Minerva - Hardware Benchmarking Tool

81 views
Skip to first unread message

Kris Gaj

unread,
Dec 5, 2017, 4:52:04 AM12/5/17
to Cryptographic competitions
Hi,

It is our pleasure to announce the release of 
     Minerva - Automated Hardware Optimization Tool,
used for benchmarking of VHDL and Verilog implementations of cryptographic algorithms, targeting most recent families of FPGA devices.

The source code and user manual of Minerva can be downloaded from:

Minerva supplements ATHENa, our older generation hardware benchmarking tool, supporting Xilinx ISE, Altera Quartus II, and Intel Quartus Prime toolsets.

The first version of Minerva aims specifically at 
                   Xilinx Vivado,
and supports Xilinx FPGAs and All Programmable Systems on Chip, at and beyond the Series 7 families: Virtex-7, Kintex-7, Artix-7, and Zynq-7000.

In particular, the Beta version of Minerva has been already used to obtain optimized results for Round 3 CAESAR Candidates, reported in the GMU Database of Results, available at
after choosing
  Family: Virtex 7, and
  Clicking on: Update.
  
Minerva is based on a heuristic optimization algorithm aimed at finding 
a) optimum requested clock frequency, and
b) best optimization strategy (one out of 25 sets of tool options, predefined in Xilinx Vivado),
leading to either 
  Optimal Throughput or Optimal Throughput-to-Area Ratio.
   
One of the important features of Minerva is its support for reproducibility of results, accomplished by creating a zip file containing all information necessary to generate identical results using just a single run of Vivado.

Additionally, Minerva takes advantage of multithreading and multi-core execution to significantly reduce the overall run time.

Our paper on Minerva has been accepted to the 2017 International Conference on Reconfigurable Computing and FPGAs, held in Cancun, Mexico, on Dec. 4-6, 2017.

At this point, you are more than welcome to try this tool by yourself! For the Quick Start instructions, please see Section 7 of the User Manual.

Any comments, questions, and suggestions for modifications & extensions are very welcome!

GMU Benchmarking Team
Reply all
Reply to author
Forward
0 new messages