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

Was kind of thinking, something based on Tahiti with some updates and improvements might be what they did, rather than something completely new.
 
Tahiti is old, power hungry and too big to be reused.
290 = Hawaii
280 = ???
270 = Bonaire?

I think that the 280 series will also feature a new chip.
 
Wynix said:
Tahiti is old, power hungry and too big to be reused.
290 = Hawaii
280 = ???
270 = Bonaire?

I think that the 280 series will also feature a new chip.
Click to expand...

Unless Hawaii is just a stop-gap to keep us entertained until the 20nm-based Pirate Islands family is released.

Since 20nm shouldn't be too far away, and since we've been hearing about PI for a while, I'd say it's possible. AMD could use a 28nm chip to work out the kinks in GCN 2.0 as well, so Hawaii could serve that purpose too.
 
Tahiti will indeed be reused, as is, for the time being.

I'm going to go out on a limb here and say the new numbering system is convoluted on purpose, for now.

To suggest that they will be reusing Pitcairn too, is simply ridiculous.
 
Tahiti is old, power hungry and too big to be reused.
Click to expand...
Its old but still very powerful.
Its big but if they are going bigger again & with its age/maturity of the node the yields must be pretty excellent by now.
It is pretty power hungry but maybe with a new stepping &/or little down-clock they can get say 90% of the performance on 75% of the power (entirely made up numbers).

A 7970 performance card at the price range of a 7850 would be a pretty bloody impressive bit of kit for the money conscious.
 
pjbliverpool said:
I'm not sure if VI is supposed to be HSA compliant, I recall seeing a slide a while back, which then referred to the Southern Islands successor as Sea Islands, which referenced the new high end discrete GPU's having HSA compliance.

But if that's true then I'm struggling to understand what it means. Would we be talking about a unified virtual memory space that addresses both system and graphics memory?

So for example the GPU could read/write directly to the system memory using the virtual address space and the CPU could read/write directly the graphics memory using the same with these reads/writes going over the PCI-E interface and thus limited to it's bandwidth or the bandwidth of the physical pool of memory if slower?

And so thus would it be valid under such a setup to consider available graphics memory bandwidth as say 200GB/s (local) + 32GB/s (PCI-e - assuming system memory was also 32GB/s or more of course)?

And vice versa for the CPU?

Would memory capacity also be valid to look at in this way? i.e. the GPU would have say 16GB of memory running at 32GB/s + 3GB running at 200GB/s?
Click to expand...

I am doubtful it will have HSA features over PCIe.

HSA seems to be for single die chips only.
 
rpg.314 said:
I am doubtful it will have HSA features over PCIe.

HSA seems to be for single die chips only.
Click to expand...


Depend the functions.. i.e, you dont want base your features on a sololy CPU / memory access compliants, if then the gpu is coupled with a cpu who cant support it... zero benefit.. or zero use ofc.. now other HSA features are not forcibly hardware based on the system/cpu side.

Somwhere they need to bring some ability and use of HSA with their GPU's and outside ARM based system or their low end SOC / APU, or it will not go up. ( not sure, but the A57 is the first to be HSA oriented ? ).

Now on professionnal computing and HPC, thats another story.. I tend to believe HSA will have a bigger impact on SOC and Professional system before "standard " PC consumer systems.. ( because on both case who are on the extreme ), you work on closed configuration system, and not on the "PC consumer" level, who is a lot more open on the hardware configuration .

HSA is just born, its a bit too early for imagine it show some real benefits anyway in 2013.
 
Last edited by a moderator:
Is there much known about potential XFire improvements? such as things they cannot add to the 7000 series.

Also I really hope the 7800 series replacement does not use tahiti, that HIS leak of the 280X has been giving me nightmares(3GB,384Bit, probably tahiti).

The reveal is the 25th right? this reveal is kind of like christmas, getting hyped only to be disappointed. :-|
 
If we are really pessimistic about this HIS leak R9-280x could be haiti yes - also because the APUs goes up to A10. On the other hand the 'x' would look weird then, if its not the top part.
And the A7-250 looks a lot like bonaire - at least a 128 bit chip due to those memory configurations, and I doubt it would make sense to replace bonaire as the highest-128bit offering already.
So maybe these products are only the already existing parts of the lineup.
 
ninelven said:
I think what they are referring to is this:



http://techreport.com/news/24737/amd-sheds-light-on-kaveri-uniform-memory-architecture
Click to expand...

Thanks for the link. It's not overly clear to me but it does seem to suggest that what I was saying earlier may be true. i.e. that both CPU and GOU would be able to read/wrte data from each others local memory via the virtual address space, albeit at different speeds depending on which bus your using - not too dissimilar to how PS4 works in that regard I guess?
 
Psycho said:
If we are really pessimistic about this HIS leak R9-280x could be haiti yes - also because the APUs goes up to A10. On the other hand the 'x' would look weird then, if its not the top part.
And the A7-250 looks a lot like bonaire - at least a 128 bit chip due to those memory configurations, and I doubt it would make sense to replace bonaire as the highest-128bit offering already.
So maybe these products are only the already existing parts of the lineup.
Click to expand...
Eh? have you not seen this yet?
Added preliminary support for AMD Radeon R7 240, R7 250, R7 260X, R9 270, R9 270X, R9 290, R9 290X
Click to expand...
 
So a bit under 30% larger than GK110 (551 mm^2 or 561 mm^2 depending on where I look) would be around 425-435 mm^2 for Hawaii.

So 2560 SPs, maybe even 2816 SPs, doesn't seem implausible.
 
Hmm, this quote sounds worrying to me:

“It’s also extremely efficient. [Nvidia's Kepler] GK110 is nearly 30% bigger from a die size point of view. We believe we have the best performance for the die size for the enthusiast GPU"

The focus on best performance for die size while also highlighting the 30% die size advantage may suggest that it doesn't have the overall performance advantage. A shame if true.
 
You must log in or register to reply here.

Similar threads

I
AMD: Pirate Islands (R* 3** series) Speculation/Rumor Thread
140 141 142
Replies
3K
Views
616K
kalelovil
K
S
AMD: Sea Islands R1100 (8*** series) Speculation/ Rumour Thread
69 70 71
Replies
1K
Views
390K
AnarchX
A
U
AMD: Southern Islands (7*** series) Speculation/ Rumour Thread
236 237 238
Replies
5K
Views
1M
BrynS
B
Back
Top