STREAM FPGA Benchmark Results

The benchmark results are given divided by the used version of the benchmark, since internal changes in the benchmark code might lead to different performance results. All performance measures are given in GB/s and the measurements were done with data arrays of 2GB. The measurements were executed 10 times and the best result is published.

The results and the used configuration is given in Table 1 and are also available as CSV.

Fig. 2 Radar plot of the normalized measurement results for all benchmarked FPGA boards

Table 1 STREAM FPGA Benchmark Results

Version	2.6	2.6	2.3	2.3	2.3	2.3
FPGA board	Alveo U280	Bittware 520N	BittWare 520N	Alveo U280	Alveo U280	PAC D5005
FPGA	Xilinx XCU280	Intel Stratix 10 GX2800	Intel Stratix 10 GX2800	Xilinx XCU280	Xilinx XCU280	Intel Stratix 10 SX
Memory Type	DDR	DDR	DDR	DDR	HBM2	SVM
SDK	2019.2	21.2.0	19.4.0	2019.2	2019.2	19.4.0
BSP/Shell	2019.2.3	20.4.0_hpc	19.2.0_hpc	2019.2.3	2019.2.3	18.1.2_svm
CPU	AMD EPYC Milan 7763	AMD EPYC Milan 7763	Intel Xeon Gold 6148	Intel Xeon Gold 6148	Intel Xeon Gold 6148	Intel Xeon Gold 6148
System	Noctua 2	Noctua 2	Noctua 1	Noctua 1	Noctua 1	Noctua 1
DATA_TYPE	float	float	float	float	float	float
VECTOR_COUNT	16	16	16	16	16	16
GLOBAL_MEM_UNROLL	1	1	1	1	1	1
DEVICE_BUFFER_SIZE	16384	32768	4096	16384	2048	1
NUM_REPLICATIONS	2	4	4	2	32	1
LUT	188124	178268	176396	20832	331904	103628
LUT percent	14.44	19	25.0	1.9	20.69	12.0
Register	298365	297342	449231	39002	574976	244354
Register percent	11.45		25.0	1.39	27.24	12.0
BRAM	853.5	3926	4029	558	1408	548
BRAM percent	42.43	33	34.0	34.19	77.7	5.0
DSP	170	128	128	160	2560	32
DSP percent	1.88	2	2.0	1.78	28.38	1.0
Frequency	411.015198	342.23	316.67	300.0	370.0	346.0
Copy (GB/s)	32.98	65.63	67.01	33.94	377.42	20.15
Scale (GB/s)	32.98	65.63	67.24	33.92	365.8	20.04
Add (GB/s)	33.88	67.78	68.9	34.58	374.03	15.04
Triad	33.89	67.80	68.9	34.57	378.88	15.12
PCIe Read	6.35	6.28	6.41	5.68	6.66	inf
PCIe Write	4.00	5.87	6.32	5.47	6.03	inf