Wednesday, May 23rd 2012
Production of AMD "Piledriver" FX Processors Begin Q3 2012
Production of AMD's next-generation FX processor family, which are based on its "Piledriver" microarchitecture, will commence in Q3 2012, according to industry sources. Some of the first client processor models based on the "Vishera" silicon, will be the eight-core FX-8350, six-core FX-6300, and quad-core FX-4320. The three model names were earlier misinterpreted with an "x" prefix from a roadmap slide.
A few more details are known about these chips. For starters, the chips will be built on the existing AM3+ package, retaining compatibility with current AM3+ platforms. The chips will also retain dual-channel DDR3-1866 MHz integrated memory controllers, and Turbo Core 2.0. The main differences here, are increases in IPC (performance to clock-speed ratio), and the implementation of resonant clock mesh technology, which increases energy efficiency.
Source:
DonanimHaber
A few more details are known about these chips. For starters, the chips will be built on the existing AM3+ package, retaining compatibility with current AM3+ platforms. The chips will also retain dual-channel DDR3-1866 MHz integrated memory controllers, and Turbo Core 2.0. The main differences here, are increases in IPC (performance to clock-speed ratio), and the implementation of resonant clock mesh technology, which increases energy efficiency.
63 Comments on Production of AMD "Piledriver" FX Processors Begin Q3 2012
terpconnect.umd.edu/~kosborn/SEALeR/Papaefthymiou.pdf
^-- stuff about resonant clock meshes
I'll be watching this one.
Keep an eye out for Steamroller, but don't hold your breath.
ewh.ieee.org/r5/denver/sscs/Presentations/2012_01_Doyle.pdf <-- It hasn't been talked about even though this design is already in Orochi B2, so I am not sure if Orochi Revision C(x) would get the 8 Gbit/s interconnect.
This is is going to be a problem for overclockers. Because of the resonant mesh design, efficiency is maximum at the resonant frequency, and it decreases when frequency increases.
edit:
There are better CMOS energy recovery schemes out there. One example is the adiabatic logic. Most chip companies will be forced to use energy recovery circuits when CMOS scaling reaches its limit.
If the Bulldozer had the performance it currently had but would be a lot more power efficient it would suddenly be appealing to me.
www.youtube.com/watch?v=c9EBhaULToU
But read over the old Bulldozer reviews from Dec 2011 there was a few multi threaded benchmarks which were faster than Nehalem and Sandy Bridge.
The problem with Bulldozer is it's leads were not consistant enough to stand out. But I hear you. Performance was disappointing.
Given that SB is up to 40% faster clock for clock than Phenom II, combine that with the fact that Bulldozer is up to 15% slower clock for clock than a Phenom II, you can see there is not a chance in hell of these Piledriver CPU's even matching SandyBridge, let alone IvyBridge.
So, myeah...
I reserve my opinion until I see the Piledriver FX chips. Glad to see they are going to stick with AM3+ for at least another generation.
Hopefully they have dealt with some of the heat and power issues. IPC is important but you can't really think there will win there for Piledriver.
Athlon II X4, just as fast in singethreaded and multi threaded games
Phenom II X6 in multi-threaded games, just as fast in singlethreaded.
Bulldozer in multi-threaded games, just as fast in singlethreaded.
LLano APU faster in games period.
Trinity APU faster in games period.Bulldozer is not 15% slower than Phenom II clock for clock . Please stop spreading misinformation.
Secondly, Anantech already says that the LOW END Trinity APU based on Piledriver is 20-25% slower than Ivy Bridge. So I would assume the HIGH END Piledriver desktop CPU would be even faster than Trinity. My conservative guess would be a 10% improvement from from a low end APU to a high end CPU. So based on that information I would say Piledriver should* be as fast as Sandybridge +/- a few % depending on the benchmark.www.anandtech.com/show/5831/amd-trinity-review-a10-4600m-a-new-hope/9
The benchmarks are out there if you look....
www.bit-tech.net/hardware/2009/08/17/amd-phenom-ii-x4-965-black-edition-review/6
EDIT: Also, to me, additional modules/cores are pure gold when recording gameplay videos. Encoding with the high compression rates, that is (the processing-intensive, not the quality-reducing type). While the game still runs as smooth as silk. YYY!
Well, there will always be fanboys everywhere, and aggressive weeding is needed every so often.
Either way, AMD needs to boost their singlethreaded heavily to compete with Sandy Bridge past the $150 dollar mark, but I think AMD has given up in that segment and focus their resources in server and also low to mid end. Bulldozer while lacking in high end gaming does very well in mid and below, trading blows with the i3 and non-k i5 when overclocked. Power consumption still needs to be improved by 50% to make it truly competitive, no just on benches.