1 / 6

The Next Generation of Microprocessors

The Next Generation of Microprocessors. Senior ASIC Design Engineer Central Engineering AMD i gor.ikodin o vic @ amd.com. Igor Ikodinovi ć. A Paradigm Shift. New inflection point in computing industry driven by: Energy efficiency becoming increasingly important

Audrey
Download Presentation

The Next Generation of Microprocessors

An Image/Link below is provided (as is) to download presentation Download Policy: Content on the Website is provided to you AS IS for your information and personal use and may not be sold / licensed / shared on other websites without getting consent from its author. Content is provided to you AS IS for your information and personal use only. Download presentation by click this link. While downloading, if for some reason you are not able to download a presentation, the publisher may have deleted the file from their server. During download, if you can't get a presentation, the file might be deleted by the publisher.

E N D

Presentation Transcript

  1. The Next Generationof Microprocessors Senior ASIC Design Engineer Central Engineering AMD igor.ikodinovic@amd.com Igor Ikodinović

  2. A Paradigm Shift New inflection point in computing industry driven by: • Energy efficiency becoming increasingly important • New applications using parallel algorithms • Scarce resources for parallel application development Market trends: • Balance of performance, energy efficiency, and cost • Equalizing hardware and software costs • The move to virtualization • Emergence of stream processing • Diverging system requirements

  3. x86 Processor Generations Different Methods for Improving Application Performance 2010 1990 1985 2000 Generation two: Homogeneous Multicores Generation three: Heterogeneous Multicores Generation one: Frequency & Architecture Improved throughput = Better performance Higher frequency = Better performance Best hardware for the task = Better performance, energy efficiency, and lower cost

  4. Fusion CPU-centric approach • Extending the standard x86 instruction set • Not optimal in terms of cost/performance/energy GPU-centric approach • All processing delegated onto GPUs • Not optimal for fine-grain parallelism Balance and optimization approach • AMD: Fusion = CPU + GPU + MM accelerators • Intel: CPU + C/GPU grid (Larrabee)

  5. Key Issues Power • Static: HVT cells, voltage islands, power gating • Dynamic: voltage & clock throttling (power controler) Performance • CPU architecture: SMP, MCM, # cores/CPU, # ALUs/core, cache hierarchy • System architecture: NUMA, system memory, point-to-point I/O links Cost • Design for yield • Test & verification: Months x 100s of platforms

  6. Design & Technology Design • Shift from custom to standard cell • Flow automation • Worldwide team collaboration Technology • New technology nodes: 45nm, 32nm, 22nm, … • New litography techniques: immersion litography • Old story: SOI (AMD CPUs) vs Bulk Si (rest of world) • High-k gates (to address gate leakage) • Work adapted to a given process (Fab)

More Related