![Page 1: Reconfigurable Computing CS294-6 Fall 1998 Dr. Andre DeHon](https://reader035.vdocument.in/reader035/viewer/2022062515/56649c815503460f94939639/html5/thumbnails/1.jpg)
Reconfigurable Computing
CS294-6
Fall 1998
Dr. Andre DeHon
![Page 2: Reconfigurable Computing CS294-6 Fall 1998 Dr. Andre DeHon](https://reader035.vdocument.in/reader035/viewer/2022062515/56649c815503460f94939639/html5/thumbnails/2.jpg)
This Class is About
• Reconfigurable Computing
• Computer Architecture
• Coping with Change
![Page 3: Reconfigurable Computing CS294-6 Fall 1998 Dr. Andre DeHon](https://reader035.vdocument.in/reader035/viewer/2022062515/56649c815503460f94939639/html5/thumbnails/3.jpg)
Outline
• What’s wrong with the status quo
• (Admin break: handouts)
• Reconfigurable Computing: What and Why
• (break)
• What this class is about
![Page 4: Reconfigurable Computing CS294-6 Fall 1998 Dr. Andre DeHon](https://reader035.vdocument.in/reader035/viewer/2022062515/56649c815503460f94939639/html5/thumbnails/4.jpg)
Big Idea
The Biggest Idea here is perhaps the simplest:
When we have 1000x the resources, we design computer
differently.
(Good architecture depends on costs.)
![Page 5: Reconfigurable Computing CS294-6 Fall 1998 Dr. Andre DeHon](https://reader035.vdocument.in/reader035/viewer/2022062515/56649c815503460f94939639/html5/thumbnails/5.jpg)
Fountainhead Pathenon Quote“Look,” said Roark. “The famous flutings on the famous columns---what are they there for? To hide the joints in wood---when columns were made of wood, only these aren’t, they’re marble. The triglyphs, what are they? Wood. Wooden beams, the way they had to be laid when people began to build wooden shacks. Your Greeks took marble and they made copies of their wooden structures out of it, because others had done it that way. Then your masters of the Renaissance came along and made copies in plaster of copies in marble of copies in wood. Now here we are making copies in steel and concrete of copies in plaster of copies in marble of copies in wood. Why?
![Page 6: Reconfigurable Computing CS294-6 Fall 1998 Dr. Andre DeHon](https://reader035.vdocument.in/reader035/viewer/2022062515/56649c815503460f94939639/html5/thumbnails/6.jpg)
What About Computer Architecture?
Are we making copies in submicron CMOS VLSI of copies in NMOS of copies in TTL of early vacuum tube computer designs?
Mainframe->Mini->super microprocessors ?
CDC->Cray1->i860->Vector microprocessors?
![Page 7: Reconfigurable Computing CS294-6 Fall 1998 Dr. Andre DeHon](https://reader035.vdocument.in/reader035/viewer/2022062515/56649c815503460f94939639/html5/thumbnails/7.jpg)
1983 Computer Architecture
• VLSI is “new” to the computer architect
• you have 15Min 4m NMOS
• want to run “all” programs
• What do you build?
2
![Page 8: Reconfigurable Computing CS294-6 Fall 1998 Dr. Andre DeHon](https://reader035.vdocument.in/reader035/viewer/2022062515/56649c815503460f94939639/html5/thumbnails/8.jpg)
What can we build in 15M
• 12Kb SRAM (1.2Kbit)
• 1500 Gate-Array Gates (10K/gate)
• 30 4-LUTs (500K /4LUT)
• 32b ALU+RF+control
2
2
2
2
![Page 9: Reconfigurable Computing CS294-6 Fall 1998 Dr. Andre DeHon](https://reader035.vdocument.in/reader035/viewer/2022062515/56649c815503460f94939639/html5/thumbnails/9.jpg)
1983
• RISC II
• MIPs
![Page 10: Reconfigurable Computing CS294-6 Fall 1998 Dr. Andre DeHon](https://reader035.vdocument.in/reader035/viewer/2022062515/56649c815503460f94939639/html5/thumbnails/10.jpg)
What has changed in 15 years?
• Technology (0.25m CMOS)
• Capacity (20G• Architecture?
2
![Page 11: Reconfigurable Computing CS294-6 Fall 1998 Dr. Andre DeHon](https://reader035.vdocument.in/reader035/viewer/2022062515/56649c815503460f94939639/html5/thumbnails/11.jpg)
Capacity
![Page 12: Reconfigurable Computing CS294-6 Fall 1998 Dr. Andre DeHon](https://reader035.vdocument.in/reader035/viewer/2022062515/56649c815503460f94939639/html5/thumbnails/12.jpg)
Architecture
• Moved memory system on chip
• 32->64b datapath
• +FPU, moved on chip
• 1->4 compute units
• …lots of “hacks” to preserve sequential model of original uP
![Page 13: Reconfigurable Computing CS294-6 Fall 1998 Dr. Andre DeHon](https://reader035.vdocument.in/reader035/viewer/2022062515/56649c815503460f94939639/html5/thumbnails/13.jpg)
Why did we build computers this way in 1983?
Yesterday’s solution becomes today’s historical curiosity.
-- Goldratt
![Page 14: Reconfigurable Computing CS294-6 Fall 1998 Dr. Andre DeHon](https://reader035.vdocument.in/reader035/viewer/2022062515/56649c815503460f94939639/html5/thumbnails/14.jpg)
Why…1983?
![Page 15: Reconfigurable Computing CS294-6 Fall 1998 Dr. Andre DeHon](https://reader035.vdocument.in/reader035/viewer/2022062515/56649c815503460f94939639/html5/thumbnails/15.jpg)
More Why?
![Page 16: Reconfigurable Computing CS294-6 Fall 1998 Dr. Andre DeHon](https://reader035.vdocument.in/reader035/viewer/2022062515/56649c815503460f94939639/html5/thumbnails/16.jpg)
![Page 17: Reconfigurable Computing CS294-6 Fall 1998 Dr. Andre DeHon](https://reader035.vdocument.in/reader035/viewer/2022062515/56649c815503460f94939639/html5/thumbnails/17.jpg)
Have our assumptions changed?
• Beware of cached answers.
• Always check your assumptions.
To stay young requires unceasing cultivation of the ability to unlearn old falsehoods.
-- Lazarus Long
![Page 18: Reconfigurable Computing CS294-6 Fall 1998 Dr. Andre DeHon](https://reader035.vdocument.in/reader035/viewer/2022062515/56649c815503460f94939639/html5/thumbnails/18.jpg)
Why…1983?
![Page 19: Reconfigurable Computing CS294-6 Fall 1998 Dr. Andre DeHon](https://reader035.vdocument.in/reader035/viewer/2022062515/56649c815503460f94939639/html5/thumbnails/19.jpg)
Example
• HP PA-RISC8500 (Hot Chips X)
• SPEC fits in on-chip cache
• What next?
• Does it make sense to keep this architecture and balance as capacity continues to grow?
![Page 20: Reconfigurable Computing CS294-6 Fall 1998 Dr. Andre DeHon](https://reader035.vdocument.in/reader035/viewer/2022062515/56649c815503460f94939639/html5/thumbnails/20.jpg)
Admin: Handouts
• Why Configurable Computing (today’s read)
• XC4K (Thursday read)
• HSRA Overview (Thursday read, hmwrk)
• Terminology
• Course Administrivia
• SPACE1 assignment
• Questionaire (return end of class)
![Page 21: Reconfigurable Computing CS294-6 Fall 1998 Dr. Andre DeHon](https://reader035.vdocument.in/reader035/viewer/2022062515/56649c815503460f94939639/html5/thumbnails/21.jpg)
Challenging our Assumptions
• General-purpose computing machines don’t have to look like processors.
• 1000x increase in single-chip silicon capacity changes the underlying design costs.
![Page 22: Reconfigurable Computing CS294-6 Fall 1998 Dr. Andre DeHon](https://reader035.vdocument.in/reader035/viewer/2022062515/56649c815503460f94939639/html5/thumbnails/22.jpg)
Early RC Successes
• Fastest RSA implementation is on a reconfigurable machine (DEC PAM)
• Splash2 (SRC) performs DNA Sequence matching 300x Cray2 speed, and 200x a 16K CM2
• Many modern processors and ASICs are verified using FPGA emulation systems
• For many signal processing/filtering operations, single chip FPGAs outperform DSPs by 10-100x.
![Page 23: Reconfigurable Computing CS294-6 Fall 1998 Dr. Andre DeHon](https://reader035.vdocument.in/reader035/viewer/2022062515/56649c815503460f94939639/html5/thumbnails/23.jpg)
What is Configurable Computing?
Short answer: Computing via post-fabrication, spatially programmed connection of processing elements.
![Page 24: Reconfigurable Computing CS294-6 Fall 1998 Dr. Andre DeHon](https://reader035.vdocument.in/reader035/viewer/2022062515/56649c815503460f94939639/html5/thumbnails/24.jpg)
Defining Terms
• Computes one function (e.g. FP-multiply, divider, DCT)
• Function defined at fabrication time
• Computes “any” computable function (e.g. Processor, DSPs, FPGAs)
• Function defined after fabrication
Fixed Function: Programmable:
![Page 25: Reconfigurable Computing CS294-6 Fall 1998 Dr. Andre DeHon](https://reader035.vdocument.in/reader035/viewer/2022062515/56649c815503460f94939639/html5/thumbnails/25.jpg)
“Any” Computation?(Universality)
• Any computation which can “fit” on the programmable substrate
• Limitations: hold entire computation and intermediate data
• Recall size/fit constraint
![Page 26: Reconfigurable Computing CS294-6 Fall 1998 Dr. Andre DeHon](https://reader035.vdocument.in/reader035/viewer/2022062515/56649c815503460f94939639/html5/thumbnails/26.jpg)
Benefits of Programmable
• Non-permanent customization and application development after fabrication
• economies of scale (amortize large, fixed design costs)
• time-to-market (evolving requirements and standards, new ideas)
![Page 27: Reconfigurable Computing CS294-6 Fall 1998 Dr. Andre DeHon](https://reader035.vdocument.in/reader035/viewer/2022062515/56649c815503460f94939639/html5/thumbnails/27.jpg)
Distinction in Instruction Binding Time
• Fabrication time --> Fixed function devices
• Beginning of product use --> Actel/Quicklogic FPGAs
• Beginning of usage epoch --> Reconfigurable FPGAs
• Every cycle --> traditional RISC processors
![Page 28: Reconfigurable Computing CS294-6 Fall 1998 Dr. Andre DeHon](https://reader035.vdocument.in/reader035/viewer/2022062515/56649c815503460f94939639/html5/thumbnails/28.jpg)
Spatial vs. Temporal Computing
Spatial Temporal
![Page 29: Reconfigurable Computing CS294-6 Fall 1998 Dr. Andre DeHon](https://reader035.vdocument.in/reader035/viewer/2022062515/56649c815503460f94939639/html5/thumbnails/29.jpg)
Density Comparison
![Page 30: Reconfigurable Computing CS294-6 Fall 1998 Dr. Andre DeHon](https://reader035.vdocument.in/reader035/viewer/2022062515/56649c815503460f94939639/html5/thumbnails/30.jpg)
Spatial/Configurable Benefits
• 10x raw density advantage over processors
• potential for fine-grained (bit-level) control --- can offer another order of magnitude benefit
![Page 31: Reconfigurable Computing CS294-6 Fall 1998 Dr. Andre DeHon](https://reader035.vdocument.in/reader035/viewer/2022062515/56649c815503460f94939639/html5/thumbnails/31.jpg)
Processor vs. FPGA Area
![Page 32: Reconfigurable Computing CS294-6 Fall 1998 Dr. Andre DeHon](https://reader035.vdocument.in/reader035/viewer/2022062515/56649c815503460f94939639/html5/thumbnails/32.jpg)
Configurable Drawbacks
• Each compute/interconnect resource dedicated to single function
• Must dedicate resources for every computational subtask
• Infrequently needed portions of a computation sit idle --> inefficient use of resources
![Page 33: Reconfigurable Computing CS294-6 Fall 1998 Dr. Andre DeHon](https://reader035.vdocument.in/reader035/viewer/2022062515/56649c815503460f94939639/html5/thumbnails/33.jpg)
Where CC interesting?
• Regular applications -- need same operation repeatedly
• High concurrency -- large number of operations can occur simultaneously
• Fine-grained data -- small operand data widths
![Page 34: Reconfigurable Computing CS294-6 Fall 1998 Dr. Andre DeHon](https://reader035.vdocument.in/reader035/viewer/2022062515/56649c815503460f94939639/html5/thumbnails/34.jpg)
Implications?
• Post-fabrication programmable computing space >> processor arch.
• With 10Gdies now and 1T on the horizon, a much wider space of computing architectures opens up.
• Major feature: more spatial processing, less multiplexing/sharing of resources.
2 2
![Page 35: Reconfigurable Computing CS294-6 Fall 1998 Dr. Andre DeHon](https://reader035.vdocument.in/reader035/viewer/2022062515/56649c815503460f94939639/html5/thumbnails/35.jpg)
Break
![Page 36: Reconfigurable Computing CS294-6 Fall 1998 Dr. Andre DeHon](https://reader035.vdocument.in/reader035/viewer/2022062515/56649c815503460f94939639/html5/thumbnails/36.jpg)
This Class
• Good Architecture is driven by media costs
• Technology advances --> Costs change
• What makes sense now, in the near future
• Theme: watch for/make note of where cost assumptions drive architecture
• Be prepared to re-evaluate/review your solutions
![Page 37: Reconfigurable Computing CS294-6 Fall 1998 Dr. Andre DeHon](https://reader035.vdocument.in/reader035/viewer/2022062515/56649c815503460f94939639/html5/thumbnails/37.jpg)
Another Quote
An organization must have some means of combating the process by which people become prisoners of their procedures. The rule book becomes fatter as the ideas become fewer. Almost every well established organization is a coral reef of procedures that were laid down to achieve some long-forgotten objective.
-- John W. Gardner
Some would argue computer architecture is falling prey to this phenomenon.
![Page 38: Reconfigurable Computing CS294-6 Fall 1998 Dr. Andre DeHon](https://reader035.vdocument.in/reader035/viewer/2022062515/56649c815503460f94939639/html5/thumbnails/38.jpg)
Who Course for?
• Programmable Architects (FPGA, processor, etc.)
• ASIC/ASP architects
• System designers who may use any of above
![Page 39: Reconfigurable Computing CS294-6 Fall 1998 Dr. Andre DeHon](https://reader035.vdocument.in/reader035/viewer/2022062515/56649c815503460f94939639/html5/thumbnails/39.jpg)
What’s it about?
• Architectures for late-bound computing systems
• Emphasis on spatial computing
• Re-exam what goes into these architectures and why
• Build up tools, techniques, and intuition for the architect and system designer
![Page 40: Reconfigurable Computing CS294-6 Fall 1998 Dr. Andre DeHon](https://reader035.vdocument.in/reader035/viewer/2022062515/56649c815503460f94939639/html5/thumbnails/40.jpg)
Topics
• Instructions • Interconnect• Compute elements• Retiming• Specialization• Control• Allocation
• Costs• Comparisons• Modeling• Mapping
Architecture Components Related Issues
![Page 41: Reconfigurable Computing CS294-6 Fall 1998 Dr. Andre DeHon](https://reader035.vdocument.in/reader035/viewer/2022062515/56649c815503460f94939639/html5/thumbnails/41.jpg)
Class Requirements
• Participation (reading, class discussion) [20%]• Weekly exercises [60%]• Project Summary [20%]• No tests/final
![Page 42: Reconfigurable Computing CS294-6 Fall 1998 Dr. Andre DeHon](https://reader035.vdocument.in/reader035/viewer/2022062515/56649c815503460f94939639/html5/thumbnails/42.jpg)
Exercises
• 4 common/intro
• 7 “project” -- explore architecture issues from class for a particular application kernel
• Grade best 9 of 11
• Please try all
• Use HSRA as starting point architecture
![Page 43: Reconfigurable Computing CS294-6 Fall 1998 Dr. Andre DeHon](https://reader035.vdocument.in/reader035/viewer/2022062515/56649c815503460f94939639/html5/thumbnails/43.jpg)
Intro Exercises
• Spatial Compute -- high throughput multiply
• Special/Spatial -- FIR
• Cycle -- IIR
• Area-Time -- AT for above
![Page 44: Reconfigurable Computing CS294-6 Fall 1998 Dr. Andre DeHon](https://reader035.vdocument.in/reader035/viewer/2022062515/56649c815503460f94939639/html5/thumbnails/44.jpg)
Project Exercises
• Kernels selected from Multimedia benchmark set (likely JPEG, GSM, ADPCM, maybe rendering…)
• Each student has different kernel (together look at variance in application requirements)
• Different aspect of focus each week
![Page 45: Reconfigurable Computing CS294-6 Fall 1998 Dr. Andre DeHon](https://reader035.vdocument.in/reader035/viewer/2022062515/56649c815503460f94939639/html5/thumbnails/45.jpg)
Project Components
• Analyze sequential version
• Spatial implementation
• Interconnect requirements
• Retiming requirements
• Power implications• Specialization
opportunities and impact
• Programming
![Page 46: Reconfigurable Computing CS294-6 Fall 1998 Dr. Andre DeHon](https://reader035.vdocument.in/reader035/viewer/2022062515/56649c815503460f94939639/html5/thumbnails/46.jpg)
Project Report
• Summarize lessons for some component/feature across class project results
• E.g. Power, AT, Interconnect
![Page 47: Reconfigurable Computing CS294-6 Fall 1998 Dr. Andre DeHon](https://reader035.vdocument.in/reader035/viewer/2022062515/56649c815503460f94939639/html5/thumbnails/47.jpg)
Next Time
• FPGA/HSRA introduction
• Take a look at reading
• HSRA for projects