PAGE 1
HLS Based HW and SW Co-Design
- f Complex IP Subsystems
An Integrated Modeling, Synthesis and Verification Methodology
Xiaojian Liu Greg Smith David Hansen Jeff Wong Louie Lee June 4th, 2013
An Integrated Modeling, Synthesis and Verification Methodology - - PowerPoint PPT Presentation
HLS Based HW and SW Co-Design of Complex IP Subsystems An Integrated Modeling, Synthesis and Verification Methodology Xiaojian Liu Greg Smith June 4th, 2013 David Hansen Jeff Wong Louie Lee PAGE 1 2 Author Page Xiaojian Liu,
PAGE 1
Xiaojian Liu Greg Smith David Hansen Jeff Wong Louie Lee June 4th, 2013
2
3
4
D-Bus C-Bus
CPU GPU Video Codec Audio Proc Video Post- Proc Camera Modem Video Display IO Interface
SoC and IP SS
5 IP Algo IP Arch SW Design Virtual Platform HW Design
6
USE (unified software environment): Multi-layer seamless context aware interface Platform VIP library ML TL modeling HL/TL verification
7 C2RTL synthesis PPA optimization HL/CA verification Transactional FEC IP Algo IP Arch Functional SW Multi-Level Transactional SS Platform
RTL Platfm Blk Design RTL SS Integration RTL SS Platform Blk Driver & API Platfm Driver Platfm API Synth C Blk / uP Descript.
8
C to RTL Synthesis & PPA Optimization Process Independent; Scheduling; Resource Sharing; Top-down Hier. Synthesis Bottom-Up Synthesis; ECO C++/SystemC/RTL; Area/Timing Estimation; Untimed synth “C” RTL Sequential Clock Gating; Memory Gating; Vt and Power Domain Aware; Power Estimation; TV uProcessor Synthesis
uP Description
ISS Compiler TC TLM RTL
9
TS TS TS TS TS TS TS TS TS/TLM TS/TLM
approach
10
Functional SW Multi-Level Transactional SS Platform
RTL Datapath Design uP RTL Design RTL SS Platform Driver & API Platfm Driver Platfm API Synth C Blks Input Stimulus Test Vectors
existing RTL methodology
11 C2RTL Synthesis
Synth C Model C Inline Coverage Model Offline RTL Coverage Model RTL Code
Automated RTL P/F and Cov Verification Relevant observation points are always either in IO library or as C variables
Transactors
C Testbench
RTL Simulator C Simulations P/F and Cov
Transactional C2RTL FEC
Transactors
Automated C2RTL Transactional FEC
Ref C Model
12 Synthesis Verification Modeling Design Algorithm Design Architecture HW/SW Partitions AlgoC Model Ref C Model SynC Model uP ISA Model uP ISS Model IPQ Validation
Synth RTL Core
Algo Evaluation ISA Verification VC Model
IP Charac- terization
uP RTL Verification Synth RTL Verification
RefC/SynC Verification An integrated algo design & hw/sw modeling process IP Tuning & U.C. Test
uP RTL Core
eAlgo SW
13
C2RTL Synthesis
IP SS Virtual Platform SW Dev Ref C Design System Validation
Traffic Performance Analysis SoC Virtual Platform System SW Dev
timing power
Other TLM IP models Transactor Syn C Design C Cov Model RTL Design
RTL Cov Model
TLM Wrapper
14
Design Ref C LoC Synth C LoC Area Hand RTL LoC Area Speed (MHz) Power
Block1 1263 1794 51.72 5152 51.26 266 Block2 1954 3722 296,436 10988 371,855 400/200 Block3 4281 12578 686,567 38110 400 Block4 3155 9498 774,726 28780 400 Block1: apple to apple comparison using the same micro-architecture Block2: HLS design uses micro-architecture optimization to increase the clock speed and decrease the area while keeping the same throughput Block3,4: LoC for Hand RTL is extrapolated (3x). It is beyond a single designer’s reach within a reasonable time period
15