RVCore Project

User Tools

Site Tools

Trace:
m1ezv
no way to compare when less than two revisions

Differences

This shows you the differences between two versions of the page.


—m1ezv [2024/12/09 01:06] (current) – [Implementation and execution on a FPGA board] 94.103.125.62
Line 1: Line 1:
 +# RVCore Project, Arch Lab, Tokyo Tech
 +
 +The RVCore Project is a research and development project
 +of the RISC-V soft processor highly optimized for FPGAs.
 +
 +**RVCoreP** (**R**ISC-**V** **Core** **P**ipelined version) is one of the RISC-V soft processor cores of the RVCore Project.
 +It is an optimized RISC-V soft processor of five-stage pipelining.
 +
 +{{:rvcorep.png?nolink&600|}}
 +
 +**RVCoreP supports the following FPGA boards!**
 +
 +- [[https://reference.digilentinc.com/reference/programmable-logic/nexys-4-ddr/reference-manual|Nexys 4 DDR board]] with Xilinx Artix-7 FPGA
 +- [[https://reference.digilentinc.com/reference/programmable-logic/arty-a7/reference-manual|Arty A7-35T board]] with Xilinx Artix-7 FPGA
 +
 +
 +
 +## What's new
 +
 +- 2020/06/25 : Released Ver.0.5.3 and support the Arty A7-35T FPGA board
 +- [[old2|2020/05/18]] : Add [[binary|the page]] how to build the RISC-V cross compiler and RISC-V binary files
 +- 2020/05/17 : Change of web page structure and release of Ver.0.5.1
 +- [[old1|2020/05/03]] : Added the setting method about New-line code
 +- 2020/03/04 : This page is released about Ver.0.4.6 !
 +
 +
 +
 +## About RVCoreP
 +
 +The main specifications of RVCoreP are shown below:
 +
 +- An optimized RISC-V soft processor
 +- Adopt **RV32I** of RISC-V as an instruction set architecture, which is the basic 32-bit integer instruction set
 +- Adopt **five-stage pipelining**
 +  - Instruction fetch (If)
 +  - Instruction decode (Id)
 +  - Instruction execution (Ex)
 +  - Memory access (Ma)
 +  - Write back (Wb)
 +- Apply **three effective optimization methods** to improve the operating frequency
 +  - Instruction fetch unit optimization including the pipelined branch prediction mechanism
 +  - ALU optimization
 +  - Data alignment and sign-extension optimization for data memory output
 +- Implemented in **Verilog HDL**
 +- Run RISC-V programs compiled with RV32I
 +  - By Verilog HDL simulation using Verilator or Icarus Verilog
 +  - On the FPGA boards including **Xilinx Artix-7 FPGA**
 +
 +## Download source file
 +
 +The latest version of RVCoreP is Ver.0.5.3: {{ :rvcorep_ver053.zip |rvcorep_ver053.zip}}
 +
 +The other old versions are as follows:
 +
 +- Ver.0.5.1: {{ :rvcorep_ver051.zip |rvcorep_ver051.zip}}
 +- Ver.0.4.6: {{ :rvcorep_ver046.zip |rvcorep_ver046.zip}}
 +
 +This source code is released under the MIT License, see LICENSE.txt.
 +
 +### Change log
 +
 +- Ver.0.5.3 : The version that supports Arty A7-35T FPGA board
 +- Ver.0.5.1 : The version supporting Verilator, Embench, pySerial, Vivado 2019.2
 +- Ver.0.4.6 : The version used in our submitted manuscript
 +
 +
 +
 +
 +## Recommended environment
 +
 +- Ubuntu 18.04 LTS
 +- [[https://www.veripool.org/wiki/verilator|Verilator]] for Verilog HDL simulation
 +- [[http://iverilog.icarus.com|Icarus Verilog]] for Verilog HDL simulation
 +- [[https://www.xilinx.com/products/design-tools/vivado.html|Xilinx Vivado Design Suite]] 2019.2 for logic synthesis
 +- [[https://reference.digilentinc.com/reference/programmable-logic/nexys-4-ddr/reference-manual|Nexys 4 DDR board]] or [[https://reference.digilentinc.com/reference/programmable-logic/arty-a7/reference-manual|Arty A7-35T board]] with Xilinx Artix-7 FPGA for placement and routing
 +- Python 3.6.9
 +- [[https://pythonhosted.org/pyserial/|pySerial]] for serial communication with FPGA board
 +
 +### Install command
 +
 +Install verilator by the following command.
 +
 +```
 +$ sudo apt install verilator
 +```
 +
 +Install iverilog by the following command.
 +
 +```
 +$ sudo apt install iverilog
 +```
 +
 +Install pySerial by the following command.
 +
 +```
 +$ sudo apt install python3-pip
 +$ pip3 install pyserial
 +```
 +
 +
 +
 +
 +## Getting started guide
 +
 +(1) Download the source code of the RVCoreP
 +
 +```
 +$ wget http://www.arch.cs.titech.ac.jp/wk/rvcore/lib/exe/fetch.php?media=rvcorep_ver053.zip -O rvcorep_ver053.zip
 +```
 +
 +(2) Extract the downloaded zip file
 +
 +```
 +$ unzip rvcorep_ver053.zip
 +$ cd rvcorep_ver053
 +```
 +
 +### Verilog HDL simulation using Verilator
 +
 +You execute the following commands on the recommended environment.
 +
 +(1) Compile source code written in Verilog HDL using Verilator
 +
 +```
 +$ make
 +verilator -DSERIAL_WCNT=2 -DNO_IP --public --top-module top --clk CLK --x-assign 0 --x-initial 0 --no-threads -O2 -Wno-WIDTH -Wno-UNSIGNED --exe sim.cpp --cc top.v main.v uart.v debug.v proc.v
 +make -j -C obj_dir -f Vtop.mk Vtop
 +cp obj_dir/Vtop simv
 +```
 +
 +The executable file `simv` is generated after the compilation is performed.
 +
 +(2) Execute the Verilog HDL simulation 
 +
 +By default, the test benchmark is executed.
 +The memory file of the test benchmark is `test/test.mem`.
 +
 +### Execution result when running `test/test.mem`
 +
 +```
 +$ make run
 +./simv
 +Run test/test.mem
 +Initializing : ..........
 +--------------------------------------------------
 +---- nqueen ----
 +Nqueen :
 +N = 6
 +The number of solutions = 4
 +----------------
 +
 +---- qsort  ----
 +Sorted Seqence :
 +59321
 +A4C86
 +AC7D3
 +B210A
 +142044
 +1DEC15
 +1EC216
 +2536B2
 +278BCF
 +34A2AC
 +----------------
 +
 +---- fib    ----
 +Fibonacci Seqence :
 +1: 1
 +2: 1
 +3: 2
 +4: 3
 +5: 5
 +6: 8
 +7: D
 +8: 15
 +9: 22
 +A: 37
 +----------------
 +
 +---- acker  ----
 +acker(0,0) = 1
 +acker(0,1) = 2
 +acker(0,2) = 3
 +acker(1,0) = 2
 +acker(1,1) = 3
 +acker(1,2) = 4
 +acker(2,0) = 3
 +acker(2,1) = 5
 +acker(2,2) = 7
 +----------------
 +
 +== elapsed clock cycles              :            35030
 +== valid instructions executed       :            28934
 +== IPC                               :            0.825
 +== branch prediction hit             :             3615
 +== branch prediction miss            :             1430
 +== branch prediction total           :             5045
 +== branch prediction hit rate        :            0.716
 +== the num of load-use stall         :             1897
 +== estimated clock cycles            :            35121
 +== r_cnt                             :         000088d6
 +== r_rout                            :         000000a0
 +- top.v:154: Verilog $finish
 +```
 +
 +You will see the above output.
 +The information such as IPC (Instructions Per Cycle) and branch prediction hit rate is output to the console after running simulation.
 +
 +
 +
 +(3) Execute the Dhrystone and Coremark benchmarks by the Verilog HDL simulation
 +
 +You compile and execute with Dhrystone and Coremark benchmarks.
 +
 +The memory files, the binary files, and the dump files are stored in the `bench/` directory.
 +
 +For Dhrystone, three configurations are prepared by the parameter `NUMBER_OF_RUNS`.
 +
 +- benchd  : NUMBER\_OF\_RUNS=500
 +- benchd2 : NUMBER\_OF\_RUNS=2000
 +- benchd3 : NUMBER\_OF\_RUNS=10000
 +
 +For Coremark, three configurations are also prepared by the parameter `ITERATIONS`.
 +
 +- benchc  : ITERATIONS=1
 +- benchc2 : ITERATIONS=2
 +- benchc3 : ITERATIONS=10
 +
 +These benchmarks can be run with the following commands.
 +
 +```
 +$ make [configuration name]
 +$ make run
 +```
 +
 +(4) Execute the [[https://embench.org|Embench]] benchmark by the Verilog HDL simulation
 +
 +You compile and execute with 19 benchmarks in Embench.
 +
 +The memory files, the binary files, and the dump files are stored in the `embench/` directory.
 +
 +These benchmarks can be run with the following commands.
 +
 +```
 +$ make [benchmark name]
 +$ make run
 +```
 +
 +The execution results of dhrystone3, coremark3, and 19 benchmarks of Embench are summarized in the file `result.txt`.
 +
 +
 +### Implementation and execution on a FPGA board
 +
 +You execute the following process in the directory of the downloaded source code on the recommended environment.
 +
 +(1) Open the project file `main.xpr` in Xilinx Vivado
 +
 +```
 +$ vivado main.xpr &
 +```
 +
 +(2) Perform logic synthesis, placement and routing, and generating bitstream using Xilinx Vivado
 +
 +Use the following set of logic synthesis and place-and-route according to the FPGA board you want to use.
 +
 +- When using Nexys 4 DDR board : synth\_1 and impl\_1
 +- When using Arty A7-35T board : synth\_2 and impl\_2
 +
 +Execute the following process according to the used set of logic synthesis and placement and routing.
 +
 +- Right click on synth_* for logic synthesis and select "Make Active"
 +- Click "Generate Bitstream" in Vivado project manager
 +
 +By default, the operating frequency of the processor is set to 160MHz.
 +
 +In the synth\_2 for Arty A7 board,
 +the following command is added to the option of synth\_2, the strategy of logic synthesis.
 +
 +- `-verilog_define ARTYA7=1`
 +
 +The source code is switched for Arty A7 board by the definition of this macro `ARTYA7`,
 +so please be careful when changing the logic synthesis strategy yourself.
 +
 +(3) Write the generated bitstream to the FPGA board
 +
 +- Click "Open Hardware Manager" in Vivado project manager to open the hardware manager
 +
 +- Click "Open target" and "Auto Connect" to recognize the FPGA board
 +
 +- Click "Program device" and specify Bitstream file
 +
 +- Click "Program" to write bitstream to FPGA board
 +
 +When the bitstream data is correctly written to the Nexys 4 DDR board,
 +the DONE LED lights up and "00000000" is displayed on the 8-digit 7-segment LEDs.
 +
 +(4) Prepare for 8M baud serial communication
 +
 +Please move to the `test/` directory for the following processes.
 +
 +```
 +$ cd test/
 +```
 +
 +For serial communication with RVCoreP, the pySerial program `serial_rvcorep.py` is used.
 +
 +This program sends a binary file to the FPGA board and outputs the received characters via pySerial.
 +
 +Before executing this program, check the setting of the port used for serial communication in the program.
 +
 +By default, `/dev/ttyUSB1` is set.
 +
 +The usage of this program is as follows:
 +
 +```
 +$ python3 serial_rvcorep.py [serial baud rate (Mbaud)] [binary file name]
 +```
 +
 +(5) Send the RISC-V program binary to the FPGA board by serial communication and execute the program
 +
 +To send the test program `test.bin` and start serial communication, execute one of the following commands.
 +
 +```
 +$ python3 serial_rvcorep.py 8 "test.bin"
 +```
 +
 +or
 +
 +```
 +$ python3 serial_rvcorep.py
 +```
 +
 +- Send the test benchmark to the FPGA board via serial communication and execute the program on the implemented processor
 +
 +The following execution result is output via serial communication.
 +
 +### Execution result when running `test.bin`
 +
 +```
 +$ python3 serial_rvcorep.py
 +serial baud rate : 8000000
 +send file : test.bin
 +---- nqueen ----
 +Nqueen :
 +N = 6
 +The number of solutions = 4
 +----------------
 +
 +---- qsort  ----
 +Sorted Seqence :
 +59321
 +A4C86
 +AC7D3
 +B210A
 +142044
 +1DEC15
 +1EC216
 +2536B2
 +278BCF
 +34A2AC
 +----------------
 +
 +---- fib    ----
 +Fibonacci Seqence :
 +1: 1
 +2: 1
 +3: 2
 +4: 3
 +5: 5
 +6: 8
 +7: D
 +8: 15
 +9: 22
 +A: 37
 +----------------
 +
 +---- acker  ----
 +acker(0,0) = 1
 +acker(0,1) = 2
 +acker(0,2) = 3
 +acker(1,0) = 2
 +acker(1,1) = 3
 +acker(1,2) = 4
 +acker(2,0) = 3
 +acker(2,1) = 5
 +acker(2,2) = 7
 +----------------
 +```
 +
 +When using Nexys 4 DDR board, The 7-segment LED shows the value of the program counter of the processor at the end of execution.
 +If you execute the test benchmark `test/test.bin`, the output to the 7-segment LEDs is `000000A0`.
 +
 +When using Nexys 4 DDR board and the button "BTNU" is pressed, the 7-segment LED shows the number of execution cycles.
 +If you execute the test benchmark `test/test.bin`, the output　to the 7-segment LEDs is `000088D6`.
 +
 +
 +This program is terminated by Ctrl-C.
 +
 +If you want to send the binary file of the program to the FPGA board again and execute it, proceed from step (3).
 +
 +
 +
 +
 +## How to build the RISC-V cross compiler and RISC-V binary files that works with RVCoreP
 +
 +Please refer the fllowing page: [[binary|How to build the RISC-V cross compiler and RISC-V binary files that works with RVCoreP]]
 +
 +
 +
 +## Publication
 +
 +- [[https://www.jstage.jst.go.jp/article/transinf/E103.D/12/E103.D_2020PAP0015/_article|Hiromu MIYAZAKI, Takuto KANAMORI, Md Ashraful ISLAM, Kenji KISE, RVCoreP: An Optimized RISC-V Soft Processor of Five-Stage Pipelining, IEICE Transactions on Information and Systems, 2020, Volume E103.D, Issue 12, Pages 2494-2503, Released December 01, 2020, Online ISSN 1745-1361, Print ISSN 0916-8532.]]
 +
 +- Hiromu Miyazaki, Takuto Kanamori, Md Ashraful Islam, Kenji Kise: RVCoreP : An optimized RISC-V soft processor of five-stage pipelining, [[https://arxiv.org/abs/2002.03568|arXiv:2002.03568 [cs.AR]]] (2020-02-10).
 +
 +
 +
 +## Contact
 +
 +[[http://www.arch.cs.titech.ac.jp/|Kise Laboratory]], Department of Computer Science, School of Computing, [[https://www.titech.ac.jp/english/|Tokyo Institute of Technology]] (Tokyo Tech)
 +
 +Maintainer : (E-mail) riscv-support (at) arch.cs.titech.ac.jp
 +
 +Contributor : Hiromu Miyazaki, Takuto Kanamori, Md Ashraful Islam, Kenji Kise
 +
 +
 +
 +## Other Project
 +
 +- [[http://www.arch.cs.titech.ac.jp/wk/rvsoc/doku.php|RVSoC]]
 +
 +Copyright (c) 2020 Kise Laboratory, Tokyo Institute of Technology
 +
 +
 +
 +
 +
 +
 +
 +
 +
  
m1ezv.txt · Last modified: 2024/12/09 01:06 by 94.103.125.62