AMD: Sea Islands R1100 (8*** series) Speculation/ Rumour Thread
Hmm...
Actually you might be right, islands in Canada which sound the same as planets might be mobile, while Öland island in Sweden might belong to desktop parts.
More likely planets are the codenames and islands in Canada are just a coincidence
But just for the sake of it, Öland actually is in a sea for what it's worth
Of course with the wrong codename...
Why?
School geography said, that Mars Island in British Columbia, N. America, Oland Island in Sweden, Europe and Venus Island in Philippines, Asia
it is a Sea Islands
UniversalTruth
Veteran
Because my inner voice says so.
http://forum.beyond3d.com/showpost.php?p=1647556&postcount=293
http://forum.beyond3d.com/showpost.php?p=1649885&postcount=298
School geography said, that Mars Island in British Columbia, N. America, Oland Island in Sweden, Europe and Venus Island in Philippines, Asia
it is a Sea Islands
Ok, I see but regarding school I cannot agree, this is too much in details and I doubt anyone ouside these particular countries was taught about these islands.
UniversalTruth
Veteran
I see what you did there.20-30% higher framerate is always a benefit.
But I'm curious if there is anything new this time around or are we looking at a improved 7900? Will we see free 8xAA, or 16AF for example? Run CCC from a smart phone/tablet while in a game (temps, fan control, etc)? Connect your mouse directly from the card (ok, maybe too extreme). Something else...
Last edited by a moderator:
I see what you did there.
But I'm curious if there is anything new this time around or are we looking at a improved 7900? Will we see free 8xAA, or 16AF for example? Run CCC from a smart phone/tablet while in a game (temps, fan control, etc)? Connect your mouse directly from the card (ok, maybe too extreme). Something else...
There might be some new HSA features (not sure what the roadmap is) and perhaps new AA modes, since they tend to add those every now and then, but beyond that I can't see anything obvious missing from GCN, apart from NVIDIA's adaptive V-Sync, maybe. Oh, apparently AMD's idle power is really bad with multiple displays, for some reason, so improving this would be most welcome.
And of course, additional power-efficiency and performance is always nice. Their Turbo could probably stand to be more aggressive too.
8870
>Compute Performance (SP / DP) 3.94TFLOPS / 246.4GFLOPS
DP 246.4GFLOPS
So the 7870 is probably the peak of DP performance for the masses.
Considering how little masses need DP power (in reality, they don't need a single bit of it really - if you need DP power, you can afford top-of-the-line consumer GPUs, and most likely you can afford top-of-the-line professional cards), I understand why nVidia already cut from 1/12 to 1/24 (IIRC?), and this would be just 1/16.
Where nV went IMO wrong was cutting the other compute performance affecting capabilities in the process, SP compute performance is more and more important when games use it more and more (disregarding how "handcrafted" DIRT Showdown max detail Forward+ lighting may be for AMD, it's still one example of SP compute performance heavy scenario for gamers)
As far as I can tell, Adaptive V-Sync while nice is pure driver feature hence not really missing from GCN as such.
It is true that Kepler now supports low idle power even with 2 monitors (but not 3!) even if they have different timings, though actually earlier drivers didn't support this and the power consumption was quite similar to AMD cards (unless both monitors used the same timings).
I don't actually know how nvidia does it since it's sort of an interesting problem: with multiple (not synced) monitors you cannot reclock the memory without causing flicker on at least one monitor (since you can only do it in vblank interval on one monitor), so to avoid the flicker the cards run the memory at the highest clocks always. That in itself causes higher power consumption but it looks like high memory clocks also need some minimum core clock so that's how you end up with a high power consumption with multimon (amd and nvidia generally were very similar there).
Though I guess one solution would be to just use the L2 cache while reclocking for scanout.
Actually, Kepler cards seem to be doing fine even with three monitors: http://techreport.com/review/22922/nvidia-geforce-gtx-670-graphics-card/7
Perhaps they just don't bother trying to avoid flickering, which frankly always seemed like a non-issue to me, unless I understand it incorrectly. I mean, it would only flicker when the memory clock is increased or decreased, right? So basically, whenever entering or exiting a video game.
Displays tend to go black for a second at those times anyway. I'll take those flickers over 35W of extra idle power any day.
Silent_Buddha
Legend
Huum this review date from may, i will ask me if the ZeroPowercore is working now with actual driver and 3 monitors. ( I need admit i have not check, but i think remember something about ZeroPowercore in some patch notes.. )
Already been done by ATI back in the 90's.
Regards,
SB
Could anyone convince the general public that connecting the mouse to the video card instead of the motherboard wouldn't reduce overall latency in online games? I say convince not necessary dispute.
Wouldn't some say that current, single video cards aren't able to saturate the 2.0 PCIe bus anyway? Even though it's far fetched at this time, it would be something that no other video card would have.
ZeroCore wasn't working because the sound was enabled on the monitors, but ZeroCore is not the problem.
The problem is idle power when the monitors are on. It's way higher than on NVIDIA.
Doesn't seem to be working all that well here (read at all), with a meager 2 monitors and no enumerated sound...Maybe ATI doesn't like Windows 8 yet or something!
Similar threads
