Presentation is loading. Please wait.

Presentation is loading. Please wait.

Computer Science, University of Oklahoma Reconfigurable Versus Fixed Versus Hybrid Architectures John K. Antonio Oklahoma Supercomputing Symposium 2008.

Similar presentations


Presentation on theme: "Computer Science, University of Oklahoma Reconfigurable Versus Fixed Versus Hybrid Architectures John K. Antonio Oklahoma Supercomputing Symposium 2008."— Presentation transcript:

1 Computer Science, University of Oklahoma Reconfigurable Versus Fixed Versus Hybrid Architectures John K. Antonio Oklahoma Supercomputing Symposium 2008 Norman, Oklahoma October 6, 2008

2 Computer Science, University of Oklahoma 2 Overview The (past) world of reconfigurable computing The (past) world of multi-core The (emerging) world of reconfigurable multi-core architectures Illustrative analysis Conclusions

3 Computer Science, University of Oklahoma 3 Drivers for reconfigurable computing Near performance of custom ASIC Near cost of commodity processor More flexible than custom ASIC “Programming” tools improving steadily Often used in embedded applications having high computational throughput requirements and strict SWAP constraints

4 SAR processing on a UAV “Predator” Jeffrey T. Muehring, “Optimal Configuration of a Parallel Embedded System for Synthetic Aperture Radar Processing,” MS Thesis, Texas Tech University, Dec

5 5 A prototype hybrid system Data Source PC Annapolis FPGA Subsystem (F) Custom Interface Cables SPARC Mercury DSP/GPP Subsystem Data Sink PC Annapolis FPGA Subsystem (B)

6 Data Sink PC Data Source PC Custom Interface Cables Mercury DSP/GPP Subsystem Annapolis FPGA Subsystem (F) Annapolis FPGA Subsystem (B) SPARC Data Sink PC Data Source PC Custom Interface Cables Mercury DSP/GPP Subsystem Annapolis FPGA Subsystem (F) Annapolis FPGA Subsystem (B) SPARC A prototype hybrid system

7 Minimum Power Configurations Resolution Velocity X T X r X a YTYrYaYTYrYa Jeffrey T. Muehring, “Optimal Configuration of a Parallel Embedded System for Synthetic Aperture Radar Processing,” MS Thesis, Texas Tech University, Dec

8 Minimum Power Jeffrey T. Muehring, “Optimal Configuration of a Parallel Embedded System for Synthetic Aperture Radar Processing,” MS Thesis, Texas Tech University, Dec

9 Computer Science, University of Oklahoma 9 Overview The (past) world of reconfigurable computing The (past) world of multi-core The (emerging) world of reconfigurable multi-core architectures Illustrative analysis Conclusions

10 Computer Science, University of Oklahoma 10 Drivers for multi-core technology path Single-core path leading to increased cost, heat, and power consumption Single-core path widens the pocessor/memory speed gap Multi-core path transparent to many application domain developers Multi-core path can improve performance of threaded software

11 11 Typical multi-core architecture* Dual Core Chip L2 Cache Core Memory Dual Core Chip L2 Cache Core *L. Chai, Q. Gao, D.K. Panda, “Understanding the Impact of Multi-Core Architecture in Cluster Computing: A Case Study with Intel Dual-Core System,” Seventh Int'l Symposium on Cluster Computing and the Grid (CCGrid), Rio de Janeiro - Brazil, May 2007.

12 Computer Science, University of Oklahoma 12 Overview The (past) world of reconfigurable computing The (past) world of multi-core The (emerging) world of reconfigurable multi-core architectures Illustrative analysis Conclusions

13 Computer Science, University of Oklahoma 13 Emerging drivers and requirements for multi-core architectures Scale to support massively data parallel (SPMD) applications Match coupling among cores with application granularity Power is a major challenge for large data centers and supercomputing facilities

14 Computer Science, University of Oklahoma 14 Hybrid architectural framework Multi-core Chip Reconfigurable Logic Core L2 Cache L2 Cache MU Core L2 Cache L2 Cache Core L2 Cache Core L2 Cache MU Reconfigurable logic

15 Computer Science, University of Oklahoma 15 Shared everything configuration Multi-core Chip Core L2 Cache L2 Cache MU Core L2 Cache L2 Cache Core L2 Cache Core L2 Cache MU Interconnection Network Reconfigurable logic

16 Computer Science, University of Oklahoma 16 Shared nothing configuration Multi-core Chip Core L2 Cache L2 Cache MU Core L2 Cache L2 Cache Core L2 Cache Core L2 Cache MU Co- Proc Co- Proc Co- Proc Co- Proc Co- Proc Co- Proc Reconfigurable logic

17 Computer Science, University of Oklahoma 17 Hybrid configuration Multi-core Chip Core L2 Cache L2 Cache MU Core L2 Cache L2 Cache Core L2 Cache Core L2 Cache MU Co-Proc Interconnection Network Reconfigurable logic

18 Computer Science, University of Oklahoma 18 Features of hybrid architecture Match core coupling and core processing capacity with application granularity –Fixed multiprocessor architecture not well matched with all application granularities –Proposed reconfigurable multi-core architecture can be configured to match core coupling with application granularity

19 Computer Science, University of Oklahoma 19 Mismatched SPMD execution core 1 core 2 core 3 core 4 core c time Core coupling too loose relative to application granularity Communication time Computation time

20 Computer Science, University of Oklahoma 20 core 1 core 2 core 3 core 4 core c Matched SPMD execution time Core coupling tightened to match application granularity Communication time Computation time

21 Computer Science, University of Oklahoma 21 Overview The (past) world of reconfigurable computing The (past) world of multi-core The (emerging) world of reconfigurable multi-core architectures Illustrative analysis Conclusions

22 Computer Science, University of Oklahoma 22 Illustrative Analysis Notation –Number of cores: c –Problem size: n –Sequential time complexity: –Parallel time complexity: –Computational complexity: –Communication complexity: –Core coupling ratio:

23 Computer Science, University of Oklahoma 23 Example Sequential Time: Parallel Time: Speedup: The value of K: related to core processing capacity The value of L: related to interconnection among cores

24 Computer Science, University of Oklahoma 24 K = 1.0, L = 1.0 Number of cores, c

25 Computer Science, University of Oklahoma 25 K = 1.5, L = 0.5 Number of cores, c

26 Computer Science, University of Oklahoma 26 K = 0.5, L = 1.5 Number of cores, c

27 Computer Science, University of Oklahoma 27 n = 1024 Number of cores, c

28 Computer Science, University of Oklahoma 28 Overview The (past) world of reconfigurable computing The (past) world of multi-core The (emerging) world of reconfigurable multi-core architectures Illustrative analysis Conclusions

29 Computer Science, University of Oklahoma 29 Conclusions Current multi-core approaches may not scale to support massive parallelism Hybrid reconfigurable multi-core approach enables trades between core coupling and core processing capacity More research needed in reconfigurable micro-architecture to support hybrid architectures


Download ppt "Computer Science, University of Oklahoma Reconfigurable Versus Fixed Versus Hybrid Architectures John K. Antonio Oklahoma Supercomputing Symposium 2008."

Similar presentations


Ads by Google