No matter what it is you won't get 100% scaling. Would be nice if that were possible though.
nvidia "D8E" High End solution, what can we expect in 2008?
- Thread starter 2900guy
- Start date
No matter what it is you won't get 100% scaling. Would be nice if that were possible though.
Well if we take a theoretical performance yes - it is 3x rv670 but as you see HD3870x2 doesn`t give 100% performance boost over single rv670
Blazkowicz
Legend
100% faster means twice as fast, 200% is three times, etc.
it'd be confusing if "50% faster" meant twice slower.
it'd be confusing if "50% faster" meant twice slower.
DegustatoR
Legend
100%+200% is three times however200% is 2 times, 300% is 3 times and 100% is one time
%ΔPerformance=100% means twice as fast. Which would pretty much mean that (P1-P0)*100/P0=100%. Percentages are generally applied to a number, value, something. So if I say so 100% * n=100*n/100=n. So 50% of something means it's half of something. 50% faster means P1 would be 150% P0. So what's all of this commotion about?
Sound_Card
Regular
Well if we take a theoretical performance yes - it is 3x rv670 but as you see HD3870x2 doesn`t give 100% performance boost over single rv670
I guess you missed my disclaimer then. I'm assuming R700 is not some tacky AFR soulution.
200% is 2 times, 300% is 3 times and 100% is one time
errr.... Thanks morgoth. Saves me some breath on this one.
Well if Fuad is right about the RV770 performance then G100 should be more than 2 times faster than RV770. So even dual-chip RV770 board won't be able to compete with it. And if that is true then i think we'll see another one of AMD's "we're slower but cheaper" things. Only this time their _dual-chip_ AFR-board will be slower than one G100 and this is gonna hurt.
But those are rumours after all
They tend to be inaccurate and even false.
And even if they are true today everything can chang in the nearest future.
So you never know... Until the product is released.
What does it mean? You are saying G100 is rumoured to be more than 2 times faster than G80?
DegustatoR
Legend
Yes.
So if it is true about G100 could you tell me why NVIDIA is launching GX2 next month when they have such a fast GPU?
DegustatoR
Legend
1. Because G100 should come quite later than GX2 and they have to beat RV670X2 with something.So if it is true about G100 could you tell me why NVIDIA is launching GX2 next month when they have such a fast GPU?
2. Because there is no hurt in G100 and GX2 coexistance on the market anyway.
OK but there is no only GX2 launching in March but GF9800GTX and GF9800GT too (GT in April according to Hardocp NVIDIA roadmap leaks
). Then what? GF9800GTX will be slower than GX2?GX2 is waste of money because of lacks of efectiveness in many situations.
OK but there is no only GX2 launching in March but GF9800GTX and GF9800GT too (GT in April according to Hardocp NVIDIA roadmap leaks
GX2 is waste of money because of lacks of efectiveness in many situations.
maybe the g100 isn't the 9800 gtx
DegustatoR
Legend
I do believe that all the GF9 line-up will use G9x chips. G94 for GF9600s (and 9500s?), G92 for GF9800s and G96/98 for anything lower then G94.OK but there is no only GX2 launching in March but GF9800GTX and GF9800GT too (GT in April according to Hardocp NVIDIA roadmap leaks
GX2 is waste of money because of lacks of efectiveness in many situations.
G100 could be a start of GF10 line-up already.
Yeah, NV's naming scheme is a mess right now and it'll get worse before it get better.
Dual chip of what? G90 is already the dual version of G80.
There's no G90; you obviously mean G92 and that would be two GPUs clustered together for that 9800GX2 thingy or whatever it's going to be called. What Vincent probably means is two cores on a single-die package. It's an interesting theory nonetheless, but at this point I have a hard time figuring out why they would need to go that route that early.