Nvidia GT300 core: Speculation

Status
Not open for further replies.
Arty said:
So I guess its sooner than later.
Click to expand...
That depends on how optimistic you feel. But the number 3 is wrong in any case.

Domell said:
I think NVIDIA needs about 20-30% faster GPU to be successful.
Click to expand...
Not really. Since both vendors will have features parity this time (DX11) it comes down to what they have beyond that. And it's pretty clear that NV has better Compute infrastructure and PhysX. So even in the case of being in the same performance league (which I myself tend to consider one of less possible variants for G300 vs Cypress situation) NV will have some advantage in features.

neliz said:
There hasn't been a noteable quantity of any GT21x part, you think they finally fixed the leakage and launch a top to bottom DX10.1 lineup before the Win7 launch?
It's been about 7 months now since GT218 and GT216-A2 had been approved for production. where are they?
Click to expand...
I don't think that you may make any assumptions about G300 family from the GT21x line launch woes.
And for what it's worth they ARE avialable, even more so than RV740.
Plus I still don't really believe that Win7 launch is that important. DX11 launch is what's important but for some time there won't be a lot of titles to use it and we've yet to see HOW the first titles will use it.
In any way right now I think that the main problem for NV will be Juniper because they'll probably have G300 against Cypress and Hemlock but Juniper -- well, they simply won't have anything to counter it in any sane way.
 
Last edited by a moderator:
DegustatoR said:
That depends on how optimistic you feel. But the number 3 is wrong in any case.
Click to expand...
If you have a fixed date already, congratulations you're one of the very very few.


I don't think that you may make any assumptions about G300 family from the GT21x line launch woes.
And for what it's worth they ARE avialable, even more so than RV740.
Click to expand...
I can buy a dozen RV740 based cards right out my front door. it's a lot less for GT216/18
Plus I still don't really believe that Win7 launch is that important. DX11 launch is what's important but for some time there won't be a lot of titles to use it and we've yet to see HOW the first titles will use it.
Click to expand...

Can't compare Vista with Win7
In any way right now I think that the main problem for NV will be Juniper because they'll probably have G300 against Cypress and Hemlock but Juniper -- well, they simply won't have anything to counter it in any sane way.
Click to expand...

GT212 will go againts HD5850, and Juniper. GT215 against Redwood. that leaves a lot of bottom feeding for GT216/218
 
Jawed said:
Wow, NVidia's pitting a 40nm refresh of GT200 against Juniper? Whoops.

Jawed
Click to expand...

Would it really be that bad? With good process scaling and a 256-bit bus with GDDR5, they'd have a roughly 260mm² chip with the same bandwidth as the GTX 285.

Factor in the power savings made possible by the 40nm process and we could have shaders clocked well over 1.5GHz.

If Juniper is indeed about as fast as the HD 4870, given its 180mm² die size, I'd say Nvidia wouldn't be in such a bad position. Of course they'd still lack DX11, but they'd have... wait for it... PhysX :D

The biggest problem is that even the HD 5850 will be running circles around that, not to mention the 5870, let alone the 5870X2...
 
With a spec of 384SPs & 96TMUs, I think it'll atleast match or beat the 5850. The one going against Juniper would most likely be the salvage parts.
 
Chalnoth said:
Well, it's not really about the number of games either, but whether or not those games that do are ones that people like.

In any case, yes, I understand the advantage of vendor-agnostic API's. I'm just trying to put it out there that nVidia isn't completely crazy for pushing PhysX, from a business perspective. Not that that's necessarily good for the consumer.
Click to expand...

I'll be the first in line to applaud companies who aggressively support their products as it's the right way to do business. But it's all in the delivery. They could demonstrate the advantages of PhysX without saying nonsense like RV870 is slower because it doesn't support it. That just makes them seem completely out of touch with reality, and certainly can't make shareholders very confident in their upcoming hardware.
 
trinibwoy said:
I'll be the first in line to applaud companies who aggressively support their products as it's the right way to do business. But it's all in the delivery. They could demonstrate the advantages of PhysX without saying nonsense like RV870 is slower because it doesn't support it. That just makes them seem completely out of touch with reality, and certainly can't make shareholders very confident in their upcoming hardware.
Click to expand...
Do they actually say that?
 
CarstenS said:
Not more than me having to tell you, that it's equally pointless to try and qualify bandwidth from quantification in an otherwise as good as unknown arch. :)
Click to expand...

True. But keep in mind that I had the relevant part of PCGH's interview with Dally in the back of my head when I replied.
 
Ailuros said:
True. But keep in mind that I had the relevant part of PCGH's interview with Dally in the back of my head when I replied.
Click to expand...

Well he could have been referring to some grand design he has in his head for 5 years down the road from now. But if he was referring to GT300 being bandwidth limited even on a 512-bit GDDR5 bus well then.....hummmm :)
 
trinibwoy said:
Well he could have been referring to some grand design he has in his head for 5 years down the road from now. But if he was referring to GT300 being bandwidth limited even on a 512-bit GDDR5 bus well then.....hummmm :)
Click to expand...

Well he obviously is foreseeing that they'll need much more bandwidth for the future. I didn't say or imply that D12U will be bandwidth limited; I just expect other factors to see a bigger increase than bandwidth itself.
 
D11 is for the GT21x stuff, unless of course you're implying some weird codename twists here.
 
neliz said:
If you have a fixed date already, congratulations you're one of the very very few.
Click to expand...
I have several numbers and there are no 3 there.

neliz said:
I can buy a dozen RV740 based cards right out my front door. it's a lot less for GT216/18
Click to expand...
Less of what? You can buy thousands of them for your notebooks if you're an OEM.

neliz said:
Can't compare Vista with Win7
Click to expand...
Not much to compare. The difference is smaller than between 95 and 98.

neliz said:
GT212 will go againts HD5850, and Juniper. GT215 against Redwood. that leaves a lot of bottom feeding for GT216/218
Click to expand...
As far as I know there is no GT212. There never was any samples of GT212. That chip never made it to tape out. Can you provide us with any reason why they would ressurect a year old chip instead of working on a 300-series middle-end GPU?
I have doubts about Redwood being able to compete with GT215 since GT215 should end up being pretty close to G92b/4850. And don't even start with DX11 -- performance is what matters most.
GT218 and GT216 will have their Cedar.

neliz said:
The Hiatus wasn't indefinite. I'm guessing around Computex time.
Click to expand...
That's one of the stupidest things i've heard for the last year or so. NV's late tactics was strange but if that turns out to be true then it'll go beyond strange right into the crazy territory.

Jawed said:
Wow, NVidia's pitting a 40nm refresh of GT200 against Juniper? Whoops.
Click to expand...
GT212 shoud've been around Cypress performance territory. That's another reason to consider that GT212 ressurection rumour highly unlikely.

Alexko said:
Isn't GT212 just a "dumb" GT200 shrink?
Click to expand...
You can't make a "dumb shrink" from 65 to 40nms. And you can't have 512-bit bus on a 1.4 bln transistors chip on 40nms.
 
Status
Not open for further replies.

Similar threads

G
After Effects Rendering PC PrePurchase
Replies
3
Views
2K
Davros
D
Back
Top