.  Nitro, a similar argument you could try and take is that you can't compare a 6 series vs. a 7 series cause they are different classes of cards.  Just because the card is built for DX10 doesn't mean a thing.  It still has to run DX 9 and do it better than a 7 series card for people to upgrade to it.  So yes it is a different class of card, but you can compare it to a 7 series.