AMD Tech Day at CES: 2018 Roadmap Revealed, with Ryzen APUs, Zen+ on 12nm, Vega on 7nm
by Ian Cutress on February 1, 2018 8:45 AM ESTRyzen APU Overclocking: A Focus on Memory Support
Also part of the demonstration about the new APUs was overclocking. For a number of years, AMD has hired Sami Makkinen, a well-known former world-record holder in various overclocking categories, to help tune the overclocking on some of its processors. So despite this being an integrated graphics platform, Sami was on hand to show the day-to-day overclocking potential. I was told that liquid nitrogen overclocking might be held at a later date…
Nonetheless, we were shown the system that would be distributed to reviewers for launch day: a mini-ITX motherboard, a Ryzen 5 2400G processor, 2x8 GB of DDR4 memory, and an AMD Wraith cooler. Using the popular benchmarking tool 3DMark Fire Strike, Sami showed a 39% increase over stock performance by overclocking only the memory and the integrated graphics:
- DDR4-2400 + 1250 MHz GPU: 2911 Points - 'stock'
- DDR4-3200 + 1250 MHz GPU: 3322 Points (+14%)
- DDR4-3200 + 1550 MHz GPU: 3596 Points (+24%)
- DDR4-3600 + 1675 MHz GPU: 4048 Points (+39%)
This shows a couple of interesting data points. First, that the DRAM speed directly influences the results a lot: the 14% gain moving to DDR4-3200 and a similar gain again up to DDR-3600 means that we are still dealing with a graphics system that is memory bandwidth limited, even with only 11 compute units. The Infinity Fabric communication pathways are tied to the DRAM speed, so these would also get a speed up – when asked if it would be possible to discern how much of the speedup is due to a pure DRAM improvement and how much from the Infinity Fabric boost, I was told that AMD could do it in house, but it would be impossible for consumers to do.
However, the DRAM speed gains match up with what we have seen with previous generations of AMD APUs, but also it means that Intel’s decision to combine a Vega GPU with high-bandwidth memory is going to be the way forward in this market.
The second point that this data brings is about memory support. One of the major criticisms of AMD’s initial launch of Ryzen was the lack of high-performance memory support, due to a young firmware. AMD has stated that the Ryzen with Vega graphics line of processors have better memory support with newer firmware revisions, capable of driving higher memory speeds. This is, of course, important to APUs. Additional updates will be coming with the new X470 motherboards in Q2.
Facebook
Twitter
RSS
131 Comments
View All Comments
mateau - Monday, January 8, 2018 - link
@Ian ..."It is our understanding that the 12nm process is essentially a 14+ process for GloFo"
Would you please a credible source for that statement. Say, EETimes or other such Industry outlet rather than an garnished from sites who do nothing other than plagiarize work that fits into the point they are try to make.
I too have read on-line media's dismissal of GloFo 12nm process. Everything that I have read on this matter is based upon an early on-line media piece highly critical to AMD that tried to minimize the impact of 12nm. GloFo on their website specifically states that 12nm is a NEW process node and it is 12nm. GloFo has announced that 12nm IS 12nm, a new shrink from 14nm. Announcing falsely would create all sorts of problems with the FTC and lawyers who do nothing but specialize in Class Action suits. Case in point Intel is getting it's clock cleaned due to their knowledge of and failure to eliminate the Meltdown hardware flaw.
As with ALL process nodes, the entire die is not fabbed with say 12nm or 14nm or for that matter the upcoming 7nm node. There are components within the die that can be and sometimes must be greater in size than the process node taped out for the silicon.
Perhaps you shaould also take the time to ask Dr. Su directly if ONE DESIGN covers ALL EPYC, RYZEN and Threadripper design? In my opinion EPYC is a design similar to Zeppelin and Ryzen but with some major design differences that make a server processor.
No sense speculating on something if you can ask AMD CEO directly.
A5 - Monday, January 8, 2018 - link
They may have been able to reduce the minimum possible feature size in a way that lets them claim a new number, but that doesn't necessarily mean that the performance of a finished product is significantly different.It's impossible to know until products are out in the world, because all we have for now is GloFo PR statements, which are going to make things look as rosy as possible.
SaturnusDK - Monday, January 8, 2018 - link
The Zeppelin die covers all Ryzen, Ryzen Pro, Threadripper, and EPYC CPUs. However, there are parts disabled on those designated Ryzen and Threadripper that are enabled on Ryzen Pro and EPYC. I do not know for certain if there is any feature differences between Ryzen Pro and EPYC but judging from AMDs own information there doesn't seem to be.Dr. Swag - Monday, January 8, 2018 - link
https://www.anandtech.com/show/11854/globalfoundri...based on improvements in the 10-20% range calling it "essentially a 14nm+ node" seems reasonable to me.
iwod - Monday, January 8, 2018 - link
Huh?Meltdown - Not relevant here.
TSMC called their 16nm++ ( or was it 16nm +++ ) as 12nm.
And FTC would have a problem with Intel because their 10nm really should be 7nm by Fab Industry measurement.
And it is not garish, it is straight from GF investor conference.
Design? Speculate? Is Goldmount the same design as Goldmount+. Do QA and feature / yield test accounts for design? And more importantly, do even any other chip maker disclose these information? There is something call trade secret.
LurkingSince97 - Thursday, January 18, 2018 - link
It is well known knowledge that GloFlo 're-branded' their 14nm+ to 12nm.There is nearly a dozen different feature sizes in a process these days, no one number captures it.
Also, GloFlo is not the only one that has done this sort of thing.
LurkingSince97 - Thursday, January 18, 2018 - link
To be more specific, GloFo's own conference / announcement talked about the name change... AMD's roadmaps had "Zen+" on a GloFo 14nm+ node, then suddenly that same roadmap changed it to "12nm" when GloFo announced the new node (and named it)coolhardware - Monday, January 8, 2018 - link
Get well soon Ian!Luposian - Wednesday, January 10, 2018 - link
Wouldn't now be a perfect time to implement changes to the predictive branch execution portions of the new processors, to eliminate issues with Spectre and Meltdown, since they're just making these processors now?LurkingSince97 - Thursday, January 18, 2018 - link
They have _finished_ the Zen 2 design. Zen already doesn't have problems with Meltdown. Spectre is a lot harder to do anything about, other than maybe adding an instruction for a branch prediction barrier. Maybe they had time to add that to Zen 2, or maybe it can be done even in Zen+ without much trouble. Your software will need to be recompiled against it, however.