Download presentation
Presentation is loading. Please wait.
Published byHilary Stewart Modified over 9 years ago
1
Array Synthesis in SystemC Hardware Compilation Authors: J. Ditmar and S. McKeever Oxford University Computing Laboratory, UK Conference: Field Programmable Logic and Applications (FPL), 2007 Presenter: Tareq Hasan Khan ID: 11083577 ID: 11083577 ECE, U of S Literature review-1 (EE 800) Literature review-1 (EE 800)
2
2 Outline Introduction to SystemC and Agility compiler Mapping SystemC Array In Register In Register In FPGA’s RAM In FPGA’s RAMResultsConclusion
3
3 SystemC High level programming language to describe hardware C++ syntax Synthesizable subset defined by Open SystemC Initiative (OSCI)
4
4 Agility SystemC Compiler Translate SystemC program to HDL language (VHDL, Verilog, RTL SystemC) Agility accepts most C++ constructs, such as: conditional statements — if, switch conditional statements — if, switch loop statements — while, do...while, for loop statements — while, do...while, for control flow — break, continue, return control flow — break, continue, return
5
5 Agility Design Flow
6
6 Outline Introduction to SystemC and Agility compiler Mapping SystemC Array In Register In Register In FPGA’s RAM In FPGA’s RAMResultsConclusion
7
7 Mapping SystemC Array SystemC Array Synthesis Registers using general purpose logic Registers using general purpose logic RAM blocks of modern FPGA RAM blocks of modern FPGAConstrain SystemC arrays offer parallel access to elements SystemC arrays are accessed in one cycle SystemC arrays have write-before-read semantics “Semantics of SystemC array must be “Semantics of SystemC array must be preserved after synthesis “
8
8 Outline Introduction to SystemC and Agility compiler Mapping SystemC Array In Register In Register In FPGA’s RAM In FPGA’s RAMResultsConclusion
9
9 Array Synthesis in Registers for Read Access Each array element produces a register in hardware The address decoder translates address x into a bit vector which controls the output multiplexer The multiplexer selects the output of the particular element that is indexed by x If an array is read several times, several address decoders and multiplexers are required Mux y = Array [ x ] x y
10
10 Array Synthesis in Registers for Write Access The output lines of the address decoder are connected to the write enables of the registers to select in which register to write If an array is written to multiple times in a same clock cycle, multiplexers are required on the inputs of the registers to select which data to write Array [ x ] = y x y
11
11 Outline Introduction to SystemC and Agility compiler Mapping SystemC Array In Register In Register In FPGA’s RAM In FPGA’s RAMResultsConclusion
12
12 Array Synthesis in FPGA’s RAM Modern FPGA contains large amount of RAM Altera Cyclone II has 512KB of SRAM Altera Cyclone II has 512KB of SRAM A SystemC Array can be accessed by several processes in a single clock cycle Multi-port RAM is needed Multi-port RAM is needed An algorithm is developed to automatically assign memory access to RAM ports An algorithm is developed to automatically assign memory access to RAM ports
13
13 Algorithm Constraints and Steps The algorithm must assign memory access to ports such a way that: A memory port is never accessed more than once in a clock cycle A memory port is never accessed more than once in a clock cycle The total number of memory ports must be minimized The total number of memory ports must be minimizedSteps 1. Control Flow Graph (CFG) 2. Access Analysis 3. Port Assignment
14
14 Control Flow Graph (CFG) Codes between two consecutive wait statements executes concurrently in a clock cycle All possible paths that might be traversed during program execution
15
15 Access Analysis and Port Assignments PathRW 111 211 302 One Read only port and Two Write only port required One Write only port and One Read-Write port requiredPathWRW111 211 311 R=Read only port W=Write only port RW = Read-Write port Each path executes in a single clock cycle OR
16
16 Access Analysis Algorithm
17
17 Outline Introduction to SystemC and Agility compiler Mapping SystemC Array In Register In Register In FPGA’s RAM In FPGA’s RAMResultsConclusion
18
18 Results Inverse Desecrate Cosine Transform (IDCT) algorithm written in C contains array of 1KB The design ported from C to SystemC Used Agility compiler to produce EDIF (Xilinx Virtex-4 FPGA)
19
19 Conclusion Mapping of SystemC Array to hardware In Resister File In Resister File Take more logic area Less performance In FPGA RAM In FPGA RAM A new algorithm to map arrays to memory has been presented More efficient in terms of logic area
20
20 Thanks
Similar presentations
© 2025 SlidePlayer.com. Inc.
All rights reserved.