AMD: Speculation, Rumors, and Discussion (Archive)

[Razor1]

Well, I'm and idiot. Then again so is AMD.

This was supposed to be the year they made a turnaround. The stock market has been betting on it. How the fuck is this a turnaround? Dual GPU for a low end card, not a mid range, but a low end thing? It's mind blowingly stupid for any company, especially one that was supposed to be on the right track.

Same thing they did with the rv680 launch.....

 

[xEx]

\And of course the stream went offline when I was about to heard about zen...

 

[SimBy]

Lisa said Polaris RX ranges from $100 to $300. Whats above $199 RX480 then?

 

[Kaotik]

And there's the dual GPU.

It's not dual GPU, but 2x RX 480 aka CF (at least until confirmed otherwise)

 

[kalelovil]

Lisa said Polaris RX ranges from $100 to $300. Whats above $199 RX480 then?

Pessimistically, she could just mean say 8GB stock models could be $229 and most expensive OC AIB models could reach $300.
Or Fury X is getting a rebrand and a massive price cut.

 

[Orion]

Nah. Just means they and the devs have more work to do with multi-GPU performance improvements.

I would guess they are seeing maybe 60-70% CF scaling, which would put a single Rx480 ~15-25% less than GTX1070.

@ what resolution or settings?
That 51% utilization is probably the limit of the Ashe's engine to distribute load on multiple GPUs, I bet 2 GTX 1080s would post roughly the same utilization rate.

http://www.hardwarezone.com.sg/feature-preview-nvidia-geforce-gtx-1080-sli-benchmarked

Maybe I'm misinterpreting utilization rate but I'd assume that 51% means around half of the hardware was being utilized. Being two chips, that means around one chip worth of hardware was utilized across both chips. So going by that one would assume a single chip at or near 100% utilization ~ 2 chips at 50% utilization ~ 1 x gtx 1080 at 98% utilization.

 

[Razor1]

Well easy way to see this what is the utilization rate of current AMD 2 card setups in AOTS?

 

[Kaotik]

They have 100% increase from base line there, where does that fit in with your 70% uplift from process?

The whole thing on the right doesn't' fit into the 1.7x at all. Are they constituting technology and baseline as finfet? Architecture and features on top of that? doesn't make much sense.

Anycase AMD technologies has a super scripted 1 next to it, its going to be specific applications.

100 % Baseline is 28nm, 170 % is 14nm, 280 % is 14nm+architecture tweaks

 

[Kaotik]

Well, I'm and idiot. Then again so is AMD.

This was supposed to be the year they made a turnaround. The stock market has been betting on it. How the fuck is this a turnaround? Dual GPU for a low end card, not a mid range, but a low end thing? It's mind blowingly stupid for any company, especially one that was supposed to be on the right track.

Again, AMD didn't say anything about dual GPU, they showed 2x RX 480 performance, not "RX 480 x2" or something similar

 

[Anarchist4000]

Lisa said Polaris RX ranges from $100 to $300. Whats above $199 RX480 then?

Well the $100 part is easily Polaris 11, although it wasn't mentioned. As for $300 definitely better cooling and stock memory seems likely enough.

It's not dual GPU, but 2x RX 480 aka CF (at least until confirmed otherwise)

Thought they said "dual" during the presentation, but may just be semantics. Powercolor showing a AIO water cooler for dual does seem to suggest two GPUs on one board though.

 

[Razor1]

100 % Baseline is 28nm, 170 % is 14nm, 280 % is 14nm+architecture tweaks

Ah ok that makes sense, I was thinking it was increases, from 0 onward.

yeah but that x2.8 is only with AMD technologies software.

And they no longer are going to hit the 2.0 x perf.watt that 14nm finfets should be giving as a max.

 

[Esrever]

Wow the PCB is actually like as short as the Nano. There will probably be very small custom cards.

 

[SimBy]

Wow the PCB is actually like as short as the Nano. There will probably be very small custom cards.

Where did you see a Nano sized PCB? All I saw was standard sized card with a blower.

 

[lanek]

oups, sorry for the size, i have just copy paste it.
VRM in front. As i have write it upper in the thread, the pcb is way shorter than the cooler.

 

[LordEC911]

yes and it will in specific applications, that is what is noted on the slide, AMDtechnologies which I just linked you to what those were.

What happened to the 60% decrease in power usage with FinFets, that is the same as 2.0 perf/watt increase. Its no longer that.

Since you are so adamant about the Nvidia 10x perf BS. They made up their own performance metric and used that to get to their magical 10x number... which their "new" architecture doesn't even support across all products.

1.7x from FinFet means that they choose a balanced approach to their design, taking into account frequency, density, and power savings.

And no, it is no longer a 2x perf/w increase from 14FinFet. It is upto a 2.8x perf/w increase from 14FinFet and the Polaris architecture.

 

[Unknown Soldier]

Where did you see a Nano sized PCB? All I saw was standard sized card with a blower.

Only one GPU with a blower

That's what I thought was mentioned too.

 

[SimBy]

VRM in front.

Thanks, not quite Nano though. Nano PCB ends exactly where PCI-Express slot ends.

Spoiler

 

[firstminion]

Ah ok that makes sense, I was thinking it was increases, from 0 onward.

yeah but that x2.8 is only with AMD technologies software.

And they no longer are going to hit the 2.0 x perf.watt that 14nm finfets should be giving as a max.

Those "up to 2x" from process alone are always with iso frequency and that's not the case.

 

[lanek]

Thanks, not quite Nano though. Nano PCB ends exactly where PCI-Express slot ends.

Effectively, there's a little difference 2cm maybe )

 

[Anarchist4000]

For the purposes of watercooling it will be close to a nano once that fan disappears.