The official GT300/Fermi Thread

**Solus Corvus** · 10-07-2009, 11:30 AM

Originally Posted by Luka_Aveiro

Seriously, why care with games on GT300? It'll smash every piece of code in any game so far released, so why marketting the obvious...?

Well gaming performance will be a factor in my decision to get GT300 or not. So I care, lol.

**Calmatory** · 10-07-2009, 11:33 AM

Originally Posted by Dante80

Or you think that 500$ cards are going to sell by themselves? Regardless of performance...

Nvidia certainly does.

I see them moving away from "gaming market" more and more, so why bother?

**v_rr** · 10-07-2009, 11:33 AM

Why DX_11 isn't listed here?

**Luka_Aveiro** · 10-07-2009, 11:37 AM

Originally Posted by v_rr

Why DX_11 isn't listed here?

Because it won't have!!!!

Or because Dx11 is such an obvious and cliché feature that it needs no market at all.

Why isn't AA listed there?

palhaço

**JoeBar** · 10-07-2009, 11:37 AM

IMHO this thread should be closed and reopen when something concrete comes up...

**Helloworld_98** · 10-07-2009, 11:40 AM

Originally Posted by v_rr

Why DX_11 isn't listed here?

It doesn't, hardware wise, support DX11, supposedly the CUDA cores will handle that.

**Tim** · 10-07-2009, 11:43 AM

Originally Posted by JoeBar

IMHO this thread should be closed and reopen when something concrete comes up...

Agreed.

**003** · 10-07-2009, 11:45 AM

Originally Posted by Helloworld_98

It doesn't, hardware wise, support DX11, supposedly the CUDA cores will handle that.

It does, hardware wise, support DX11. Whether through fixed function units vs. highly efficient and scalable cores is very irrelevant (and both are still "hardware", mind you), as it will still end up faster than RV870 and that's really all that matters.

**Helloworld_98** · 10-07-2009, 11:46 AM

Originally Posted by JoeBar

IMHO this thread should be closed and reopen when something concrete comes up...

but where else will we post our ramblings about how the GT300 is gonna suck compared to the 5xxx overall?

**Calmatory** · 10-07-2009, 11:51 AM

Originally Posted by 003

as it will still end up faster than RV870 and that's really all that matters.

...to you.

**~~eleeter~~** · 10-07-2009, 11:54 AM

Originally Posted by 003

...as it will still end up faster than RV870 and that's really all that matters.

Well except that is not all that matters. GT200 was the fastest single chip of its generation, but Nvidia got a

kicking of a lifetime as far as margins go.

**Sly Fox** · 10-07-2009, 12:03 PM

Originally Posted by eleeter

Well except that is not all that matters. GT200 was the fastest single chip of its generation, but Nvidia got a

kicking of a lifetime as far as margins go.

Yup.

**VoodooProphetII** · 10-07-2009, 12:12 PM

If only.....ahhh shame.

**golden_boy** · 10-07-2009, 12:33 PM

Originally Posted by Luka_Aveiro

Because it won't have!!!!

Or because Dx11 is such an obvious and cliché feature that it needs no market at all.

Why isn't AA listed there?

palhaço

You had some sense in your "petit" joke/sarcasm but these slides are showing something that doesn't appeal to the masses (the ones that truly represent big sales), that is, gaming...

cuidado com a língua

**BenchZowner** · 10-07-2009, 12:46 PM

Originally Posted by Helloworld_98

It doesn't, hardware wise, support DX11, supposedly the CUDA cores will handle that.

Good lord, why would you say that when you obviously don't know about it ?

Not as misleading as other people but ffs, if you don't know something at all, better say nothing.

**Jamesrt2004** · 10-07-2009, 12:49 PM

Originally Posted by BenchZowner

Good lord, why would you say that when you obviously don't know about it ?

Not as misleading as other people but ffs, if you don't know something at all, better say nothing.

+1

for everyone else... it DOES support dx11 btw.... it jsut handles a couple features like tesselation through a software "layer" first.. not a biggie maybe a small performance hit but still 100% complies

**zerazax** · 10-07-2009, 01:09 PM

Originally Posted by Jamesrt2004

+1

for everyone else... it DOES support dx11 btw.... it jsut handles a couple features like tesselation through a software "layer" first.. not a biggie maybe a small performance hit but still 100% complies

Curious but how would you know it's a "small" hit?

Considering PhysX has its own performance hit, and sometimes it isn't that small, so how do you know that is so? Of course, Nvidia knows there aren't any tesselation games out right now it seems, but its still that's quite an assumption

**003** · 10-07-2009, 01:33 PM

Originally Posted by zerazax

Curious but how would you know it's a "small" hit?

Considering PhysX has its own performance hit, and sometimes it isn't that small, so how do you know that is so? Of course, Nvidia knows there aren't any tesselation games out right now it seems, but its still that's quite an assumption

The difference is that ATI has a fixed function unit for tessellation, while nvidia does it on the SPs. Nothing wrong with that if it's still faster. Granted we don't know yet how it will perform, but I would say with a lot of certainty that it will be faster/better than RV870.

**marten_larsson** · 10-07-2009, 01:38 PM

Originally Posted by 003

The difference is that ATI has a fixed function unit for tessellation, while nvidia does it on the SPs. Nothing wrong with that if it's still faster. Granted we don't know yet how it will perform, but I would say with a lot of certainty that it will be faster/better than RV870.

You can't have a slightest idea if tesselation is going to be faster on dedicated hardware or "CUDA core"...

**Luka_Aveiro** · 10-07-2009, 01:43 PM

Originally Posted by golden_boy

You had some sense in your "petit" joke/sarcasm but these slides are showing something that doesn't appeal to the masses (the ones that truly represent big sales), that is, gaming...

cuidado com a língua

For that, there will be the "8800GT" GT300 derivative.

I expect it to be around 250-280$, faster than HD5850 and almost as fast as HD5870.

**AMDDeathstar** · 10-07-2009, 01:44 PM

Originally Posted by 003

The difference is that ATI has a fixed function unit for tessellation, while nvidia does it on the SPs. Nothing wrong with that if it's still faster. Granted we don't know yet how it will perform, but I would say with a lot of certainty that it will be faster/better than RV870.

Trying to do graphics,phycis and tessellation your going to have to take a hit somewhere

**Luka_Aveiro** · 10-07-2009, 01:48 PM

Originally Posted by AMDDeathstar

Trying to do graphics,phycis and tessellation your going to have to take a hit somewhere

Not necessarily, if you can efficiently code in order to make idle units be close to 0 at each clock

**marten_larsson** · 10-07-2009, 01:51 PM

Originally Posted by Luka_Aveiro

For that, there will be the "8800GT" GT300 derivative.

I expect it to be around 250-280$, faster than HD5850 and almost as fast as HD5870.

With what specs since you can predict performance that well...

**LordEC911** · 10-07-2009, 01:53 PM

Originally Posted by Luka_Aveiro

For that, there will be the "8800GT" GT300 derivative.

I expect it to be around 250-280$, faster than HD5850 and almost as fast as HD5870.

Which we might see, best case scenario, late Q2/Q3 '10.

**Luka_Aveiro** · 10-07-2009, 02:06 PM

Originally Posted by marten_larsson

With what specs since you can predict performance that well...

Something around 416SP, 384 bit 1536MB GDDR3, 650mhz core clock and some sugar and spice with everything nice!!!!

I am not predicting performance based on specs, but based on "strategy", where it should be placed at, imo. It's just an opinion.

Originally Posted by LordEC911

Which we might see, best case scenario, late Q2/Q3 '10.

I would bet for late Q1/Q2 if "full" GT300 launches indeed at November, but, hey, I don't know a thing.

Thread: The official GT300/Fermi Thread

Thread Tools

Search Thread

Rate This Thread

Display

Bookmarks

Bookmarks

Posting Permissions