The basis for Beema and Mullins is the Puma core. This is a highly optimized revision of Jaguar, and in some ways can be considered a new design. All of the basics in terms of execution units, caches, and memory controllers are the same. What AMD has done is go through the design with a fine toothed comb and make it far more efficient per clock than what we have seen previously. This is still a 28 nm part, but the extra attention and love lavished upon it by AMD has resulted in a much more efficient system architecture for the CPU and GPU portions.
The parts will be offered in two and four core configurations. Beema will span from 10W to 25W configurations. Mullins will go all the way down to “2W SDP”. SDP essentially means that while the chip can be theoretically rated higher, it will rarely go above that 2W envelope in the vast majority of situations. These chips are expected to be around 2X more efficient per clock than the previous Jaguar based products. This means that at similar clock speeds, Beema and Mullins will pull far less power than that previous gen. It should also allow some higher clockspeeds at the top end 25W area
AMD hasn’t disclosed how much the underlying architecture has changed, and I would guess the Puma cores are actually quite similar to Jaguar cores, but the net result is a 2X improvement in performance per Watt according to AMD. They arrive at that number by dividing the performance in a few common benchmarks by the rated TDP of the APUs. Now that’s a bit contrived, as a 25W TDP APU may not actually be drawing 25W during the tests, but we’ll just ignore the marketing for now and focus on the important metrics.
Looks like it's what Kabini/Temash should have been. If they do hit those targets, they might actually have something that's competitive with Airmont.
Hopefully we actually get some form of turbo on this, and not just on one or two SKUs.
Homeles,
You really think AMD can simply break the laws of physics? On a 28nm HKMG process, AMD is at a significant disadvantage to a chip built on 14nm (2nd gen ) FinFETs.
Unless you think Intel's architects are dumb as dirt and that AMD's are godly, I don't see how AMD could be competitive in perf/watt against Airmont.
Hey, I didn't mention a metric. Maybe they can compete on price.Homeles,
You really think AMD can simply break the laws of physics? On a 28nm HKMG process, AMD is at a significant disadvantage to a chip built on 14nm (2nd gen ) FinFETs.
Unless you think Intel's architects are dumb as dirt and that AMD's are godly, I don't see how AMD could be competitive in perf/watt against Airmont.
IF the 2x perf/watt is true, then maybe they can compete - at least for high resolution platforms (since AMD GFX will be better - Airmont is still IMG Tech, AFAIK). AMD desperately needs some design wins in the space.
Are these coming from TSMC or GFL?
http://images.anandtech.com/doci/7514/slide-07.jpg
AMD likes to imitate Intel. Now they have SDP as well.
Cherry Trail will have Intel Gen 8 Graphics, same as Broadwell (Bay Trail is based on Ivy Bridge's Gen 7 Graphics).
I believe the rumor is that the smartphone version of silvermont will be IMG graphics. All the other versions are intel's graphics.
Homeles,
You really think AMD can simply break the laws of physics? On a 28nm HKMG process, AMD is at a significant disadvantage to a chip built on 14nm (2nd gen ) FinFETs.
Unless you think Intel's architects are dumb as dirt and that AMD's are godly, I don't see how AMD could be competitive in perf/watt against Airmont.
BayTrail is Intels Gen7 GFX.
please dont derail this thread with talk of airmont! try to stay on or close to the topic!
Sorry, talking about the primary competitor of a product in this thread is not derailing. It is common sense. Beema will be competing against silvermont for tablet design wins. Unless you just want everyone here to praise marketing slides nonstop (don't think so), it is going to be discussed because it is relevant to the discussion, period.
Oh, from what I read (admittedly in a forum discussion), Bay Trail is IMG Tech, Intel's GFX still use too much power (with reasonable performance) for a mobile SoC. I don't know if that's only specific versions only (i.e. ultra mobile). Maybe @ 14nm, Intel will get good enough perf/watt to use their own tech.