The NEORV32 Processor (RISC-V)

Project maintainers


Name: neorv32
Created: Jun 23, 2020
Updated: Jul 20, 2020
SVN Updated: May 11, 2021
SVN: Browse
Latest version: download (might take a bit to start...)
Statistics: View
Bugs: 1 reported / 1 solved
Star8you like it: star it!

Other project properties

Development status:Stable
Additional info:Design done, FPGA proven, Specification done
WishBone compliant: Yes
WishBone version: B.4
License: BSD


The NEORV32 RISC-V Processor

Processor Check riscv-arch-test license release datasheet


neorv32 Overview

The NEORV32 Processor is a customizable microcontroller-like system on chip (SoC) that is based on the RISC-V NEORV32 CPU. The processor is intended as auxiliary processor in larger SoC designs or as ready-to-go stand-alone custom microcontroller.

:books: For detailed information take a look at the NEORV32 data sheet (pdf). The asciidoc sources can be found in docs/src_adoc. The latest automatic build can be downloaded as artifacts from the Build Data Sheet GitHub workflow. The doxygen-based documentation of the software framework is available online at GitHub-pages.

:label: The project’s change log is available as in the root directory of this repository. To see the changes between stable releases visit the project's release page.

:rocket: The boards folder provides exemplary setups targeting various FPGA boards to get you started.

:spiral_notepad: Check out the project boards for a list of current ideas, TODOs, features being planned and work-in-progress.

:bulb: Feel free to open a new issue or start a new discussion if you have questions, comments, ideas or bug-fixes. Check out how to contribute.

Key Features

  • RISC-V 32-bit rv32 NEORV32 CPU, compatible to
  • Configurable RISC-V-compatible CPU extensions
    • A - atomic memory access instructions (optional)
    • B - Bit manipulation instructions (optional) :construction:
    • C - compressed instructions (16-bit) (optional)
    • E - embedded CPU (reduced register file size) (optional)
    • I - base integer instruction set (always enabled)
    • M - integer multiplication and division hardware (optional)
    • U - less-privileged user mode (optional)
    • X - NEORV32-specific extensions (always enabled)
    • Zfinx - Single-precision floating-point extensions (optional)
    • Zicsr - control and status register access instructions (+ exception/irq system) (optional)
    • Zifencei - instruction stream synchronization (optional)
    • PMP - physical memory protection (optional)
    • HPM - hardware performance monitors (optional)
  • Full-scale RISC-V microcontroller system / SoC NEORV32 Processor with optional submodules
    • optional embedded memories (instructions/data/bootloader, RAM/ROM) and caches
    • timers (watch dog, RISC-V-compatible machine timer)
    • serial interfaces (SPI, TWI, UARTs)
    • general purpose IO and PWM channels
    • external bus interface (Wishbone / AXI4)
    • dedicated NeoPixel(TM) LED interface
    • subsystem for custom co-processors
    • more ...
  • Software framework
  • Full-blown data sheet (pdf)
  • Completely described in behavioral, platform-independent VHDL - no primitives, macros, etc.
  • Fully synchronous design, no latches, no gated clocks
  • Small hardware footprint and high operating frequency

Design Principles

  • From zero to hello_world: Completely open source and documented.
  • Plain VHDL without technology-specific parts like attributes, macros or primitives.
  • Easy to use – working out of the box.
  • Clean synchronous design, no wacky combinatorial interfaces.
  • Be as small as possible – but with a reasonable size-performance trade-off.
  • Be as RISC-V-compliant as possible.
  • The processor has to fit in a Lattice iCE40 UltraPlus 5k low-power FPGA running at 22+ MHz.


The processor is synthesizable (tested on real hardware using Intel Quartus Prime, Xilinx Vivado and Lattice Radiant) and can successfully execute all the provided example programs including the CoreMark benchmark and the custom NEORV32 processor check (sw/example/cpu_test, see the status report in the according GitHub workflow).

RISC-V Architecture Tests: The processor passes the official rv32_m/C, rv32_m/I, rv32_m/M, rv32_m/privilege and rv32_m/Zifencei riscv-arch-test tests. More information regarding the NEORV32 port of the riscv-arch-test test framework can be found in riscv-arch-test/

Project componentCI status
NEORV32 processorProcessor Check
SW Framework Documentation (online at GH-pages)Doc@GitHub-pages
Build data sheet from asciidoc sourcesBuild Data Sheet
Pre-built toolchainsTest Toolchains
RISC-V architecture testriscv-arch-test


The full-blown data sheet of the NEORV32 Processor and CPU is available as pdf file: :page_facing_up: NEORV32 data sheet.

NEORV32 Processor Features

The NEORV32 Processor provides a full-scale microcontroller-like SoC based on the NEORV32 CPU. The setup is highly customizable via the processor's top generics and already provides the following optional modules:

  • processor-internal data and instruction memories (DMEM / IMEM) & cache (iCACHE)
  • bootloader (BOOTLDROM) with UART console and automatic application boot from SPI flash option
  • machine system timer (MTIME), RISC-V-compatible
  • watchdog timer (WDT)
  • two independent universal asynchronous receivers and transmitters (UART0 & UART1) with optional hardware flow control (RTS/CTS)
  • 8/16/24/32-bit serial peripheral interface controller (SPI) with 8 dedicated chip select lines
  • two wire serial interface controller (TWI), with optional clock-stretching, compatible to the I²C standard
  • general purpose parallel IO port (GPIO), 32xOut & 32xIn, with pin-change interrupt
  • 32-bit external bus interface, Wishbone b4 compatible (WISHBONE)
  • wrapper for AXI4-Lite Master Interface (see AXI Connectivity)
  • PWM controller with 4 channels and 8-bit duty cycle resolution (PWM)
  • ring-oscillator-based true random number generator (TRNG)
  • custom functions subsystem (CFS) for tightly-coupled custom co-processor extensions
  • numerically-controlled oscillator (NCO) with three independent channels
  • smart LED interface (NEOLED) - WS2812 / NeoPixel(c) compatible
  • system configuration information memory to check hardware configuration by software (SYSINFO)

[back to top]

NEORV32 CPU Features

The NEORV32 CPU implements the official RISC-V specifications (2.2) including a subset of the RISC-V privileged architecture specifications (1.12-draft)

More information regarding the CPU including a detailed list of the instruction set and the available CSRs can be found in the :page_facing_up: NEORV32 data sheet.

General Features

  • Modified Harvard architecture (separate CPU interfaces for data and instructions; NEORV32 processor: Single processor-internal bus via I/D mux)
  • Two stages in-order pipeline (FETCH, EXECUTE); each stage uses a multi-cycle processing scheme
  • No hardware support of unaligned accesses - they will trigger an exception
  • BIG-ENDIAN byte-order, processor's external memory interface allows endianness configuration to connect to system with different endianness
  • All reserved or unimplemented instructions will raise an illegal instruction exception
  • Privilege levels: machine mode, user mode (if enabled via U extension)
  • Official RISC-V open-source architecture ID

[back to top]

A - Atomic memory access extension

  • Supported instructions: LR.W (load-reservate) SC.W (store-conditional)

[back to top]

B - Bit manipulation instructions extension

  • :construction: work-in-progress :construction:
  • :warning: this extension has not been officially ratified yet!
  • :books: more information can be found here: RISC-V B spec.
  • Compatible to v0.94-draft of the bit manipulation spec
  • Software support via intrinsic library (see sw/example/bit_manipulation)
  • Zbb base instruction set: CLZ CTZ CPOP SEXT.B SEXT.H MIN[U] MAX[U] ANDN ORN XNOR ROL ROR[I] zext(pseudo-instruction for PACK rd, rs, zero) rev8(pseudo-instruction for GREVI rd, rs, -8) orc.b(pseudo-instruction for GORCI rd, rs, 7)
  • Zbs single-bit instructions: SBSET[I] SBCLR[I] SBINV[I] SBEXT[I]
  • Zba shifted-add instructions: SH1ADD SH2ADD SH3ADD

[back to top]

C - Compressed instructions extension

  • Jump and branch instructions: C.J C.JAL C.JR C.JALR C.BEQZ C.BNEZ
  • Memory instructions: C.LW C.SW C.LWSP C.SWSP
  • System instructions: C.EBREAK (requires Zicsr extension)
  • Pseudo-instructions are not listed

[back to top]

E - Embedded CPU version extension

  • Reduced register file (only the 16 lowest registers are implemented)

[back to top]

I - Base integer instruction set

  • ALU instructions: LUI AUIPC ADD[I] SLT[I][U] XOR[I] OR[I] AND[I] SLL[I] SRL[I] SRA[I] SUB
  • Jump and branch instructions: JAL JALR BEQ BNE BLT BGE BLTU BGEU
  • Memory instructions: LB LH LW LBU LHU SB SH SW
  • System instructions: ECALL EBREAK FENCE
  • Pseudo-instructions are not listed

[back to top]

M - Integer multiplication and division hardware extension

  • Multiplication instructions: MUL MULH MULHSU MULHU
  • Division instructions: DIV DIVU REM REMU
  • By default, the multiplier and divider cores use an iterative bit-serial processing scheme
  • Multiplications can be mapped to DSPs via the FAST_MUL_EN generic to increase performance

[back to top]

U - Privileged architecture - User mode extension

  • Requires Zicsr extension
  • Privilege levels: M (machine mode) + less-privileged U (user mode)

[back to top]

X - NEORV32-specific CPU extensions

  • The NEORV32-specific extensions are always enabled and are indicated via the X bit set in the misa CSR.
  • 16 fast interrupt request channels with according control/status bits in mie and mip and custom exception codes in mcause
  • mzext CSR to check for implemented Z* CPU extensions (like Zifencei)
  • All undefined/umimplemented/malformed/illegal instructions do raise an illegal instruction exception

[back to top]

Zfinx - Single-precision floating-point extension

  • :warning: this extension has not been officially ratified yet!
  • :books: more information can be found here: RISC-V Zfinx spec.
  • Software support via intrinsic library (see sw/example/floating_point_test)
  • Fused multiply-add instructions (F[N]MADD.S & F[N)MSUB.S) are not supported!
  • Computational instructions: FADD.S FSUB.S FMUL.S FSGNJ[N/X].S FCLASS.S FDIV.S FSQRT.S
  • Comparison instructions: FMIN.S FMAX.S FEQ.S FLT.S FLE.S
  • Conversion instructions: FCVT.W.S FCVT.WU.S FCVT.S.W FCVT.S.WU
  • Additional CSRs: fcsr frm fflags

[back to top]

Zicsr - Privileged architecture - CSR access extension

  • Privilege levels: M-mode (Machine mode)
  • CSR access instructions: CSRRW[I] CSRRS[I] CSRRC[I]
  • System instructions: MRET WFI
  • Pseudo-instructions are not listed
  • Counter CSRs: [m]cycle[h] [m]instret[m] time[h] [m]hpmcounter*[h](3..31, configurable) mcounteren mcountinhibit mhpmevent*(3..31, configurable)
  • Machine CSRs: mstatus[h] misa(read-only!) mie mtvec mscratch mepc mcause mtval mip mvendorid marchid mimpid mhartid mzext(custom)
  • Supported (sync.) exceptions (implementing the RISC-V specs):
    • Misaligned instruction address
    • Instruction access fault (via timeout/error after unacknowledged bus access)
    • Illegal instruction
    • Breakpoint (via ebreak instruction)
    • Load address misaligned
    • Load access fault (via timeout/error after unacknowledged bus access)
    • Store address misaligned
    • Store access fault (via unacknowledged bus access after timeout)
    • Environment call from U-mode (via ecall instruction in user mode)
    • Environment call from M-mode (via ecall instruction in machine mode)
  • Supported interrupts:
    • RISC-V non-maskable interrupt nmi (via external signal)
    • RISC-V machine timer interrupt mti (via processor-internal MTIME unit or external signal)
    • RISC-V machine software interrupt msi (via external signal)
    • RISC-V machine external interrupt mei (via external signal)
    • 16 fast interrupt requests, 6+1 available for custom usage

[back to top]

Zifencei - Instruction stream synchronization extension

  • System instructions: FENCE.I (among others, used to clear and reload instruction cache)

[back to top]

PMP - Privileged architecture - Physical memory protection

  • Requires Zicsr extension
  • Configurable number of regions (0..63)
  • Additional machine CSRs: pmpcfg*(0..15) pmpaddr*(0..63)

[back to top]

HPM - Privileged architecture - Hardware performance monitors

  • Requires Zicsr extension
  • Configurable number of counters (0..29)
  • Additional machine CSRs: mhpmevent*(3..31) [m]hpmcounter*[h](3..31)

[back to top]

:warning: Non-RISC-V-Compatible Issues and Limitations

  • CPU and Processor are BIG-ENDIAN, but this should be no problem as the external memory bus interface provides big- and little-endian configurations
  • misa CSR is read-only - no dynamic enabling/disabling of synthesized CPU extensions during runtime; for compatibility: write accesses (in m-mode) are ignored and do not cause an exception
  • mip CSR is read-only - pending IRQs can be cleared using mie
  • The physical memory protection (PMP) only supports NAPOT mode yet and a minimal granularity of 8 bytes
  • The A extension only implements lr.w and sc.w instructions yet. However, these instructions are sufficient to emulate all remaining AMO operations

[back to top]

FPGA Implementation Results


This chapter shows exemplary implementation results of the NEORV32 CPU for an Intel Cyclone IV EP4CE22F17C6N FPGA on a DE0-nano board. The design was synthesized using Intel Quartus Prime Lite 20.1 ("balanced implementation"). The timing information is derived from the Timing Analyzer / Slow 1200mV 0C Model. If not otherwise specified, the default configuration of the CPU's generics is assumed (e.g. no physical memory protection, no hardware performance monitors). No constraints were used at all.

Results generated for hardware version

CPU ConfigurationLEsFFsMemory bitsDSPs (9-bit)f_max
rv32i98040910240123 MHz
rv32i + Zicsr183585610240124 MHz
rv32im + Zicsr2443113410240124 MHz
rv32imc + Zicsr2669114910240125 MHz
rv32imac + Zicsr2685115610240124 MHz
rv32imac + Zicsr + u2698116210240124 MHz
rv32imac + Zicsr + u + Zifencei2715116210240122 MHz
rv32imac + Zicsr + u + Zifencei + Zfinx4004181210247121 MHz

Setups with enabled "embedded CPU extension" E show the same LUT and FF utilization and identical f_max as the according I configuration. However, the size of the register file is cut in half.

[back to top]

NEORV32 Processor-Internal Peripherals and Memories

Results generated for hardware version (mandatory core modules in bold).

ModuleDescriptionLEsFFsMemory bitsDSPs (9-bit)
Boot ROMBootloader ROM (4kB)31327680
BUSKEEPERProcessor-internal bus monitor11600
BUSSWITCHBus mux for CPU instr. and data interface49800
CFSCustom functions subsystem----
DMEMProcessor-internal data memory (8kB)182655360
GPIOGeneral purpose input/output ports676500
iCACHEInstruction cache (1x4 blocks, 256 bytes per block)22015481920
IMEMProcessor-internal instruction memory (16kB)621310720
MTIMEMachine system timer28920000
NCONumerically-controlled oscillator25422600
NEOLEDSmart LED Interface (NeoPixel/WS28128) 4xFIFO34730900
PWMPulse_width modulation controller716900
SPISerial peripheral interface13812400
SYSINFOSystem configuration information memory101000
TRNGTrue random number generator13210500
TWITwo-wire interface774400
UART0/1Universal asynchronous receiver/transmitter 0/117613200
WDTWatchdog timer604500
WISHBONEExternal memory interface12910400

[back to top]

NEORV32 Processor - Exemplary FPGA Setups

:information_source: Check out the boards folder for exemplary setups targeting various FPGA boards.

The following tables show exemplary processor implementation results for different FPGA platforms. The processor setups use the default peripheral configuration (like no CFS and no TRNG), no external memory interface and only internal instruction and data memories. IMEM uses 16kB and DMEM uses 8kB memory space.

Results generated for hardware version

VendorFPGABoardToolchainCPU ConfigurationLUT / LEFF / REGDSP (9-bit)Memory BitsBRAM / EBRSPRAMFrequency
IntelCyclone IV EP4CE22F17C6NTerasic DE0-NanoQuartus Prime Lite 20.1rv32imc + u + Zicsr + Zifencei3813 (17%)1904 (8%)0 (0%)231424 (38%)--119 MHz
LatticeiCE40 UltraPlus iCE40UP5K-SG48IUpduino v2.0Radiant 2.1 (Synplify Pro)rv32ic + u + Zicsr + Zifencei4397 (83%)1679 (31%)0 (0%)-12 (40%)4 (100%)c 22.15 MHz
XilinxArtix-7 XC7A35TICSG324-1LArty A7-35TVivado 2019.2rv32imc + u + Zicsr + Zifencei + PMP2465 (12%)1912 (5%)0 (0%)-8 (16%)-c 100 MHz


  • The "default" implementation strategy of the according toolchain is used.
  • The Lattice iCE40 UltraPlus setup uses the FPGA's SPRAM memory primitives for the internal IMEM and DMEM (each 64kb).
  • The clock frequencies marked with a "c" are constrained clocks. The remaining ones are f_max results from the place and route timing reports.
  • The Upduino and the Arty board have on-board SPI flash memories for storing the FPGA configuration. These device can also be used by the default NEORV32 bootloader to store and automatically boot an application program after reset (both tested successfully).
  • The setups with PMP implement 2 regions with a minimal granularity of 64kB.
  • No HPM counters are implemented.

[back to top]


CoreMark Benchmark

The CoreMark CPU benchmark was executed on the NEORV32 and is available in the sw/example/coremark project folder. This benchmark tests the capabilities of a CPU itself rather than the functions provided by the whole system / SoC.

Hardware:       32kB IMEM, 8kB DMEM, no caches, 100MHz clock
CoreMark:       2000 iterations, MEM_METHOD is MEM_STACK
Compiler:       RISCV32-GCC 10.1.0 (rv32i toolchain)
Compiler flags: default, see makefile
Optimization:   -O3
Peripherals:    UART for printing the results

Results generated for hardware version

CPU (including Zicsr extension)Executable SizeOptimizationCoreMark ScoreCoreMarks/MHz
rv32i28 756 bytes-O336.360.3636
rv32imc22 008 bytes-O368.970.6897
rv32imc + FAST_MUL_EN + FAST_SHIFT_EN22 008 bytes-O390.910.9091

The FAST_MUL_EN configuration uses DSPs for the multiplier of the M extension (enabled via the FAST_MUL_EN generic). The FAST_SHIFT_EN configuration uses a barrel shifter for CPU shift operations (enabled via the FAST_SHIFT_EN generic).

When the C extension is enabled, branches to an unaligned uncompressed instruction require additional instruction fetch cycles.

[back to top]

Instruction Cycles

The NEORV32 CPU is based on a two-stages pipelined architecutre. Each stage uses a multi-cycle processing scheme. Hence, each instruction requires several clock cycles to execute (2 cycles for ALU operations, ..., 40 cycles for divisions). The average CPI (cycles per instruction) depends on the instruction mix of a specific applications and also on the available CPU extensions. By default the CPU-internal shifter (e.g. for the SLL instruction) as well as the multiplier and divider of the M extension use a bit-serial approach and require several cycles for completion.

The following table shows the performance results for successfully running 2000 CoreMark iterations, which reflects a pretty good "real-life" work load. The average CPI is computed by dividing the total number of required clock cycles (only the timed core to avoid distortion due to IO wait cycles; sampled via the cycle[h] CSRs) by the number of executed instructions (instret[h] CSRs). The executables were generated using optimization -O3.

Results generated for hardware version

CPU (including Zicsr extension)Required Clock CyclesExecuted InstructionsAverage CPI
rv32i5 595 750 5031 466 028 6073.82
rv32imc2 981 786 734611 814 9184.87
rv32imc + FAST_MUL_EN + FAST_SHIFT_EN2 265 135 174611 814 9483.70

The FAST_MUL_EN configuration uses DSPs for the multiplier of the M extension (enabled via the FAST_MUL_EN generic). The FAST_SHIFT_EN configuration uses a barrel shifter for CPU shift operations (enabled via the FAST_SHIFT_EN generic).

When the C extension is enabled branches to an unaligned uncompressed instruction require additional instruction fetch cycles.

[back to top]

Top Entities

The top entity of the NEORV32 Processor (SoC) is rtl/core/neorv32_top.vhd, which provides a Wishbone b4-compatoible bus interface.

:information_source: It is recommended to use the processor setup even if you want to use the CPU in stand-alone mode. Simply disable all the processor-internal modules via the generics and you will get a "CPU wrapper" that already provides a minimal CPU environment and an external memory interface (like AXI4). This setup also allows to further use the default bootloader and software framework. From this base you can start building your own processor system.

Use the top's generics to configure the system according to your needs. Each generic is initilized with the default configuration. Detailed information regarding the interface signals and configuration generics can be found in the :page_facing_up: NEORV32 data sheet (pdf).

All signals of the top entity are of type std_ulogic or std_ulogic_vector, respectively (except for the processor's TWI signals, which are of type std_logic). Leave all unused output ports unconnected and tie all unused input ports to zero.

Alternative top entities, like the simplified "hello world" test setup or CPU/Processor wrappers with resolved port signal types (i.e. std_logic), can be found in rtl/top_templates.

[back to top]

AXI4 Connectivity

Via the rtl/top_templates/neorv32_top_axi4lite.vhd wrapper the NEORV32 provides an AXI4-Lite compatible master interface. This wrapper instantiates the default NEORV32 processor top entitiy and implements a Wishbone to AXI4-Lite bridge.

The AXI4-Lite interface has been tested using Xilinx Vivado 19.2 block designer:


The processor was packed as custom IP using neorv32_top_axi4lite.vhd as top entity. The AXI interface is automatically detected by the packager. All remaining IO interfaces are available as custom signals. The configuration generics are available via the "customize IP" dialog. In the figure above the resulting IP block is named "neorv32_top_axi4lite_v1_0". (Note: Use Syntheiss option "global" when generating the block design to maintain the internal TWI tri-state drivers.)

The setup uses an AXI interconnect to attach two block RAMs to the processor. Since the processor in this example is configured without IMEM and DMEM, the attached block RAMs are used for storing instructions and data: the first RAM is used as instruction memory and is mapped to address 0x00000000 - 0x00003fff (16kB), the second RAM is used as data memory and is mapped to address 0x80000000 - 0x80001fff (8kB).

[back to top]

Getting Started

This overview is just a short excerpt from the Let's Get It Started section of the NEORV32 documentary:

:page_facing_up: NEORV32 data sheet

0. Build the Documentation

This step is optional since there are pre-built versions of the processor data sheet and the software documentation. If you want to build the documentation by yourself:

NEORV32 Data Sheet

To build the data sheet open a console and navigate to the project's docs folder. Run $ sh (make sure asciidoctor-pdf is installed). This will take all the asciidoc sources from docs/src_adoc to generate docs/NEORV32.pdf.

Software Framework Documentation

Make sure doxygen is installed. Open a console and navigate to the project's docs folder and run $ doxygen Doxyfile. This will create (if not already there) a new folder docs/doxygen_build/html where doxygen will generate the HTML-based documentation pages. Open docs/doxygen_build/html/files.html to get started.

1. Get the Toolchain

At first you need a RISC-V GCC toolchain. You can either download the sources and build the toolchain by yourself, or you can download a prebuilt one and install it.

To build the toolchain by yourself, follow the official build instructions. Make sure to use the ilp32 or ilp32e ABI.

Alternatively, you can download a prebuilt toolchain. I have uploaded the toolchains I am using to GitHub. These toolchains were compiled on a 64-bit x86 Ubuntu 20.04 LTS (Ubuntu on Windows, actually). Download the toolchain of choice: :octocat:

You can also use the toolchains provided by SiFive. These are 64-bit toolchains that can also emit 32-bit RISC-V code. They were compiled for more sophisticated machines (rv32imac) so make sure the according NEORV32 hardware extensions are enabled.

:warning: Keep in mind that – for instance – a rv32imc toolchain only provides library code compiled with compressed and mul/div instructions! Hence, this code cannot be executed (without emulation) on an architecture without these extensions!

To check everything works fine, make sure GNU Make and a native GCC compiler are installed. Test the installation of the RISC-V toolchain by navigating to an example program project like sw/example/blink_led and running:

neorv32/sw/example/blink_led$ make check

2. Download the NEORV32 Project

Get the sources of the NEORV32 Processor project. The simplest way is using git clone (suggested for easy project updates via git pull):

$ git clone

Alternatively, you can either download a specific release or get the most recent version of this project as *.zip file.

3. Create a new FPGA Project

:information_source: If want to use a script-based exemplary project setup check out the boards folder, which provides exemplary setups targeting various FPGA boards.

Create a new project with your FPGA design tool of choice. Add all the *.vhd files from the rtl/core folder to this project. Make sure to add these files to a new design library called neorv32.

You can either instantiate the processor's top entity or one of its wrappers in your own project. If you just want to try thing out, you can use the simple test setup (rtl/top_templates/neorv32_test_setup.vhd) as top entity.

neorv32 test setup

This test setup instantiates the processor and implements most of the peripherals and some ISA extensions. Only the UART0 communications lines, clock, reset and some GPIO output signals are propagated as actual top entity interface signals. Basically, it is a FPGA version of a "hello world" example:

  entity neorv32_test_setup is
    port (
      -- Global control --
      clk_i       : in  std_ulogic := '0'; -- global clock, rising edge
      rstn_i      : in  std_ulogic := '0'; -- global reset, low-active, async
      -- GPIO --
      gpio_o      : out std_ulogic_vector(7 downto 0); -- parallel output
      -- UART0 --
      uart0_txd_o : out std_ulogic;       -- UART0 send data
      uart0_rxd_i : in  std_ulogic := '0' -- UART0 receive data
  end neorv32_test_setup;

4. Compile an Example Program

The NEORV32 project includes several example program project from which you can start your own application. There are example programs to check out the processor's peripheral like I2C or the true-random number generator. And of course there is also a port of Conway's Game of Life available.

Simply compile one of these projects using

neorv32/sw/example/blink_led$ make clean_all exe

This will create a NEORV32 executable neorv32_exe.bin in the same folder, which you can upload via the bootloader.

5. Upload the Executable via the Bootloader

Connect your FPGA board via UART to your computer and open the according port to interface with the fancy NEORV32 bootloader. The bootloader uses the following default UART configuration:

  • 19200 Baud
  • 8 data bits
  • 1 stop bit
  • No parity bits
  • No transmission / flow control protocol (raw bytes only)
  • Newline on \r\n (carriage return & newline) - also for sent data

Use the bootloader console to upload the neorv32_exe.bin executable gerated during application compiling and run your application.

<< NEORV32 Bootloader >>

BLDV: Mar 23 2021
HWV:  0x01050208
CLK:  0x05F5E100
USER: 0x10000DE0
MISA: 0x40901105
ZEXT: 0x00000023
PROC: 0x0EFF0037
IMEM: 0x00004000 bytes @ 0x00000000
DMEM: 0x00002000 bytes @ 0x80000000

Autoboot in 8s. Press key to abort.

Available CMDs:
 h: Help
 r: Restart
 u: Upload
 s: Store to flash
 l: Load from flash
 e: Execute
CMD:> u
Awaiting neorv32_exe.bin... OK
CMD:> e

Blinking LED demo program

Going further: Take a look at the Let's Get It Started! chapter of the :page_facing_up: NEORV32 data sheet.

[back to top]


I'm always thankful for help! So if you have any questions, bug reports, ideas or if you want to give any kind of feedback, feel free to open a new issue, start a new discussion on GitHub or directly drop me a line.

Here is a simple guide line if you'd like to contribute to this repository:

  1. :star: this repository :wink:
  2. Check out the project's code of conduct
  3. Fork this repository and clone the fork
  4. Create a feature branch in your fork: git checkout -b awesome_new_feature_branch
  5. Create a new remote for the upstream repo: git remote add upstream
  6. Commit your modifications: git commit -m "Awesome new feature!"
  7. Push to the branch: git push origin awesome_new_feature_branch
  8. Create a new pull request

[back to top]

This project is released under the BSD 3-Clause license. No copyright infringement intended. Other implied or used projects might have different licensing - see their documentation to get more information.


If you are using the NEORV32 or parts of the project in some kind of publication, please cite it as follows:

S. Nolting, "The NEORV32 RISC-V Processor",

BSD 3-Clause License

Copyright (c) 2021, Stephan Nolting. All rights reserved.

Redistribution and use in source and binary forms, with or without modification, are permitted provided that the following conditions are met:

  1. Redistributions of source code must retain the above copyright notice, this list of conditions and the following disclaimer.
  2. Redistributions in binary form must reproduce the above copyright notice, this list of conditions and the following disclaimer in the documentation and/or other materials provided with the distribution.
  3. Neither the name of the copyright holder nor the names of its contributors may be used to endorse or promote products derived from this software without specific prior written permission.


Our website contains links to the websites of third parties ("external links"). As the content of these websites is not under our control, we cannot assume any liability for such external content. In all cases, the provider of information of the linked websites is liable for the content and accuracy of the information provided. At the point in time when the links were placed, no infringements of the law were recognisable to us. As soon as an infringement of the law becomes known to us, we will immediately remove the link in question.

Proprietary Notice

"Artix" and "Vivado" are trademarks of Xilinx Inc.

"Cyclone" and "Quartus Prime Lite" are trademarks of Intel Corporation.

"iCE40", "UltraPlus" and "Radiant" are trademarks of Lattice Semiconductor Corporation.

"AXI", "AXI4" and "AXI4-Lite" are trademarks of Arm Holdings plc.

"NeoPixel" is a trademark of Adafruit Industries.

[back to top]



RISC-V - Instruction Sets Want To Be Free!

Continous integration provided by :octocat: GitHub Actions and powered by GHDL.

Open Source Hardware Logo

This project is not affiliated with or endorsed by the Open Source Initiative ( /

[back to top]

Made with :coffee: in Hannover, Germany :eu: