Table 3 Scalability analysis: runtime overhead across different cluster scales.

From: Attention-based workload prediction and dynamic resource allocation for heterogeneous computing environments

Cluster size (nodes)

Workloads

Prediction time (ms)

Allocation time (ms)

Total cycle (ms)

Memory usage (GB)

20

100

45 ± 3

28 ± 5

73 ± 6

2.1

50

250

67 ± 4

58 ± 8

125 ± 9

3.4

100

500

98 ± 6

112 ± 12

210 ± 14

5.2

200

1000

156 ± 9

198 ± 18

354 ± 21

8.7

500

2500

287 ± 15

423 ± 35

710 ± 42

18.3