Skip to main content
Journal of Big Data

Table 10 Alveo U250 platform experimental results in comparison to a software baseline and state-of-the-art FPGA works. Reported numbers are all in Seconds

From: Distributed large-scale graph processing on FPGAs

Dataset

Sequential (CPUa)

OpenMP (CPUa)

GridGraph (CPUa)

cuGraph (GPUb)

Vitis Library (FPGA)

Our work (FPGA)

LiveJournal

27.01

5.49

3.54

5.28

79.79

2.78

Web-UK-2005

275.44

185.4

34.9

90.73

N/Ac

20.6

Twitter

1443

658.5

88.5

Failedd

N/Ac

55.6

Friendster

2258

950

141

Failedd

N/Ac

95.4

  1. a CPU details are described in Table 9
  2. b The GPU used for the experiments is a NVIDIA Volta V100
  3. c N/A indicates that the mentioned study didn’t report this dataset evaluation
  4. d The experiment hit the GPU memory limit
Back to article page