AMD: Volcanic Islands R1100/1200 (8***/9*** series) Speculation/ Rumour Thread

Each L2 slice can service an L1 cache line request per clock, and Hawaii has 16 slices. Each request would be for a 64 byte cache line.
 
According to quite respected guy on AMD business here in Finland, 20nm will be only used for ARM-based products
As a side note, AMD announced in mid 14 the first wave of Project Skybridge, which consists of a 20nm A57 and 20nm Cat APU...
 
Has anyone ever seen something useful about Iceland / Topaz? I'd actually be interested in how this thing compares to Mars / Oland. At least I think it might already be shipping, though it's difficult to find some definitive answer thanks to the magic wonders of marketing (should I call that tactic reverse renaming maybe?) which would make all product names identical to these using Oland.
(It does appear though now in the open-source amdgpu linux driver which supports all 3 VI chips - Tonga, Iceland/Topaz, Carrizo but that doesn't really tell you much more - but yes it really is a 6 CU, 8 rop, 128bit chip.)
 
https://www.techpowerup.com/211973/...-to-14-nm-with-arctic-islands-gpu-family.html

AMD's next-generation GPU family, which it plans to launch some time in 2016, codenamed "Arctic Islands," will see the company skip the 20 nanometer silicon fab process from 28 nm, and jump straight to 14 nm FinFET. Whether the company will stick with TSMC, which is seeing crippling hurdles to implement its 20 nm node for GPU vendors; or hire a new fab, remains to be seen. Intel and Samsung are currently the only fabs with 14 nm nodes that have attained production capacity. Intel is manufacturing its Core "Broadwell" CPUs, while Samsung is manufacturing its Exynos 7 (refresh) SoCs. Intel's joint-venture with Micron Technology, IMFlash, is manufacturing NAND flash chips on 14 nm.

Named after islands in the Arctic circle, and a possible hint at the low TDP of the chips, benefiting from 14 nm, "Arctic Islands" will be led by "Greenland," a large GPU that will implement the company's most advanced stream processor design, and implement HBM2 memory, which offers 57% higher memory bandwidth at just 48% the power consumption of GDDR5. Korean memory manufacturer SK Hynix is ready with its HBM2 chip designs.

So I guess stuff this year is 28nm
 
Weren't all the 20nm processes deemed unsuitable for GPUs regardless of foundry?

The thing is other fundries was not really on this type of chip.. so their process was more or less only for SOC.

Well, lets be honest, if any 20nm part will have been made, we will allready got them at this point.

Anyway, 2016 could be really a fun year with a jump to 14-16nm.
 
Interesting... So which GPUs exactly support this feature - Tonga, yes? (As you hint yourself.) Any others? Radeon Fury series? Others...? :)
 
It showed up in a slide for Fury as well.

Carrizo's GPU shares their ISA, and has some features outlined that would help, but I have not seen that bullet point for the APU, specifically.
 
Huum, this said, you will need an Firepro S series for enable the hardware multi-user, no ?
 
Last edited:
There is an existing "Radeon Sky" series, already.
Not sure but they seem to be on Pitcairn, Tahiti and Hawaii respectively.

Multi-user Windows isn't that cheap either - though even XP Home was perfectly multi-user, but only one desktop can be shown and interacted with at a time.
I've always thought having four copies of a multiplayer game running on a single PC would be very neat. Monitors, keyboards etc. can be directly attached or on underpowered and disparate desktops, laptops and others on the network.
If running a single copy of Windows rather than four ones you even save a few gigabytes of memory.
It's not new stuff by this point, only it's always been kept away from consumers.
 
Back
Top