AMD: Speculation, Rumors, and Discussion (Archive)

Status
Not open for further replies.
Some 36 CU GPU in the device ID hole above Hawaii (memory specs should be ignored, SiSoft has a bad detection on AMD GPUs): http://ranker.sisoftware.net/show_r...e6d7e7d1e1d3e1c7b588b89efb9ea393b5c6fbc3&l=en
Details for Result ID 67DF:C4 (2304SP 36C 1GHz, 16kB L2, 3GB DDR3 5GHz 256-bit) (OpenCL)
Mabye a stock cleaner for Hawaii or something new.

The initial Baffin PCI IDs include 0x67E0, 0x67E1, 0x67E8, 0x67E9, 0x7EB, and 0x7FF. The Ellesmere IDs are 0x67C0 and 0x67DF.
http://phoronix.com/scan.php?page=news_item&px=AMD-Open-Sourced-Polaris

So its Ellesmere / Polaris 10.

And there are even some new benchmarks: http://ranker.sisoftware.net/show_d...a09dac8ae2dfeaccb489b89efb9ea393b5c6fbc3&l=en

Now running with 800MHz / 6Gbps.
 
Last edited:
3GB DDR3 5GHz? Holy shit that's some serious overclock on DDR3!
 
There's more surprise to come... was on a meeting with AMD retail chanel yesterday ( online meeting with Roy as guest ) ( mostly about VR and next products )
 

So I was wrong in the vega thread to believe that baffin was the bigger polaris chip. They look pretty pricey then.

It's not out of the realm of possibilities that Ellesmere is the smaller one, even if it is unlikely since there's only 2 of them and ton of Baffin models
 
It's not out of the realm of possibilities that Ellesmere is the smaller one, even if it is unlikely since there's only 2 of them and ton of Baffin models

Well, AMD have confirmed that Polaris 10 is the bigger chip(AT even think same goes for the Vega chips) and the open source driver release asks them to change the earlier codenames.

drm/amdgpu: change ELM/BAF to Polaris10/Polaris11
Adjust to preferred code names. Signed-off-by: Flora Cui <Flora.Cui@amd.com> Reviewed-by: Alex Deucher <alexander.deucher@amd.com>

https://cgit.freedesktop.org/~agd5f...s&id=f3737d519277961f295f0a17e337f63542d17efb

Michael We're calling them Polaris 10 (was Ellesmere) and Polaris 11 (was Baffin) now, any chance you could tweak the article accordingly ?

Thanks !

https://www.phoronix.com/forums/for...-open-source-driver-code-for-next-gen-polaris

edit:

More info from pcgh, plus information about greenland from another AMD employee linkedin account(though some are more cunning about keeping hardware secrets).

https://translate.google.co.in/tran...llesmere-Baffin-Greenland-1190310/&edit-text=

IP tables for the new chips,

https://www.forum-3dcenter.org/vbulletin/showpost.php?p=10980597&postcount=1966
 
Last edited:
So based on the all the leaks AMD's Nextgen Finfet Lineup will (roughly) look like this:

P11/Baffin: 20 CUs 128-bit GDDR5 ~ 115 mm²
P10/Ellesmere: 40 CUs 256-bit GDDR5(X) ~ 220 mm²
V10/Greenland: 64 CUs (Gen 9 Graphics IP) (1/2 rate DP) 2048-bit HBM2 ~ 360 mm²

If Nvidia can get GP100 (or a derivative of it) into the Desktop they have this one in the bag.
 
But the same gfx IP level as Fiji is surprising.

Feature Level 12.0 confirmed!

Aren't you guys jumping the gun here a little?

The lineup I posted is based on "confirmed" leaks and reasonable speculation. AMD is not responsible for the overinflated expectations flying around in various fora, since they have yet to make any serious performance claims. Apparently everyone has already forgotten the Fiji launch. I expect to see many long faces once Polaris reviews hit the net.
 
Feature Level 12.0 confirmed!



The lineup I posted is based on "confirmed" leaks and reasonable speculation. AMD is not responsible for the overinflated expectations flying around in various fora, since they have yet to make any serious performance claims. Apparently everyone has already forgotten the Fiji launch. I expect to see many long faces once Polaris reviews hit the net.

I mean everything is possible, of course. But saying Polaris is a 'completely new architecture' with all the mentioned improvements and then delivering basically just a Fiji die shrink, goes far beyond what AMD ever pulled. All the rebrands would pale in comparison to this.

I would put just a little more faith in Raja Koduri. But again, everything is possible I guess.

 
If Polaris is just a current GPU shrunk down to 14nm then AMD will be dead to me. So I'm hoping - well, expecting really - that that isn't the case, because we need some decent competition in the dGPU marketplace. Otherwise NV will be free to do whatever they like, and we know what a bunch of shitty people they are over there. They'd abuse the hell out of their market position, you know this is true.
 
If Polaris is just a current GPU shrunk down to 14nm then AMD will be dead to me.

Polaris is certainly more than just a shrink if only by virtue of the fact that there is no such thing anymore (since 55 nm at least). The term "completely new architecture" can be stretched very far if necessary. It looks like Polaris is based on the cancelled 20nm lineup with a couple things pulled in from the Finfet Generation. Greenland/Vega10 will bring larger architectural changes later on.

What is it you don't like about the lineup I posted at #961 above? Seems reasonable to me, rather like the inital 28nm rollout.
 
With GDDR5X, that should be fine.

But the same gfx IP level as Fiji is surprising.

Earlier discussions tried to sort out the usage of each entry:

https://forum.beyond3d.com/threads/...peculation-rumors-and-discussion.56719/page-9

There seem to be some tweaks, but it's going back to a minor bump in version from Tonga for everything but multimedia. Is DCE related to compute? That seems to be inherited from Carrizo. (edit: Or is it the display controller?)
The lack of a *new* tag on things like the rasterizer and color blocks could be what is housed under the GFX label.
 
Last edited:
If Polaris is just a current GPU shrunk down to 14nm then AMD will be dead to me
If Polaris is just a shrink but it manages to perform well with improved perf/W, with just low level fixes, then why not?

GCN on Fiji is a great architecture, it's just power hungry.
 
Status
Not open for further replies.
Back
Top