Nios® V Processor Reference Manual

ID 683632
Date 5/26/2023
Public

A newer version of this document is available. Customers should click here to go to the newest version.

Document Table of Contents

3.1. Processor Performance Benchmarks

Table 43.   Nios® V/g Processor Performance Benchmarks in Intel FPGA Devices for Intel® Quartus® Prime Pro Edition Software
FPGA Used fMAX (MHz) Logic Size (ALM) Architecture Performance
DMIPS/MHz Ratio CoreMark/MHz Ratio
Intel® Cyclone® 10 233.71 2511 0.866 1.467
Intel® Arria® 10 240.79 2504
Intel® Stratix® 10 272.70 2556
Intel Agilex® 7 321.50 2480
Table 44.  Benchmark Parameters for Intel® Quartus® Prime Pro Edition Software
Parameter Settings/Description
Intel® Quartus® Prime seed Maximum performance result are based on 10 seed sweep from Intel® Quartus® Prime Pro Edition software version 23.1.
Device speed grade Fastest speed grade from each Intel FPGA device family.
Defined peripherals
  • Nios® V/g processor core (without debug module, 4 KB instruction cache, and 4 KB data cache).
  • 128 KB on-chip memory for the instruction and data bus.
  • JTAG UART Intel® FPGA IP.
Toolchain Version
  • riscv32-unknown-elf-gcc (GCC) version 12.1.0
  • CMake Version: 3.23.2
Compiler configuration
  • Compiler flags: -03
  • Assembler options: -Wa -gdwarf2
  • Compile options: -Wall -Wformat-security -march=rv32ima_zicbom -mabi=ilp32
Intel uses the same Intel® Quartus® Prime design example for maximum performance benchmark(fMAX) and logic size benchmarks. However, the compiler settings are different for each benchmarks:
  • fMAX benchmark: superior_performance_optimized_placement_effort
  • Logic size benchmark: area_aggressive
Note: Results may vary depending on the version of the Intel® Quartus® Prime software, the version of the Nios® V processor, compiler version, target device and the configuration of the processor. Additionally, any changes to the system logic design might change the performance and LE usage. All results are generated from design built with Platform Designer.