Benchmark

Datasets

The dataset used in the benchmark can be downloaded from https://sparse.tamu.edu.

Performance

For representing the resource utilization in each benchmark, we separate the overall utilization into 2 parts, where P stands for the resource usage in platform, that is those instantiated in static region of the FPGA card, as well as K represents those used in kernels (dynamic region). The target device is set to Alveo U280.

Table 1 Performance on FPGA
Architecture Dataset Latency(ms) Timing LUT(P/K) BRAM(P/K) URAM(P/K) DSP(P/K)
SPMV (U280) nasa2910 0.0512565 256MHz 165.475K/220.98K 323/211 64/64 4/900

These are details for benchmark results and usage steps.

Test Overview

Here are benchmarks of the Vitis Sparse Library using the Vitis environment.

Vitis Sparse Library

  • Download code

These sparse benchmarks can be downloaded from vitis libraries master branch.

git clone https://github.com/Xilinx/Vitis_Libraries.git
cd Vitis_Libraries
git checkout master
cd sparse
  • Setup environment

Specifying the corresponding Vitis, XRT, and path to the platform repository by running following commands.

source <intstall_path>/installs/lin64/Vitis/2021.1_released/settings64.sh
source /opt/xilinx/xrt/setup.sh
export PLATFORM_REPO_PATHS=/opt/xilinx/platforms

Python3 environment: follow the steps as per https://xilinx.github.io/Vitis_Libraries/blas/2020.2/user_guide/L1/pyenvguide.html to set up Python3 environment.