*cough*What a load of BS*cough*
https://ieeexplore.ieee.org/document/979585/
and "High performance fully-depleted tri-gate CMOS transistors"
Fully-depleted (FD) tri-gate CMOS transistors with 60 nm physical gate lengths on SOI substrates have been fabricated. These devices consist of a top and two side gates on an insulating layer. The transistors show near-ideal subthreshold gradient and excellent DIBL behavior, and have drive current characteristics greater than any non-planar devices reported so far, for correctly-targeted threshold voltages. The tri-gate devices also demonstrate full depletion at silicon body dimensions approximately 1.5 - 2 times greater than either single gate SOI or non-planar double-gate SOI for similar gate lengths, indicating that these devices are easier to fabricate using the conventional fabrication tools
It was to hard says Intel. With Samsung, STM, GlobalFoundries saying otherwise.
In "A 153Mb-SRAM Design with Dynamic Stability Enhancement and Leakage Reduction in 45nm High-K Metal-Gate CMOS Technology"
"The design contains
fully integrated dynamic forward-body-bias to achieve lower voltage operation while keeping low the area and power overhead. The dynamic sleep design, which was developed at the 65nm node (Zhang et al., 2005), is further enhanced with op-amp-based active-feedback control and on-die programmable reference-voltage generator. The new sleep design reduces the effect of PVT variation, leading to further power reduction. The modular architecture of the design also enables the 16KB-subarray to be used directly as the building block for a 6MB L2 cache in the Core^TM 2 CPU (George, 2007). The design operates over 3.5GHz at 1.1V."
The return of body biasing will re-utilize older engineers for new designs.
scis.scichina.com/en/2016/061402.pdf
"Contrary to the common belief that FDSOI technology only holds promise in niche low-power applications, the results of the above ASIC chip demonstrate its benefit for mid- and high-performance applications. Any application that is concerned about active power, whether an always-on IoT/wearable or a high performance server/networking chip, should consider the implications of FBB and low voltage operation offered by FDSOI."
Unlike, Samsung and GlobalFoundries. Intel's UTB can be done with Shin-Etsu Handotai. Which is super aggressive and flaunting their UTB ready muscles.
2011 and then onward marked a shift in Intel. 2.5D/3D ICs allow UTB and Tri-gate to be used together for the best advantages of each.
Why are you linking to something written about the original Hammer (K8) architecture back in 2001, a decade before Bulldozer came out? Hammer wasn't a CMT architecture and I didn't see any mention of CMT or VISC in that link.
SoftMachines based their designs off that. So, Intel has;
1. Jim Keller who built or touched upon the foundation of the VISC architecture, the CPU core.
2. SoftMachines group who built VISC and had path finding to Ordos.
As of currently: TigerLake is Ordos(WLC) and AlderLake is Ordos+(OCC).
Sanjay Jha was also considered the first time before BK, and he is in the short-list a second time.
The roadmaps built by Michael Mayberry also indicate that Intel for 2.5D/3D must also open up.
- Quantum Well and UTB-SOI
- FinFETs, GAA: Nanowire/Nanosheet, and Dots/Vertical.
For example,
- stacked SRAM with UTB-SOI means low variation, low area, lower overhead, lower voltage.
- Successor to "Tangle Lake" , cheaper Qubits, etc.
- Intel's Lolhi which has a die size of 60 mm squared. Would probably be better on UTBB SOI, etc.
Sanjay Jha getting the position would put some credence to Intel-GlobalFoundries "stuff". IBM would probably fling their support in on Sanjay Jha if they get SOI FinFET/Nanosheet. Even, if they consider it a deal with their competitor.
There is something against Sanjay Jha however:
http://roshmere.com/
^He is the CEO of this.
http://www.computerworld.in/news/intels-alexis-bjorlin-explains-why-silicon-photonics-matter
https://www.ofcconference.org/en-us...re-demonstrates-800gbps-per-wavelength-all-s/
imho, that is more likely to be purchased than Ampere with Renee James.
http://ieeetv.ieee.org/technology/i...-not-wait-for-the-killer-app?rf=channels|104&
https://ieeetv.ieee.org/conference-...world-forum-santa-clara-2018?rf=channels|104&
Data-centric Intel anyone?
While not sacrificing the Foundry?
Facebook
Twitter