Die Size: 19% larger
Memory Bandwidth: 37.5%
Memory Size: 50%
Frequency: (9%)
So my question is: What is holding GCN, specifically Tahiti, back?
Z/stencil rate?
Die Size: 19% larger
Memory Bandwidth: 37.5%
Memory Size: 50%
Frequency: (9%)
So my question is: What is holding GCN, specifically Tahiti, back?
So my question is: What is holding GCN, specifically Tahiti, back? As others have noted it seems Pitcairn at 212mm^2 scaled up to Tahiti size would, in theory, be faster. So is Tahiti not "balanced" correctly?
That is indeed one aspect AMD hasn't touched for quite some time. Nvidia has had twice the Z/stencil rate per ROP since G80, and AMD stopped increasing it after RV770 for some reason. It's a bit strange, seems like a relatively low-hanging fruit to me.Z/stencil rate?
That is indeed one aspect AMD hasn't touched for quite some time. Nvidia has had twice the Z/stencil rate per ROP since G80, and AMD stopped increasing it after RV770 for some reason. It's a bit strange, seems like a relatively low-hanging fruit to me.
Surprised AMD haven't budged on pricing yet, I guess they are happy having the more expensive and often slower GPU compared to Nvidia.
Could be enough, if they can add a Boost on top of that. This could be doable with current hardware, but with new Bios. (or not?)
how does it go at bit coin mining they could be selling out just from those crazy nutters.
My Sapphire 7970 OC just arrived today; it does 1GHz using less than factory voltage along with 1,450Mhz memory speed. Most people see over 1200MHz on this card, so I don't see why they'd need boost -- they just need more base clock.
You do realize that Tahitis have at least 4 different "factory voltages" (IIRC)
nVidia has 1:8 color:depth in ROPs nowadays, AMD has 1:4, if my memory serves me right
AMD could have had something like this ready for the Kepler launch.
With any luck we'll get an improved cooling solution, improved components, and ~1.2GHz stock clock.
Or maybe they'll try to lower voltage to improve perf/watt vs. Kepler.
Anyway, got a link?
Kepler has a slight edge as it is clocked higher at 1006MHz plus Boost verus 925MHz for GCN.
The reason was posted here: early wafers indicated 28nm was going to be more troublesome than it has turned out.This chip was clocked low for a reason and long term stability is likely one of them.
So Leo is the new power virus?A lot of people have been having problems with the Leo demo with their overclocks even though they are ok for games.
The reason was posted here: early wafers indicated 28nm was going to be more troublesome than it has turned out.
But 1.2 GHz seems like wishful thinking in the near term.
So Leo is the new power virus?
Could be. The irony. Vrm are reportedly get crazy hot and cards with some people run 100mhz lower with Leo compared to what they thought was stable before.
So.. erm.. any speculation as to when will AMD lower the MSRP of the HD7900s?
To me the real problem with pricing is in the mid end, 78xx are pretty expansive and the new hd 77xx are not that compelling. Nvidia needs to put some pressure on AMD in the 200$ /250 range.Well, there has been almost no supply of 680 to put any pressure on.
Starting to wonder if AMD cant get away with 499 7970 long term instead of 449 though. For these reasons:
-Not always slower than 680
-Better compute/bitcoin
-3GB Ram/better at high res/more bandwidth
-Traditional overclocking ability and seemingly higher overclocking ceiling
All the above would just be strengthened if indeed AMD kicks default clocks up to 1GHZ as rumored
Surprised AMD haven't budged on pricing yet, I guess they are happy having the more expensive and often slower GPU compared to Nvidia.