Saturday, July 20th 2013
AMD Kaveri APU Delayed To 1H 2014?
AMD's next generation Kaveri APUs might be delayed until early 1H 2014. The Steamroller architecture based APU is officially slated for Q4 2013, but a new report says that Kaveri based APUs are yet to achieve the mass production levels required for a full-scale launch. Of course, AMD could still launch the APUs in limited supply, with the supply eventually improving over the months, but that doesn't ever go well with consumers.
"AMD originally expected to start supplying its Heterogeneous System Architecture (HAS) Kaveri APUs in the second half of 2013, but according to the company's latest plans, the CPU maker will only provide two A10 and one A8 APU pilot production samples to its clients in December 2013, indicating that Kaveri APU-based PC products may have difficulties showing up in the retail channel before April 2014, the sources said."
If AMD is indeed facing manufacturing issues (Kaveri is based on the 28nm bulk process), then they should do a full-scale launch when ready, instead of doing yet another half-baked product.
Kaveri will bring in major changes in the memory field by introducing HUMA, a memory architecture which will allow for unified sharing of RAM between the CPU and GPU, much like the APU implementation in Sony's PlayStation 4.
Source:
DigiTimes
"AMD originally expected to start supplying its Heterogeneous System Architecture (HAS) Kaveri APUs in the second half of 2013, but according to the company's latest plans, the CPU maker will only provide two A10 and one A8 APU pilot production samples to its clients in December 2013, indicating that Kaveri APU-based PC products may have difficulties showing up in the retail channel before April 2014, the sources said."
If AMD is indeed facing manufacturing issues (Kaveri is based on the 28nm bulk process), then they should do a full-scale launch when ready, instead of doing yet another half-baked product.
Kaveri will bring in major changes in the memory field by introducing HUMA, a memory architecture which will allow for unified sharing of RAM between the CPU and GPU, much like the APU implementation in Sony's PlayStation 4.
20 Comments on AMD Kaveri APU Delayed To 1H 2014?
Hope reports of Broadwell being delayed turn out to be true
AMD 28nm
Vs.
Intel 14nm
What is going on with AMD?
AMD is at the mercy of chip foundries like TSMC and Global Foundries (which used to be AMD's fab business).
When you consider the sizes of each company, it's amazing that AMD can stay somewhat competitive.
sorry for my english
Kaveri will in fact ship in Q4 and engineering samples have already been tested by key industry players. AMD has said for some time that the earliest Kaveri would START shipping is Q4 of '13 and in fact some product will ship in Q4 and the rest starting in H1 of '14 as always planned.
All roadmaps from AMD/Intel are subject to changes based on many factors. No one got too excited when Ivy Bridge was delayed for 3 months nor with Haswell being delayed 2+ months.
BTW, for those not in the tech loop... 14nm vs. 28 nm results in minute differences in actual CPU performance. Reduced die shrink below 32nm only shows tangible gains in power consumption. Every die shrink has diminishing gains, thus 14nm gets you a very tiny power reduction over 28nm and virtually no CPU performance gain, as we have seen with Ivy Bridge and Haswell die shrinks at 22nm.
AMD is 1-2 nodes behind but that's not just AMD, that the entire industry. So the reality is that the industry is not behind intel but rather intel is a step ahead of everyone. The industry as a whole is the norm and intel is the exception.
Intel is not sandbagging intentionally. It's just corporate complacency when they're on top. The executives are just enjoying profits but they did not stop developing tech. In fact, it's a wise business decision to hold the tech and patent it. They should only use it to keep themselves one step ahead of the competition to maximize revenue and minimize expense. This strategy gives them a lot of headroom when they hit a crunch. Intel can literally stop R&D for 2 years and still be at least on par with the industry in the server/pc market. That is a good buffer to have when the industry or economy goes to the toilet and you need to lay off 1/4 of your employees to keep yourself in the black.
On the other hand, intel just admitted that they are too slow in the transition to the mobile space. This just proves the fact that they are more overconfident than intentionally sandbagging.
God damn it, if this is true, I won't hide, I'm kind of pissed...