1. Berlin, Germany – January 21st, 2013 A2B: A F RAMEWORK FOR F AST P ROTOTYPING OF R ECONFIGURABLE S YSTEMS Christian Pilato, R. Cattaneo, G. Durelli, A.A. Nacci, M.D. Santambrogio, D. Sciuto Politecnico di Milano Dipartimento di Elettronica. Informazione e Bioingegneria, Italy W ORKSHOP ON R ECONFIGURABLE C OMPUTING (WRC)

2. Berlin, Germany – January 21st, 2013 2 Motivations  The design of reconfigurable systems is a difficult task Interactions between the different phases have to be taken into account  Decision in the frontend phase may highly affect the backend implementation E.g.: Mapping onto reconfigurable regions and floorplacing of the tasks may generate low-quality solutions due to a wrong partitioning or assignment of implementations  The optimal design methodology (and the number of its iterations) cannot be known in advance A2B is an ongoing project at Politecnico di Milano to assist the design of such complex systems

3. Berlin, Germany – January 21st, 2013 3 Agenda  Framework Overview  How do we explore the design space?  How do we generate a design solution?  Conclusions and Future Work

4. Berlin, Germany – January 21st, 2013 4 Framework Overview Evaluation Exploration  Inputs: Information about the target device (.XML) Application source files (.C)  Decision Making (Exploration): Task graph generation Library generation Mapping, Scheduling, Floorplacing Architectural modification  Refinement (Evaluation): Specification of the platform Generation of the SW code  Output: Project files ready for the synthesis with back-end tools

5. Berlin, Germany – January 21st, 2013 5 XML Exchange Format XML file  The entire project can be represented through an XML file Architecture: components’ characteristics (e.g., reconfigurable regions), … Applications: source code files and profiling information Library: task implementations with the characterization (time, resources,...) Partitions: task graph, mapping and scheduling, … modular organization  It allows a modular organization of the framework, but also the sharing of information among the different phases The phases can be applied in any order to progressively optimize the design The designer can perform as many iterations as he/she wants to refine the solution  Specific details of the target architecture are taken into account only in the refinement phase (interactions with backend tools)

6. Berlin, Germany – January 21st, 2013 6 Task Graph and Library Generation  Application source code files can be analyzed to extract the task graphs Profiling information can drive the generation of such solutions  Task graph will be then specified in the XML file as processing nodes connected by data transfers Currently they can be designed by hand, but automated methodologies for automatic parallelization will be investigated in the future Transformations to improve the description by splitting/merging the tasks  LLVM-based compiler to extract the DFG of each task Estimation of required resources (including bit-width analysis) Interaction with HLS synthesis tools to obtain more accurate results  Generated implementations are then store in the XML file to offer opportunities to the mapping phase and information to the floorplacer

7. Berlin, Germany – January 21st, 2013 7 Mapping, Scheduling and Floorplacing one or more configurations  We generate one or more configurations where each task of the applications is analyzed and assigned to An available and admissible implementation A component of the architecture (e.g., processor or reconfigurable region)  This supports to hardware sharing “share” implementations across different tasks (hardware sharing) task relocation move a task implementation to another processing element at run-time (task relocation)  Tasks assigned to the same reconfigurable region are analyzed to determine its constraints and requirement of resources Floorplacing of the regions to determine their positions and the feasibility of the mapping

8. Berlin, Germany – January 21st, 2013 8 Architecture Exploration  An additional step can be included to explore the target architecture Adding/removing processing elements Modifying their parameters Determining the proper interconnection topology  It can affect: task graph transformations and library generation mapping and floorplacing: modification to the computational resources (especially the number of reconfigurable regions)  It allows a progressive refinement of the solution and a concurrent customization of both architecture and application E.g.: mapping and floorplacing can suggest which resources should be added

9. Berlin, Germany – January 21st, 2013 9 Supported Platforms  Virtex-5 XC5VLX110T Two XCF32P Platform Flash PROMs (32Mbyte each) SystemACE™ Compact Flash configuration controller 64-bit wide 256Mbyte DDR2 small outline DIMM (SODIMM)  MPC Research Platform MaxWorkstation Intel i7 2600s@2.8GHz, 16GB RAM, 500GB HDD Max3 dataflow engine (DFE) Virtex 6 SX475T FPGA, 24GB memory DFE connected to CPU via PCI Express gen2 x8 XUPV5 Reconf. Area DDR2 (256MB) CPU0 CPU1 CPU MAX3 DFE DRAM (16GB) Interface FPGA Compute FPGA DRAM (24GB)

10. Berlin, Germany – January 21st, 2013 10 Solution Refinement for the Target Platforms CPU Compiler.c.xml Bitstream Generation HLS (MaxJ-VHDL) -Source code for CPU -DFGs for HW tasks -Mapping configurations Bitstream Generation exec bin bit Manual VHDL Implementations DFG-C HLS (C-VHDL) Manual MaxJ Implementations FPGA-based embedded system MaxWorkstation The code can be always further optimized by hand; e.g., glue code for data transfers MaxIDE DFG-MaxJ

11. Berlin, Germany – January 21st, 2013 11 Graphical User Interface (GUI)  Practical GUI to support the designer, to limit the errors in the interactions with the XML and to allow custom design methodologies

12. Berlin, Germany – January 21st, 2013 12 Conclusions and Future Work  A2B is a modular framework to design reconfigurable systems Easy to plug alternative methods for each of the phase Possibility to perform progressive refinement of both application and architecture  A2B is becoming part of a larger project (ASAP – Advanced Synthesis of Applications and Platforms) Refinement will also include the generation of SystemC TLM models of the target system for (co-)simulation and early validation Closer interaction with actual synthesis (e.g., high-level synthesis) Automated methodologies to accelerate the design Our goal is to make them available to the community (open-source) as soon as possible (a.k.a. ASAP)!

13. Berlin, Germany – January 21st, 2013 Thank you! Christian Pilato pilato@elet.polimi.it the European Community’s Seventh Framework Programme, FASTER project. Research partially funded by the European Community’s Seventh Framework Programme, FASTER project.