This system is still much more complicated than necessary. ATI's new naming scheme (no suffixes) is supurb.

The current and abandoned naming scheme is sending the wrong message.

Several years ago friends of mine who are knowledge-lacking computer users ask me which graphics card to buy... The first thing they say is "Wait a minute, why get a 7900gt when I can get a 8300? The number is higher and is cheaper!" Then I have to explain to them that that is not how the numbering works.

What they need to do is have 2 numbers. E.g. Consider "a 80GTX2". The "80" is the product performance number, the GTX is the suffix, (if one should even be there at all... they are stupid anyway), and the second number, "2", is the generation. You might see how this would be equivalent to a gtx280.

This scheme suggests that when you get a new generation, you can just update the 3rd number.

Any thoughts?

Well, the way I have seen it, it already kind of is named like that. The 8300GT for instance:
8 is the series
300 is the "level" (ie higher number is more power)
the suffix could be droped.
Different revisions could change the last number (8301, 8302 etc.)

For all these years i dont know if Nvidia actually did any good marketing at all apart from confusing the market with MASS of Products.

Just use Numbers, Drop All the Letters.
I must argee, out of ATI Graphics, Intel CPU and Nvidia, ATI graphics has the best naming scheme.

I think this will definitely help, but it is still not great. "lotharamious" has a good point, in that no matter what letters and numbers they use, the performance indicator should come *BEFORE* the generation, so that people can recognize that an older generation performance part is going to be better than a newest-generation budget part.

ATI by far has the best nameing scheme
2900
3850
3870
3870 X2
4850
4870
4870 X2
very easy to figure out witch one is the best 2,3,4 is the series the level is well 900, 850, and 870,

nvidia on the other hand

7900 gt
7900 gtx
7950 gt
7950 gtx

8300
8600 gt
then you have the 8800 gs with 300 and some ram and then the 8800gt with 640 mb ram then the new revision 8800 gt with i think 512 that was almost as good as the 8800 gtx with 864 mb ram then the 8800 ultra with 864 mb ram.

then you get into the 9000 sereis witch isn;t even a huge boost in performance from the 8000 series just some more futures like tripple SLI. and the same thing happands you have 9800gx2 and then the old 8800gx2 and if you were to read that wrong it looks like a gtx. witch i don't beleive if you are going to buy that card you have to know something at least about computer and could igure it out pretty easly

i am an ATI fan i am canadian and they have worked great for me in the past and still work great for me now. i have dual 4870 x2's and can run any game i want for at least a couple years. and i don't know how you can handle that shitty interface to setup the nvidia graghics cards i hate it ( have a 8600 gt in my laptop, nvidia right now is better for laptops i know that) the ati catiylst controle ceneter works so much better. aspecialy when useing your 1080P tv as a moniter so many more options to use for the. use the HDMI out of my laptop is just hell when hooking it up to watch a movie. one of these days nvidia will do somehting a little better i hope.

Its not just Nvidia. ATI is just as bad. They need to stop polluting the market with 6 different SKU's of the same card that only differ from each other by maybe 5 fps. It was so much easier back in the GeForce 4 and earlier days. You only had the Ti4200, Ti4400, and Ti4600 to choose from.

they were way more confusing

"It was so much easier back in the GeForce 4 and earlier days. You only had the Ti4200, Ti4400, and Ti4600 to choose from."

I hope that was a joke - the GF4 series contained the infamous GeForce 4 MX, which, though labeled as a GF4, was really a GF2 and offered no shader support (breaking from the rest of the series AND the previous gen GF3).

the whole point of nvidia doing what they do with their names is to fool people that are noobs. they don't have a new generation card for at least a few years now and are all the same gpu cards with different names and are over clocked differently with price tags to go with the overclock. you would be surprised how many people fall victim to nvidia`s naming scam. they are doing it all over again so you think there old cards are new with the new names e.g. all g8 g9 etc. cards now will have GT200 series names. nvidia is the biggest scammers I have ever come across. there prices are ridicules to, well until AMD/ATi put them in there place. only now do we see cheaper NV cards, which was a joke B4. plus all g8/g9 and even gt200 series cards have problems once they reach their thermal limit a few times (they are all faulty gpu cards). that is the biggest scam which makes the names look like a good thing. now run along and tell your mates or famil;y not to buy nvidia cards.

Those companies need to have a very profitable -evil reason- to use those gut-hate confusing labeling.

I hate also a lot the Intel naming for his processors.

I wish there existed an independent cpu and GPU index forced to being put in the naming.

Since The G80 (ex. 8800GTX) They had G92 (ex. 8800GT) G94 (ex. 9600GT) and G200 (ex. GTX280)

Because I keep track on everything I know what's best, but many don't.

It is very hard to create a perfect naming scheme, but this would do (for renaming the old series):

Nv8700 (for the 8800GT)
Nv8800 (For the 8800GTS)
Nv8900 (for the 8800GTX)
Nv8900Ultra (for the 8800GTX Ultra)

Nv9600 (for the 9600GT)
Nv9700 (for the 9800GT)
Nv9800 (for the 9800GTX)
Nv9900 (for the 9800GTX )
Nv9900X2 (for the 9800gx2)

And for the new series:


G250 (for the gtx260)
G260 (for the gtx 260 55nm)
G270 (for the gtx260 core 216)
G280 (for the gtx280)
G290 (for the gtx280 )
(and if any updates is neccesary for the competition they just add a '5' ex. (G265)

Well ignoring the "I know what's best, but many don't" comment, your naming convention still doesn't address one of the biggest complaints; that the series low-end card is often slower than previous series cards yet sports a larger number (implying it's faster or better in some way to unknowing user).

For instance if nVidia releases an entry level 9300 card, which presumably would be slower than the 8800 cards, what then?

In some respects the old AMD cpu naming conventions were great; an Athlon 64 4000 may not have run at 4.0Ghz, but you definitely knew how it stacked up to a previous core-generation Athlon 64 3500 or 4400. The naming was relative to the performance, which was great.

"the company would lay off about 7% of its staff..."

sounds like Nvidia just layed off its numbering system department.

:P