We are Still Among the Living
AMD talked extensively about their upcoming 2013 releases.
The day after the official AMD presentation we were able to sit down with Leslie Sobon for a good hour and really dig into the products we are expecting throughout this next year. AMD did not officially announce any products, but they revealed more details about products on their roadmaps.
To say that AMD is in a somewhat precarious situation is an understatement. This does not necessarily mean that they won’t survive for some years. This was never mentioned to us by AMD, but we can assume that it is not in ATIC’s best interest to let AMD flounder too much. AMD is still GLOBALFOUNDRIES largest customer, and ATIC believes that they can become a fabrication giant in the next few years. So, while AMD is hitting some hard times, they will be around for some time to come in spite of their issues.
CPUs
Believe it or not, AMD is still a CPU company with some relevant producxts. While Intel has the advantage in x86 performance and process technology, AMD has a distinct advantage in the integrated graphics portion. While Trinity was a big step in the right direction in terms of performance and power consumption, it was not enough to boost their flagging marketshare. Throughout the 2013 they are working on several products that will help to change their fortunes.
The first product that we will likely see is the Jaguar core based Kabini APUs. These are the next generation, low power APUs which will replace the Brazos 2.0 products that we currently are seeing. These quad core and dual core parts are manufactured by TSMC on their 28 nm process. Kabini will be the first APU to include the new GCN architecture that we currently see in the HD 7700 series and above. AMD will be breaking new ground in offering a true quad core part at price points unseen so far.
Overall Kabini should offer around 50% greater performance than the previous Brazos generation of produts. This is a combination of CPU and graphics performance rather than one or the other. AMD believes that there should be around a 20% increase in IPC over the previous Bobcat architecture and a pretty hefty jump in graphics performance. Going from a 40 nm manufacturing node with Brazos to the latest 28 nm process for Kabini should net some major improvements. Jaguar is a bigger CPU design as well as the graphics portion being used. The die size should be comparable to current Brazos chips, but we will again see some thermal improvements as well as greater IPC.
Temash is a subset of Kabini and is intended for the tablet market. The 2C version can scale down to 3.9 watts, which is in the fanless tablet range. The 4C product will go up to 15 watts which is aimed at the convertible market. Temash features a configurable TDP which will allow its use in convertibles. When plugged into the base unit the TDP can be much higher and take advantage of the second battery typically included in the base unit (or AC power) to give better performance. When in tablet mode it will clock down to a lower TDP to maximize battery life. 15 watts is the full power envelope while 8 watts is the convertible mode.
The 4C version will be released in Q2 while the 2C version will be released during the mid-year timeframe. These are most definitely budget oriented parts, but with the push for ultra mobile notebooks and tablets, AMD is hitting a good time frame for the release of these parts. At their tent AMD showed off quite a few Temash based products and their performance was impressive, especially in graphical applications.
On the desktop and traditional notebook market we will see the release of the Richland APU. This is a Trinity based product, but “optimized” for greater performance. The optimizations on the CPU portion are essentially some basic BIOS tweaks as well as small improvements in manufacturing which should allow higher clockspeeds at the same current TDPs. So we will see a couple hundred MHz improvements at the 17, 35, 65, and 95 watt TDPs. The low level BIOS/microcode changes are expected to improve the IPC of these products by a measureable amount, but do not expect anything truly massive. My expectations are within the 3 to 5% range for some if not most workloads.
The graphics side is a bit more of a rework. It is a redesigned graphics portion that will still use the VLIW4 architecture. I doubt that there will be more stream units, but they are changing the layout to improve efficiency. They are going from the current 3 SIMD units to 6 units. There are many other smaller tweaks to the design, but AMD expects a pretty sizeable increase in graphics performance without inflating the die size of the Richland APUs.
The biggest news of this meeting is that AMD detailed their Kaveria APU to us for the first time. Kaveri is a next generation CPU which implements the improved Steamroller architecture and is the second APU to include the GCN graphics architecture. While Kabini implements GCN, it is not a fully enabled HSA part. I do not believe that it has the extensive interconnect technology to be able to do things like the shared address space that HSA dictates. Kaveri will be the first fully HSA enabled part from AMD (and anyone else for that matter).
This particular part will be fabricated by GLOBALFOUNDRIES on their 28 nm node. It will be made on bulk silicon, so no SOI for this part. I was quite surprised that GF would be in a position in the next half year to start ramping production on their 28 nm process. AMD expects to start shipping Kaveri in a late Q4 timeframe this year. These parts will be desktop at first and will transition to mobile in 2014. AMD wants (and needs) to get these parts out in a timely manner, and they are pulling in the launch as much as possible. Hence the desktop first release while they refine production to be able to adequately address the mobile space. Achieving good bins and yields at the higher TDP is easier than trying to hit those numbers for a 35 watt and below product line.
FM2 and FM2+ will be the primary platform for AMD for quite some time. It was essentially confirmed that Trinity has the ability to run PCI-E 3.0 speeds, but the parts were never certified due to a mix of time, budget, and personnel constraints. FM1 did not support PCI-E 3.0 while FM2 does, hence the change in sockets. No timeline has been given for when PCI-E 3.0 will be implemented on the AMD platform, but we were assured that when it makes sense for them to do so we will see products supporting this technology.
AM3 is a big question internally for AMD. These parts are partitioned off of the server market and they are not interchangeable with FM2 (HT vs. PCI-E connections). There will be no updates for some time other than potentially a speed bump down the road with the current Vishera product line. Kaveri will be the first Steamroller part and it is yet unknown if Steamroller will be making it to AM3 or if FM2 will simply replace it in the enthusiast market.
Graphics
AMD has launched the 8500 and 8600 parts to the market, and these are primarily mobile. The last generation HD 7600 and below parts were rebranded VLIW5 parts. So these products are the first GCN based parts for this market in both mobile and discrete desktop. These are not next generation GCN parts or even updates.
The Solar System code name is simply that of notebook parts under the Sea Islands umbrella. These again are not next generation parts as compared to the current 7700 series and above. AMD did not give a timeline as to when they will be refreshing their desktop parts. We have seen rumors of the HD 8900 series, but we do not yet know if this is a new part or simply a rebranded 7900 that coincides with OEM/ODM product refreshes.
In Closing
Kabini will be a very major step for AMD and one that will help them stay competitive in the x86 market. The Richland APU will also improve AMD’s standing, but it will certainly not put AMD back in the running for the performance crown. It will not be until Kaveri that we will see a part that will be more competitive with the current Intel lineup. Unfortunately for AMD, Intel will have released their Haswell parts into the market which promises not only improved CPU performance but also a significant bump in graphics performance and capabilities.
PC Perspective's CES 2013 coverage is sponsored by AMD.
Follow all of our coverage of the show at https://pcper.com/ces!
Does their New stuff support
Does their New stuff support “Connected standby” for Windows 8?
3 Things:
1. The first
3 Things:
1. The first sentence right under CPU’s “Believe it or not, AMD is still a CPU company with some relevant producxts.” Is it supposed to be “producxts”?
2. What is GCN?
3. What is HSA?
GCN = Graphics Core Next
It’s
GCN = Graphics Core Next
It’s the current GPU microarchitecture from AMD you normally find in Radeon graphics cards, introduced last year. It’s a complete break from the former microarchitecture, a VLIW5/VLIW4 based microarchitecture.
HSA = Heterogeneous System Architecture
It’s the evolution of AMD’s Fusion concept, integrating the GPU into the main processor.
GCN = graphics core next
HSA
GCN = graphics core next
HSA = heterosystem architecture
Temash will be the most
Temash will be the most significant step for AMD (Not Richland or Kabini to some extent). Tamesh will be 4 cores and within 5W, the dual core will be 3W.
The ONLY decent product
The ONLY decent product coming out from AMD. AMD Richland IS A JOKE. AMD either needs to get in the game OR GET OUT.
You, again… the stupid
You, again… the stupid monkey.
I’m glad to hear this, aside
I’m glad to hear this, aside from the common “We need AMD” comments, I want them to be along in a healthy way.
If nothing else it will still allow them to be a top player in the graphics market through solid income from all of these low power/mobile parts.
“They are going from the
“They are going from the current 3 SIMD units to 6 units.”
My understanding was Trinity already has 6 SIMD units? 6*64 = 384?
You are correct. I am not
You are correct. I am not sure why AMD said that they were going from 3 to 6. I will inquire further and see what they have to say.
Thanks for the informative
Thanks for the informative post Josh, I believe it is in everyone’s best interests from a competition stand point that AMD survive and grow into the future. They do need to get these parts into the market quicker though.
Great work Josh as
Great work Josh as always.
any word on Am3+ versions of richland/steamroller and anything on the chipset front am3+/fm2+?
What is the on sale date for richland/kabani?
Will kabani be offered in fm2 package?
Amd DDR4 plans?
lastly 28nm/20nm/14nm fdsoi plans
all am3 parts listed as “a
all am3 parts listed as “a big question internally for AMD. ”
with no new updates.
I remember following with anticipation to release when amd first started talking a combined cpu/gpu until i saw the modest spec.
now i am concerned that this modest spec will become the proc for the “enthusiast” of the future
its disappointing to see AMD
its disappointing to see AMD keep slipping on execution and timelines. with a late Q4 release kaveri does not in anyway become a 2013 product. if they had released by computex timeframe then AMD had a chance of being competitive. Haswell will kill Richland on the CPU side and could match or exceed Richland’s GPU. intel is confident about haswell GT3 performance
http://www.anandtech.com/show/6600/intel-haswell-gt3e-gpu-performance-compared-to-nvidias-geforce-gt-650m
2013 is going to be very difficult for AMD. Kabini and Temash will help them do better at the low end sub USD 500 notebook market. But in the USD 500+ notebook and traditional desktop space Haswell will destroy AMD’s products. Its in AMD’s best interests that they get Kaveri in real volume by early – mid Q3 or the bleeding is going to get worse.
Why slipping on execution?
Why slipping on execution? The last information was, Kaveri in 2014. Q4 2013 would be earlier. And probably you won’t see a 2C/4T Haswell before that, the competition of the 2M/4T Kaveri. There is just no 4M APU from AMD that could be a competition to 4C Haswell. At least CPU-wise. And no, Haswell’s IGP can’t compete with Richland, especially not in desktops. It’s even doubtful for the GT3 version in notebooks. Beside the facts that GT3 is an expensive part and Richland will live only half a year or so. Not to mention that Kabini/Temash will wipe the floor with any Atom based processor. So, I think it’s quite impressive what AMD does on 32/28 nm, while Intel already uses 22 nm. Imagine how the processors would look like if vice versa. I’m really looking forward to Kabini and Kaveri. Could be the processors for my next system.
Intel will have Haswell out
Intel will have Haswell out before Q4 2013. Intel CPU’s will destroy AMD in performance. Wrong again AMD Richland’s IGP is a re-brand so Intel haswell’s IGP will be so much better. AMD does sucky work on the desktop with 32nm/28nm. Their low power parts are decent thought but Intel will have 10W and below haswell to yet again Trounce AMD. AMD Re-branding the IGP is plain stupid and dumb.
This is just silly. Have you
This is just silly. Have you looked at any of the performance estimates for Richland. Its IGP will still dominate Haswell. Intel is getting like 50% better than crap today, meanwhile Richland is getting about 40% better than an already dominant position. This may be partially from use of dual channel DDR3-1866 RAM, which significantly improves IGP performance.
AMD Vishera at 32nm tops several benchmarks over i5s and i7s, and that is sucky compared to Intel’s 22nm parts? Intel wont actually be below 10W in TDP with Haswell, but maybe SDP. And in that range I would be very skeptical of its performance.
AMD did not just rebrand their IGP, that is just a dumb comment. Quit being a troll.
AMD has no chance even with
AMD has no chance even with Kaveri because Intel broadwell will be out by the time Kaveri gets out if AMD survives that long. AMD is really just shooting itself in the foot AGAIN. AMD doesn’t have ANYTHING I WANT TO BUY they are NOT Creative nor innovative. They are sticking to the tried and true and they will go down for that very reason.
Are you just a troll, answer
Are you just a troll, answer is yes. AMD is not creative? Does it surprise you to know that Intel uses many of AMD’s patents to make its CPUs? Most of Intel’s performance increases over the years have been from following in AMD’s footsteps. AMD manages to make the worlds most powerful graphics and supercomputers, Intel cannot claim that. I guess Intel is the one that really isnt all that creative.
And how much better do expect Broadwell to be? And what do you expect the Kaveri performance to be? If you dont know either of those, how can you even compare?
Just to be
Just to be condescending..
Ooooh so Intel isnt creative? When AMD’s processors can’t encode videos , compress .rar’s , and a slew of other goodies as fast as Intels processors can ((with hyperthreading and lower clock speeds))? WHo created Hypertrheading I wonder?
Both companies copy each other, there is no doubt in that, alot of stuff is copied yes.
AMD’s Hecta and Octo processors run to hot , eat to much power and do not perform as well as Intels processors do in certain formats. Inital price wise good, but Over all..the power bill will show just how power hungry those are..
When technology is able to utilize that many cores to the utmost effectiveness then I am sure AMD will have a big jump in sales on their high core processors.
This is all my opinion so take it how ever you wish.