Sign in with
Sign up | Sign in
Your question

GTX4xx Slower than other technologies...

Last response: in Graphics & Displays
Share
March 3, 2010 4:56:58 AM

Sad to say we are seeing 6 core boards, Ipads, usb 3.0, the cypress chip, H55's, P55's, ION2,ect roll out faster than Nvidia's Fermi which took almost 2 years to be released! In that span of time, so much technology has come and been properly introduced to the mass market...

We still have yet to have any OFFICIAL Specs from Nvidia on what their GF100 chip can do for us... and why we should buy it?
a b U Graphics card
March 3, 2010 5:20:49 AM

Wait and see dude, the official release date is less than a month away...

I for one predict AMD/ATI's 5x00 prices to drop or at least stabilize to their intended retail market value. TSMC's 40nm production line will also be ramped into full gear by the end of March, so dx11 card shortages will be coming to an end. This hopefully means the revival of the GPU war and lower prices for consumers... :D 
m
0
l
a b U Graphics card
March 3, 2010 10:57:00 AM

liquidsnake718 said:
Sad to say we are seeing 6 core boards, Ipads, usb 3.0, the cypress chip, H55's, P55's, ION2,ect roll out faster than Nvidia's Fermi which took almost 2 years to be released! In that span of time, so much technology has come and been properly introduced to the mass market...

We still have yet to have any OFFICIAL Specs from Nvidia on what their GF100 chip can do for us... and why we should buy it?


G100 was a new arch, with major internal revisions. Of course it took longer to develop then a 4000 series refresh.
m
0
l
Related resources
a c 271 U Graphics card
a c 171 Î Nvidia
March 3, 2010 1:55:20 PM

gamerk316 said:
G100 was a new arch, with major internal revisions. Of course it took longer to develop then a 4000 series refresh.

Not forgetting TSMC's troubles, which may not completely resolved despite what their CEO keeps telling the press.

http://www.digitimes.com/news/a20100226PD210.html
m
0
l
a b U Graphics card
March 3, 2010 3:12:34 PM

^^ Saw that too; TSMC is predicting mid-May for everything to get sorted out.
m
0
l
March 10, 2010 5:34:46 AM

If and when AMD brings out their fusion CPU/GPU discrete card out, and this might be the case by the 4th quarter of 2010.... then Nvidia's GF100Chip and the gpu line it carries will already be old news and severely outdated....

Ati would have had 2 FULL GPU cycles while Nvidia will still be figuring out its market strategy to boost sales or by that time they will be releasing cheaper gtx420, or 450's......
m
0
l
a b U Graphics card
March 10, 2010 5:46:14 AM

Flame war + troll attack! Mend the fences!!!

On topic:
A usual GPU architecture takes 1.5-2 years to make. The real problem with Fermi is that it was not meant to be from the start a gaming card. So a change had to be made at some point so that it could enter this market. That alone took some time to do and in my opinion is the reason Fermi is very late to the party. I have no doubt Fermi will be great in the future, but I have serious doubts it will be able to compete at the current moment.
m
0
l
March 10, 2010 10:21:46 AM

:kaola: 

I hope some one has an oversight and makes a mainboard which does not require a CPU. So we have a pc which boots up from the GPGPU. Then Fermi will be the best.
m
0
l
March 20, 2010 1:09:59 AM

Quote:
Gotta support you on this one. From the beginning Fermi was designed for GPGPU applications, an evolution above the previous folding and data crunching results. When fermi news was being release half a year ago, it all centered around data-crunching applications. It was designed to be a supercomputer packed into a graphics card, hence the native C++ support and Tesla applications.

Once the GF100 chip gets sold as a TESLA card, it'll blow the previous hardware options out of the water. Any data center, university reseach department, or anaysis department will want to buy these things... in bulk. Because it'll be magnitudes cheaper than any other supercomputing towers using Xeno/Opteron servers yet with easily programmable applications.

I am not worried about Nvidia's ability to sell fermi chips. I just don't think they'll sell in the GPU market. All winds point to the GTX400 cards being overpriced for their FPS results... assuming a $500/700 price tag for the 470/480 respectively.

I agree that this would help with the Tesla and perhaps Quadro markets for animation, graphics design, and data center applications, and perhaps Nvidia is thinking at the wholesale AND enterprise level. The quadro cards were so expencive and the cheapest you could get a decent one was about $499 but they go up to $8000!!!! Actually pricing for these GF100 Fermi cards are suprisingly low for new cards! Yesterday there was an article here at Toms stating that the prices for the GTX470 will be around $360, while the GTX480 will only be $499. If this is any indication of the TRUE release prices, then I have to say, for the first time, Nvidia has hit the spot in terms of competitive pricing with ATi and this has not been the case in the last 6 years or so! Let the DX11 and onwards GPU's finally begin.
m
0
l
March 30, 2010 2:44:12 AM

Best answer selected by liquidsnake718.
m
0
l
!