This paper introduces McPAT, an integrated power, area, and timing modeling framework that supports comprehensive design space exploration for multicore and manycore processor configurations ranging from 90nm to 22nm and beyond. At the microarchitectural level, McPAT includes models for the fundamental components of a chip multiprocessor, including in-order and out-of-order processor cores, networks-on-chip, shared caches, integrated memory controllers, and multiple-domain clocking. At the circuit and technology levels, McPAT supports critical-path timing modeling, area modeling, and dynamic, short-circuit, and leakage power modeling for each of the device types forecast in the ITRS roadmap including bulk CMOS, SOI, and double-gate transistors. McPAT has a flexible XML interface to facilitate its use with many performance simulators. Combined with a performance simulator, McPAT enables architects to consistently quantify the cost of new ideas and assess tradeoffs of different architectures using new metrics like energy-delay-area2 product (EDA2P) and energy-delay-area product (EDAP). This paper explores the interconnect options of future manycore processors by varying the degree of clustering over generations of process technologies. Clustering will bring interesting tradeoffs between area and performance because the interconnects needed to group cores into clusters incur area overhead, but many applications can make good use of them due to synergies of cache sharing. Combining power, area, and timing results of McPAT with performance simulation of PARSEC benchmarks at the 22nm technology node for both common in-order and out-of-order manycore designs shows that when die cost is not taken into account clustering 8 cores together gives the best energy-delay product, whereas when cost is taken into account configuring clusters with 4 cores gives the best EDA2P and EDAP.

/pdf/mcpat-an-integrated-power-area-and-timing-modeling-framework-38b785sayc.pdf

McPAT: an integrated power, area, and timing modeling framework for multicore and manycore architectures

An object is to provide a semiconductor device of which a manufacturing process is not complicated and by which cost can be suppressed, by forming a thin film transistor using an oxide semiconductor film typified by zinc oxide, and a manufacturing method thereof. For the semiconductor device, a gate electrode is formed over a substrate; a gate insulating film is formed covering the gate electrode; an oxide semiconductor film is formed over the gate insulating film; and a first conductive film and a second conductive film are formed over the oxide semiconductor film. The oxide semiconductor film has at least a crystallized region in a channel region.

Semiconductor device, and manufacturing method thereof

Concern about the performance of wires wires in scaled technologies has led to research exploring other communication methods. This paper examines wire and gate delays as technologies migrate from 0.18-/spl mu/m to 0.035-/spl mu/m feature sizes to better understand the magnitude of the the wiring problem. Wires that shorten in length as technologies scale have delays that either track gate delays or grow slowly relative to gate delays. This result is good news since these "local" wires dominate chip wiring. Despite this scaling of local wire performance, computer-aided design (CAD) tools must still become move sophisticated in dealing with these wires. Under scaling, the total number of wires grows exponentially, so CAD tools will need to handle an ever-growing percentage of all the wires in order to keep designer workloads constant. Global wires present a more serious problem to designers. These are wires that do not scale in length since they communicate signals across the chip. The delay of these wives will remain constant if repeaters are used meaning that relative to gate delays, their delays scale upwards. These increased delays for global communication will drive architectures toward modular designs with explicit global latency mechanisms.

The future of wires

This paper presents the current state of understanding of the factors that limit the continued scaling of Si complementary metal-oxide-semiconductor (CMOS) technology and provides an analysis of the ways in which application-related considerations enter into the determination of these limits. The physical origins of these limits are primarily in the tunneling currents, which leak through the various barriers in a MOS field-effect transistor (MOSFET) when it becomes very small, and in the thermally generated subthreshold currents. The dependence of these leakages on MOSFET geometry and structure is discussed along with design criteria for minimizing short-channel effects and other issues related to scaling. Scaling limits due to these leakage currents arise from application constraints related to power consumption and circuit functionality. We describe how these constraints work out for some of the most important application classes: dynamic random access memory (DRAM), static random access memory (SRAM), low-power portable devices, and moderate and high-performance CMOS logic. As a summary, we provide a table of our estimates of the scaling limits for various applications and device types. The end result is that there is no single end point for scaling, but that instead there are many end points, each optimally adapted to its particular applications.

/pdf/device-scaling-limits-of-si-mosfets-and-their-application-299ftwsxth.pdf

Device scaling limits of Si MOSFETs and their application dependencies

This paper focuses on approaches to continuing CMOS scaling by introducing new device structures and new materials. Starting from an analysis of the sources of improvements in device performance, we present technology options for achieving these performance enhancements. These options include high-dielectric-constant (high-k) gate dielectric, metal gate electrode, double-gate FET, and strained-silicon FET. Nanotechnology is examined in the context of continuing the progress in electronic systems enabled by silicon microelectronics technology. The carbon nanotube field-effect transistor is examined as an example of the evaluation process required to identify suitable nanotechnologies for such purposes.

Beyond the conventional transistor

Two key process features that are used to make 45 nm generation metal gate + high-k gate dielectric CMOS transistors are highlighted in this paper. The first feature is the integration of stress-enhancement techniques with the dual metal-gate + high-k transistors. The second feature is the extension of 193 nm dry lithography to the 45 nm technology node pitches. Use of these features has enabled industry-leading transistor performance and the first high volume 45 nm high-k + metal gate technology.

45nm High-k + metal gate strain-enhanced transistors

Methods for semiconductor device fabrication are provided Features are created using spacers Methods include creating a pattern comprised of at least two first features on the substrate surface, depositing a first conformal layer on the at least two first features, depositing a second conformal layer on the first conformal layer, partially removing the second conformal layer to partially expose the first conformal layer, and partially removing the first conformal layer from between the first features and the second conformal layer thereby creating at least two second features Optionally the first conformal film is partially etched back before the second conformal film is deposited

Feature size reduction

A transistor gate comprises a substrate having a pair of spacers disposed on a surface, a high-k dielectric conformally deposited on the substrate between the spacers, a recessed workfunction metal conformally deposited on the high-k dielectric and along a portion of the spacer sidewalls, a second workfunction metal conformally deposited on the recessed workfunction metal, and an electrode metal deposited on the second workfunction metal. The transistor gate may be formed by conformally depositing the high-k dielectric into a trench between the spacers on the substrate, conformally depositing a workfunction metal atop the high-k dielectric, depositing a sacrificial mask atop the workfunction metal, etching a portion of the sacrificial mask to expose a portion of the workfunction metal, and etching the exposed portion of the workfunction metal to form the recessed workfunction metal. The second workfunction metal and the electrode metal may be deposited atop the recessed workfunction metal.

Recessed workfunction metal in CMOS transistor gates

We report a very high performance 100 nm gate length CMOS transistor structure operating at 1.2-1.5 V. These transistors are incorporated in a 180 nm logic technology generation. Various process enhancements are incorporated to significantly improve transistor current drive capability relative to the results published by Yang et al. (1998). Unique transistor features responsible for achieving high performance are described. NMOS and PMOS devices demonstrate drive current of 1.04 mA//spl mu/m and 0.46 mA//spl mu/m respectively at 1.5 V and 3 nA//spl mu/m I/sub OFF/. These are the best drive currents reported to date at fixed I/sub OFF/. They represents 10% drive current improvement for both NMOS and PMOS devices relative to the results published by Yang without any change in gate-oxide thickness. High performance is demonstrated down to 1.2 V. Inverter delay of less than 10 psec is reported at 1.5 V at very moderate I/sub OFF/ values.

100 nm gate length high performance/low power CMOS transistor structure

A device includes a number of fins. Some of the fins have greater heights than other fins. This allows the selection of different drive currents and/or transistor areas.

Michael K. Harper

Papers

45nm High-k + metal gate strain-enhanced transistors

Feature size reduction

Recessed workfunction metal in CMOS transistor gates

100 nm gate length high performance/low power CMOS transistor structure

Substrate fins with different heights