AMD: Speculation, Rumors, and Discussion (Archive)

Status
Not open for further replies.
M400 series' specs are already in the website.
Those do seem like Tonga and Bonaire rebrands.
 
Ok does anyone know if AMD developed their own HBM Controller?
NW Logic's controller goes back before Fiji-Fury release.
But yeah it is going to be a pig to fully prove with evidence available on the internet.

Cheers
i can't dig around right now, since i'm on a phone in another country, but IIRC it has been stated that AMD was responsible for the physical memory controller design when they co-developed HBM with Hynix. That in mind, why on earth would they use 3rd party memory controller in Fiji?
 
Yeah your probably right.
I am probably conflating this due to Northwest Logic working with SK Hynix and were one of the very first with an HBM controller completed; they had their controller ready back in Q1 of 2014 and also proven using Hynix modules (way before they were available to market).

Cheers
 
Anandtech confirms they're all rebrands:

http://www.anandtech.com/show/10322/amd-unveils-radeon-m400-series-for-notebooks


Since AMD has placed Tonga as Radeon M485X and Bonaire as M470/X, I imagine Polaris 11 will be Radeon M480/X and Polaris 10 will be Radeon M490/X.

Also, they seem to confirm (or guess, maybe?) that Topaz is, in fact, a GCN3 GPU. Oland is a GPU from early 2013, so if this is right then Topaz really is a new (and rather obscure) GPU.


On a side note, it'll be really weird if the M470 consumes substantially more power while performing substantially worse than a M480.
 
Last edited by a moderator:
Yeah your probably right.
I am probably conflating this due to Northwest Logic working with SK Hynix and were one of the very first with an HBM controller completed; they had their controller ready back in Q1 of 2014 and also proven using Hynix modules (way before they were available to market).

Cheers
Still, AMD has been involved with HBM design since it began, and it's not the first memory standard they (co)developed. I see no reason why AMD would use 3rd party memory controller in anything
 
Let me get this straight. AMD announces this thing on twitter on their main account https://mobile.twitter.com/amd/status/731549678341398528 to then announce on reddit that they are not going to show/disclose anything substantial?!?

Who is running their communication department?!
A Magic 8 ball :)
It is incredible how AMD can employ such great and amazing engineers, and yet screw up in other departments (not just marketing).
Cheers
 
Last edited:
Let me get this straight. AMD announces this thing on twitter on their main account https://mobile.twitter.com/amd/status/731549678341398528 to then announce on reddit that they are not going to show/disclose anything substantial?!?

Who is running their communication department?!
I wanted to say the same thing.
If this event was all along a professional meeting to brief their channel and partners, so why do you announce it publicly ?
They've just created hope, hype and buzz for nothing. Even worst, now people is frustrated.
I have no idea how such big company can be so incompetent in marketing, it's a disaster, the kind of example you can teach in school in the category "not to do"
 
Are there compute efficiency issues with GCN?
GCN would definitely need more VGPRs per CU. Nvidia doubled VGPRs per thread in Pascal, and they have less VGPR pressure.

Is is hard to write complex shaders with good occupancy on GCN. I mostly blame the compiler (it's does aggressive load->use latency hiding, wasting lots of VGPRs). Hardware with more registers would make the software much less of a problem. It doesn't help that GCN has the widest waves of modern GPUs and executes instructions in 4 cycles (IPC per thread is low -> registers are allocated for long time). GCN scalar unit is great, and hopefully Polaris extends it further (so that the compiler can better offload coherent calculations and coherent data to it).
 
Status
Not open for further replies.
Back
Top