Jo

1
Reconfigurable Computing CHALLENGES
Pedro C. Diniz

• João M. P. Cardoso

Portugal
The High Level Conference on Nanotechnologies,
Braga, Portugal, 20 November 2007
2
Outline

• Reconfigurable Computing
• Challenges
• Role of Reconfigurable Technologies in Future
  Execution Environments
• Overall Actions

3
Reconfigurable (Custom) Computing

• Hardware resources can be configured to a
  specific architecture
• Specialized Functional Elements and Processing
  Elements
• Interconnect between Nodes Custom to Data flow
  in the application
• Configurable on-chip memories (size, data-width,
  indexing)
• Execution Models (Pipelined, Multithreading,
  VLIW) all possible in the same reconfigurable
  fabric

4
Reconfigurable (customizable) Fabrics
5
Reconfigurable (customizable) Fabrics
6
Reconfigurable (Custom) Computing

• Orders of magnitude of speed-up over traditional
  computing systems
• When? Customization is the key
• High operation- and task-level parallelism
• Increased by storage organization (data
  replication/distribution over multiple on-chip
  memories)
• Non-Standard Numeric Formats (fixed-point, etc.)
• Custom Routing

7
Are Architectures Merging?Multi-(Many)-core vs.
Reconfigurable

• Regularity of Reconfigurable Fabrics (e.g.,
  FPGAs) allow them to ride Moore Law
• Unbelievable large number of devices
• Hard-macro cores can be plugged-in

Multicore
Manycore
Reconfigurable Fabrics
8
Reconfigurable Computing Execution Models

• Data travel on paths statically or dynamically
  defined
• Many on-chip Memories

Configurable memory
Configurable logic
9
Reconfigurable Computing Managing Data
Availability

• Many on-chip Memories
• Each Array May be accessed in Parallel
• Custom Pipelining
• On-chip configurable memories can be adapted to
  communication needs
• Replication Increases Data Availability
• By writing to memories in Tandem using a
  customized bus

10
Reconfigurable (Custom) Computing

• Benefits
• Reconfiguration is ideal for fast prototyping and
  early evaluation of realistic performance
• Performance
• Tolerate Defects
• Costs
• Added complexity of execution models makes
  programming very hard (we have not yet solved the
  parallel programming problem yet, sort of)

11
Reconfigurable Computing
Many companies Cray, SGI, SRC, ARC,
PACT, PicoChip, Tilera, etc.
Based on source Bezdek, J.C, Fuzzy models - what
are they, and why, IEEE Trans. on Fuzzy Systems,
1993.
Reconfigurable Computing has already achieved
this point!
12
Reconfigurable Computing

• The Sony PSP Example
• Reconfigurable Architecture Virtual Mobile
  Engine (VME) audio
• 24-bit data width
• 166 MHz
• Single-cycle context switch

http//www.hotchips.org/archives/hc16/3_Tue/8_HC16
_Sess8_Pres1_bw.pdf
13
Reconfigurable Computing

• The Sony PSP Example
• Reconfigurable Architecture Virtual Mobile
  Engine (VME) audio
• 24-bit data width
• 166 MHz
• Single-cycle context switch

http//www.hotchips.org/archives/hc16/3_Tue/8_HC16
_Sess8_Pres1_bw.pdf
14
Challenges
15
Programming

• Years of efforts on parallelizing compilers
  yielded meager returns on the potential for
  concurrent execution
• Movement in industry for new concurrent
  programming paradigms and languages (upc, X-10,
  Fortress, etc.)

Showstopper Programming is excruciatingly
painful How to make devices like FPGAs easily
programmable is a hard research problem, still.
Application Code
Reconfigurable Architectures and Execution Models
Compilation, Synthesis and Optimization
16
Programming

• Future reconfigurable architectures will
  exacerbate all the programming problems
• Issues
• How can programming languages help the compiler?
• How can architectures help the compiler and tools?

Application Code
Reconfigurable Architectures and Execution Models
Compilation, Synthesis and Optimization
17
What is needed?
The Looming Software Crisis due to the MULTICORE
Menace Saman Amarasinghe, MIT
18
What is needed?
19
What is needed?
Advances in Computer Architecture
Success of Reconfigurable Computing
Advances in Tools
Advances in Programming Languages
Advances in Compilers
20
Role of Reconfigurable Technologies in Future
Execution Environments
21
Technology
Based on Robinett et al., Computing with a
Trillion Crummy Components, COMMUNICATIONS OF
ACM, Sept. 2007
1950s
2015s?
1960s
Technology
Key
22
Technology

• Uniprocessor Performance (SPECint)

From Hennessy and Patterson, Computer
Architecture A Quantitative Approach, 4th
edition, 2006 Based on a slide by Saman
Amarasinghe, MIT
All major manufacturers moving to multicore
architectures
Number of Transistors
Nanotechnology may impose a step to
Reconfigurable Computing!
Understandable first step
23
Technology

• Success of memories
• Scalable, highly regular structures
• New cells not working do not compromise chip,
  reduce size
• Reconfigurable computing architectures
• Matrix oriented
• Similar scalable, regular structures
• Cells not working do not compromise chip, reduces
  number of available resources

24
Nanotechnology and Reconfigurable Computing a
perfect match?

• Future computer substrates of nanotechnology-based
  devices
• likely have structure similar to current
  reconfigurable architectures
• Implication
• Most solutions to reconfigurable computing are
  likely to be applicable to nanotechnology
• Issues exacerbated by unreliability
• Tools should reconfigure architectural layer
  based on defects/faults

Nanoarray proposed by Andre DeHon, 2002
25
Future Nanofabrics Exploiting Reconfigurability?

• Changing Application Requirements
• Input Application can have widely varying
  requirements
• First handling some touch-pad interaction, next
  doing video processing
• Real-time versus off-line needs.
• Unreliable Computing substrates
• Defects/Faults
• Comments
• No killer-app still for reconfigurability,
  despite 10 years of searching
• Emerging substrates might prove to be a key
  ground for reconfiguration as either
• Cost of detection and correction of faults is
  simply too high
• The environment is inherently unreliable

26
Future Nanofabrics Programming for
Reconfigurability?

• No Clear Good Approach Today
• Programming Languages and Environments Too Rigid
• Change and Failures are never a first class
  citizen
• Shall we expose some (but not all!) aspects of
  recovery to the programmer?
• Some times failures might not be critical
• Need to offer a system with graceful performance
  degradation

27
Nanotechnology and Reconfigurable Computing

• We have been here before! (a déjà vu)

28
Overall Actions

• We are at an unique opportunity in time
• New comers in the game do not need to go through
  all the steps of the ladder
• Opportunity for EU to take the leadership
• Key investments (joint efforts on)
• Advances in Computer Architecture
• Advances in Programming Languages
• Advances in Compilers
• Advances in Tools

29
Thank You!

• João M. P. Cardoso
• jmpc_at_acm.org
• http//prosys.inesc-id.pt/jmpc