LLM Inference on Llama 3.2 Google Pixel 9 Pro XL 1B (test)
530.02Prefill Throughput (min) (tokens/sec)ET Vulkan
Evaluation Results
| Method | Links | |||||
|---|---|---|---|---|---|---|
| ET VulkanHardware=GPU, GS=1282026.05 | 530.02 | 540.08 | 23.74 | 24.03 | 676 | |
| ET XNNPACKHardware=CPU, GS=1282026.05 | 270.61 | 296.98 | 34.24 | 35.03 | 742 | |
| ONNXHardware=CPU, GS=1282026.05 | 256.34 | 267.86 | 37.98 | 39.35 | 659 | |
| ET XNNPACKHardware=CPU, GS=322026.05 | 235.51 | 245.92 | 31.29 | 32.99 | 821 | |
| ET VulkanHardware=GPU, GS=322026.05 | 192.77 | 197.38 | 23.43 | 23.54 | 920 | |
| llama.cppHardware=CPU, GS=322026.05 | 158.1 | 202.4 | 29.6 | 30.9 | 728 | |
| ONNXHardware=CPU, GS=322026.05 | 133.25 | 156.72 | 22.55 | 27.07 | 769 | |
| LiteRTHardware=CPU, GS=1282026.05 | 117 | 117.51 | 21.48 | 21.65 | 612 | |
| LiteRTHardware=CPU, GS=322026.05 | 94.03 | 99.19 | 19.93 | 20.74 | 667 | |
| llama.cppHardware=GPU, GS=322026.05 | 36 | 38.7 | 15.7 | 18.3 | 728 |