NV18 And NV28

mboeller said:
The MX440 will have an core with 275MHz and 256MHz memory, so this card will have higher performance. c't managed to overclock their test-card to 351MHz core / 338MHz memory. So this card will be an really high performance DX7 card for overclockers (if all MX440 can be overclocked so much ). Otherwise the new MX440 will now be (roughly) as fast as the R9000 but lacks the DX8 features. So we will see how this turns out for Nvidia.
Click to expand...

The MX brand since introduction and to this present day should be called the "rip off brand". They have been slow, underperforming parts all along that sell because stupid people see GF4 and don't realize they're getting a dumbed down version of the better card. Most people probably never realize how badly their card performs though, so no big loss. I still hate these "value" cards though. As if you can get a seriously performing card for <$100. :rolleyes:

alexsok said:
NV30 is late not because of any problems nVidia were having with the design, but because of the problems with 0.13 not being mature enough, which is the main and the only reason for the constant delays.
Click to expand...

I think the point is: if it's late, it's late. ;)

I just hope they finish it soon. If ATi gets the R9500 to market before the NV30, I think it's going to be bad for Nvidia.
 
It ain't even a dumbed down version. Think geforce4 MX(aka geforce 2.1 at best), thet biggest slap on the face at consumers to this day. Completely overpriced and underspeced.

Oh yeah, and where is the almighty 63 million transistor NV18?? :LOL: :rolleyes:

NV30 is late not because of any problems nVidia were having with the design, but because of the problems with 0.13 not being mature enough, which is the main and the only reason for the constant delays.
Click to expand...


Yes it was all an evil conspiracy against Nvidia. Give me a break. Nvidia saw R300 on .15micron with full floating point pipelines and ran back to the drawing board so scared, they forgot to plug the earlier leaks about it being taped out. That's what you get for resting on their laurels for so long

I guess the 3dfx butchering is biting them in the ass now. Huge cards, BIG production delays exactly like 3dfx. It seems to have paid off dont ya think? :LOL:
 
Nagorak said:
The MX brand since introduction and to this present day should be called the "rip off brand". They have been slow, underperforming parts all along that sell because stupid people see GF4 and don't realize they're getting a dumbed down version of the better card.
Click to expand...

While I agree that the GF4 MX isn't a hot card I couldn't disagree more on the GF2 MX. It had excellent value with good performace when it first appeared. It also had all the features of its big brother. I wouldn't call this a "rip off".

Patrik
 
alexsok said:
I was wondering myself about ti4600 (ti4400 ceized production, so no suprise there), but i thought that an agp 8x variant of ti4600 is a possibility...
Click to expand...

I think there is a good chance nVidia are waiting for the radeon 9500 to show its colours before deciding on the final specs on a AGP 8X 4600 so they can position it correctly against ATI's offerings. This way they can avoid a repeat of the Radeon 8500/MX460/Ti4200 situation.

Patrik
 
glappkaeft said:
I think there is a good chance nVidia are waiting for the radeon 9500 to show its colours before deciding on the final specs on a AGP 8X 4600 so they can position it correctly against ATI's offerings.
Click to expand...

They don't compare at all and never will. The ATI-9500 is DX9 compliant, Nvidia-4600-8xAGP isn't. As I believe an Nvidia rep said, this fall you're either DX9 or you're not. No two ways about it.

--|BRiT|
 
BRiT said:
glappkaeft said:
I think there is a good chance nVidia are waiting for the radeon 9500 to show its colours before deciding on the final specs on a AGP 8X 4600 so they can position it correctly against ATI's offerings.
Click to expand...

They don't compare at all and never will. The ATI-9500 is DX9 compliant, Nvidia-4600-8xAGP isn't. As I believe an Nvidia rep said, this fall you're either DX9 or you're not. No two ways about it.
Click to expand...

Considering that there will be no DX9 games to play the next 2 years or so the cards are most definately comparable IMO. It will be different in a year or so when the first DX9 games will be on the horizon but until then they will compete on performace (and IQ, FSAA and aniso).

Patrik
 
NV30 is late not because of any problems nVidia were having with the design, but because of the problems with 0.13 not being mature enough, which is the main and the only reason for the constant delays.
Click to expand...

Its NVIDIA problem for 'not' choosing 0.15, ATI did fine with R300 on 0.15, so should NVIDIA. Perhaps R300 is what Geforce 4 should have been.
 
NV30 is late not because of any problems nVidia were having with the design, but because of the problems with 0.13 not being mature enough, which is the main and the only reason for the constant delays.
Click to expand...


I also would like to add to V3's post that there are other designs that have been shipping for a "while" now on .13 micro process. So its a combination of nV complexed desgin and a still un-mature process. However since no one was forcing nV to use a .13 process they have to be held accountable for thier chocies...
 
gkar1 said:
Oh yeah, and where is the almighty 63 million transistor NV18?? :LOL: :rolleyes:
Click to expand...
In fact it was supposed to have 81M transistors (if one read the slide wrong, that is). And where is the NV28 with 86M transistors?

And no, not even the Crush18 will have that many transistors. The nForce1 had some 40M transistors, both NB and SB included.
 
I know its marketing speak, but this stands out;

'Today’s fast-paced “fly-through” applications and games will see significant improvements in overall performance due to the doubled AGP 8X speed. By combining the performance and scalability of AGP8X GPU and a high performance 128MB DDR frame buffer; the GeForce4 Ti 4200 with AGP 8X GPU delivers more than two times the graphics performance of competitive products'

As I know 2 9700 reviews are in the offing at this site, any chance they will involve testing on an 8x AGP mobo?
 
As I know 2 9700 reviews are in the offing at this site, any chance they will involve testing on an 8x AGP mobo?
Click to expand...

- AGP 8X Compatibility

There have been many reports that some current AGP8X motherboards do not allow stable operation with AGP 8X graphics cards. As you can see from the below image, we had no trouble running the MX440-8X at AGP 8X on our SiS648-based Abit SR7-8X motherboard which we reviewed recently.

prolinkmx4408x_agp8x.jpg
Click to expand...
 
Which makes the MX440 a complete dof as its still not faster than the 9000Pro.

Anyway I wasnt talking about compatibility, I was talking about B3D's normal testing of performance gains going from 2x to 4x AGP being extended to 8x AGP.
 
Anyway I wasnt talking about compatibility, I was talking about B3D's normal testing of performance gains going from 2x to 4x AGP being extended to 8x AGP.
Click to expand...

Oh... the gain is close to zero actually, as AGP 8x offers practically nothing (speed wise) in current games over AGP 4x.
 
Anyway I wasnt talking about compatibility, I was talking about B3D's normal testing of performance gains going from 2x to 4x AGP being extended to 8x AGP.
Click to expand...

Oh, that would be something we'd look at if we can. I suspect that by the time we get an 8X MB others will have done it already though!
 
Here is an article of NV28/18 vs NV25/17 vs R9000/Pro (Chinese):

http://www.pcpop.com.cn/read.asp?id=584

Actually, NV28/18 is nothing more than NV25/17 with AGP 8x. If and only if volume vertex data and/or texture data excess the capacity of the local video memory, AGP 8x can show its muscle. However, there is almost no game which needs more than 64MB video memory currently, including UT2003.
 
You must log in or register to reply here.
Back
Top