1 / 15

Design methods to facilitate rapid growth of SoCs Arvind

System-on-a-chip. Design methods to facilitate rapid growth of SoCs Arvind Computer Science & Artificial Intelligence Lab Massachusetts Institute of Technology. The biggest SoC drivers. Explosive growth in markets for cell phones game boxes sensors and actuators.

andres
Télécharger la présentation

Design methods to facilitate rapid growth of SoCs Arvind

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. System-on-a-chip Design methods to facilitate rapid growth of SoCs Arvind Computer Science & Artificial Intelligence Lab Massachusetts Institute of Technology http://csg.csail.mit.edu/arvind

  2. The biggest SoC drivers • Explosive growth in markets for • cell phones • game boxes • sensors and actuators Functionality and applications are constrained primarily by: - cost - power/energy constrains http://csg.csail.mit.edu/arvind/

  3. WLAN RF WCDMA/GSM RF WLAN RF WLAN RF Comms. Processing Application Processing Current Cellphone Architecture Today’s chip becomes a block in tomorrow’s chip IP reuse is essential Hardware/software migration IP = Intellectual Property Complex, High Performance but must not dissipate more than 3 watts http://csg.csail.mit.edu/arvind/

  4. An under appreciated fact • If a functionality (e.g. H.264) is moved from a programmable device to a specialized hardware block, the power/energy savings are 100 to 1000 fold Power savings  more specialized hardware but our mind set • Software is forgiving • Hardware design is difficult, inflexible, brittle, error prone, ... http://csg.csail.mit.edu/arvind/

  5. IBM Cell Processor On-chip memory banks Application-specific processing units General-purpose processors Structured on-chip networks SoC Trajectory:multicores, heterogeneous, regular, ... Can we rapidly produce high-quality chips and surrounding systems and software? http://csg.csail.mit.edu/arvind/

  6. Things to remember • Design costs (hardware & software) dominate • Within these costs verification and validation costs dominate • IP reuse is essential to prevent design-team sizes from exploding design cost = number of engineers x time to design http://csg.csail.mit.edu/arvind/

  7. Almost complete reliance on post-design verification for quality Mind set Common quotes • “Design is not a problem; design is easy” • “Verification is a problem” • “Timing closure is a problem” • “Physical design is a problem” http://csg.csail.mit.edu/arvind/

  8. Make Inspect Rework Defect Defect Defect Through the early 1980s: The U.S. auto industry • Sought quality solely through post-build inspection • Planned for defects and rework and U.S. quality was… http://csg.csail.mit.edu/arvind/

  9. … less than world class • Adding quality inspectors (“verification engineers”) and giving them better tools, was not the solution • The Japanese auto industry showed the way • “Zero defect” manufacturing http://csg.csail.mit.edu/arvind/

  10. New mind set:Design affects everything! • A good design methodology • Can keep up with changing specs • Permits architectural exploration • Facilitates verification and debugging • Eases changes for timing closure • Eases changes for physical design • Promotes reuse  It is essential to Design for Correctness http://csg.csail.mit.edu/arvind/

  11. New ways of expressing behavior to reduce design complexity • Decentralize complexity:Rule-based specifications (Guarded Atomic Actions) • Lets you think one rule at a time • Formalize composition: Modules with guarded interfaces • Automatically manage and ensure the correctness of connectivity, i.e., correct-by-construction methodology Strong flavor of Unity Bluespec • Smaller, simpler, clearer, more correct code for both simulation AND synthesis http://csg.csail.mit.edu/arvind/

  12. data_inpush_req_npop_req_nclkrstn data_outfullempty Reusing IP Blocks Example: Commercially available FIFO IP block No machine verification of such informal constraints is feasible These constraints are spread over many pages of the documentation... http://csg.csail.mit.edu/arvind/

  13. theFifo.enq(value1); Enqueue arbitration control theFifo.deq(); value2 = theFifo.first(); enab not full rdy enab not empty rdy not empty theFifo.enq(value3); rdy Dequeue arbitration control theFifo.deq(); value4 = theFifo.first(); Bluespec promotes compositionthrough guarded interfaces Self-documenting interfaces; Automatic generation of logic to eliminate conflicts in use. theModuleA theFifo n enq theModuleB deq FIFO n first http://csg.csail.mit.edu/arvind/

  14. module interface Bluespec: State and Rules organized into modules All state (e.g., Registers, FIFOs, RAMs, ...) is explicit. Behavior is expressed in terms of atomic actions on the state: Rule: guard  action Rules can manipulate state in other modules only via their interfaces. http://csg.csail.mit.edu/arvind/

  15. The Plan • BS2- Simple Example: GCD • BS3- Combinational Circuits: IFFT • BS4- Architectural Exploration: 802.11a • BS5- The IP Lookup problem; subtle concurrency issues • BS6- Bluespec Semantics and Scheduling Primitives Bluespec is available in two versions: BSV – Bluespec in System Verilog ESEPro – Bluespec in SystemC These lectures will use BSV syntax http://csg.csail.mit.edu/arvind/

More Related