[FUD] GT300 is codenamed Fermi

[Nightcover]

Named after a nuclear reactor


The chip that we ended up calling GT300 has internal codename Fermi. The name might suit it well as Enrico Fermi was the chap that came up with the first nuclear reactor.

The new Nvidia chip is taped out and running, and we know that Nvidia showed it to some important people. The chip should be ready for very late 2009 launch. This GPU will also heavily concentrate on parallel computing and it will have bit elements on chip adjusted for this task. Nvidia plans to earn a lot of money because of that.

The chip supports GDDR5 memory, has billions of transistors and it should be bigger and faster than Radeon HD 5870. The GX2 dual version of card is also in the pipes. It's well worth noting that this is the biggest architectural change since G80 times as Nvidia wanted to add a lot of instructions for better parallel computing.

The gaming part is also going to be fast but we won’t know who will end up faster in DirectX 11 games until we see the games. The clocks will be similar or very close to one we've seen at ATI's DirectX 11 card for both GPU and memory but we still don't know enough about shader count and internal structure to draw any performance conclusions.


Of course, the chip supports DirectX 11 and Open GL 3.1

http://www.fudzilla.com/content/view/15713/1/

[Arkangyl]

So we're looking at high heat output and instability?

Biggest architecture change since G80 sounds good though, sort of the only 'change' since then though (others seem to just be additions)

[Bodkin]

Hopefully this is an indication of 2009 launch. i.e. stuff starts leaking well before launch. If the net had stayed silent much longer it would have been confirmation of a 2010 card. Still FUD is FUD so not to optimistic

[BulldogPO]

Interesting times ahead.

[rusty]

as long as it's not Chernobyl it's ok

[zalbard]

as long as it's not Chernobyl it's ok

[ArneMS]

as long as it's not Chernobyl it's ok

Haha, good one!

[Helloworld_98]

So we're looking at high heat output and instability?

I'd hardly call nuclear fission unstable, 1 failure in 55 years with 100s of reactors isn't exactly a high number.

[AAbenson]

im fully expecting a NVIDIA limited xmas launch of just one product line at around 400$ ,enough for nV to claim a hard launch title.
a mass nV launch with complete availability of all product lines imo will be coming next year possibly having to compete against an ATI top product line refresh.
the really important thing for the 2 companies is that once again ATI will do better financially and gain further marketshare.
...but yea,nV will have the braggin rights of claiming the "performance crown" once again which in reality is pretty pointless as we have seen with current nV top line and the 4890

as long as it's not Chernobyl it's ok

Chernobyl was a reactor thats main pupose was to enrich uranium for the military,production of electricity was an afterthought.
Compareable reactor with the same purpose ie mixed military/civilian would be the one in South Carolina which was "officially" closed sometime ago,it was also used for "reenriching" spent nuclear fuel.

[tdream]

I'd hardly call nuclear fission unstable, 1 failure in 55 years with 100s of reactors isn't exactly a high number.

Uh that's what makes it work, instability releases energy. That is high when you consider if you have a meltdown the land is fked for the next 1000+ years.

[Glow9]

It sounds like a 5870 with Nvidia branding.

[Bodkin]

I'd hardly call nuclear fission unstable, 1 failure in 55 years with 100s of reactors isn't exactly a high number.

Three mile island, Windscale, K-19?

[GrossmeisterB]

I'd hardly call nuclear fission unstable, 1 failure in 55 years with 100s of reactors isn't exactly a high number.

There have been a lot more than ONE failure in 55 years!!!!
The one you are talking about is the biggest known failure in civilian usage...what about Hiroshima and Nagasaki?
What about all the storage facilities in germany with leakages? There's a lot of radiation going into the ground and the atmosphere!
What about Sellafield 1957? Idaho 1961?

[trinibwoy]

Lol those were BOMBS dude, not reactor failures

[GrossmeisterB]

Lol those were BOMBS dude, not reactor failures

I think those Japanese are still laughing about those bombs
We were talking about exposition of nuclear radiation - no matter if it were bombs or power plants or some storage facility!
So stop making fun of me!

And yes, I do work for a company who owns more than one powerplant here in germany, and yes, I do have visited a nuclear powerplant several times, so I think I know what I am talking about here...

[Jamesrt2004]

"has billions of transistors and it should be bigger and faster than Radeon HD 5870. The GX2 dual version of card is also in the pipes."

so this just mean it has high power consumption + cost + heat output... but is going to be damn fast,

surprising there making a gx2 for this "supposedly" massive chip unless is not made from 2 of the top chips and maybe something one step down, think gtx 295 was 2 275's not 285's
we shall see not long to go now it seems

[trinibwoy]

Man, that's a toss up right now. Some people are saying we'll know more soon and some people are saying no availability for 6 months. It's a pretty good bet that the truth is somewhere in the middle

[Origin_Unknown]

*sigh* once again another news topic goes wonky and off topic

[Ozzfest05]

Do you guys honestly believe NV gt300 will beat a 5870x2 which is right around the corner im asking out of curiosity because the 5870 is a beast.
Nevermind 5870x2 but as the article stated there will be a gx2 version.
Is nvidia smart enough to make it single PCB and not make a dual PCB then a conversion once again like the 295
I hope to see some pricing wars through this competitiong however.
Im all for both companies neck and neck its better for my wallet lol.

[trinibwoy]

It really comes down to how much headroom there is in engineering. Everybody looks at manufacturing process and die size etc, etc. But does anybody ever ask if engineers are making the best use of available transistor budget? If we look at just the usual stuff and assume that both Nvidia and AMD engineers are doing the best they can with available technology, then no it won't beat the X2. If we allow for them finding novel ways to do things then there's a slight possibility.

[>HyperlogiK<]

There just isn't enough technical info available for anybody not under an NDA make any kind of informed guess; sometimes these huge monolithic chips are beasts (G80, G200, R200) other times they are over-hot failures (NV30, R600). There is the rumour from Fuad that it will beat the ATI offering and because that is the only slight hint of a performance comparison available to me [and because it's plausible to think we might get a repeat of the previous generation] I would guess that it might be the case, but it would be a big and not very well founded guess.

[Solus Corvus]

Lol, all you guys think of are nukes. Fermi helped kick start atomic knowledge. Work in that field can improve semiconductors as well.

This card has a lot to live up to if they really used that codename. Personally I would like the dual gpu version to be a single pcb, full shader count/speed cores, and come stock with watercooling, lol.

[Russian]

As long as this thing doesnt require the same level of cooling as a nuclear reactor, im good.

[Bodkin]

...and come stock with watercooling, lol.

mmmmmm +1

[Generic user #2]

what about Hiroshima and Nagasaki?

its not a failure it is MEANT to explode.