mlboydaisuke commited on
Commit
6a48dff
·
verified ·
1 Parent(s): 3183689

Add measured Raspberry Pi 5 (CPU) section

Browse files

Numbers from the 2026-08-31 Raspberry Pi 5 sweep: LiteRT benchmark_model
(litert-cli-nightly 0.2.0.dev20260805), CPU/XNNPACK, 4 threads, 10 warm-up +
50 timed runs x 3 invocations per file, cooldown between invocations, thermal
throttle state verified 0x0 on every run. Median + spread, never single-shot.
Measurements are joined to this repo by exact .tflite file name.

Files changed (1) hide show
  1. README.md +11 -0
README.md CHANGED
@@ -153,3 +153,14 @@ Measured on a **Samsung Galaxy S26** (Snapdragon 8 Elite Gen 5 / SM8850, Hexagon
153
  **The NPU rows ran the published file unchanged.** LiteRT compiled it for the Hexagon on the device at first load. Those first compiles took 7.3 s to 13.7 min here. The `load` column above is the cached load every later run pays. Recipe and the runtime libraries it needs: [NPU guide](https://github.com/john-rocky/hf-to-litertlm/blob/main/docs/android-npu.md).
154
 
155
  GPU wiring: [GPU guide](https://github.com/john-rocky/hf-to-litertlm/blob/main/docs/android-gpu.md).
 
 
 
 
 
 
 
 
 
 
 
 
153
  **The NPU rows ran the published file unchanged.** LiteRT compiled it for the Hexagon on the device at first load. Those first compiles took 7.3 s to 13.7 min here. The `load` column above is the cached load every later run pays. Recipe and the runtime libraries it needs: [NPU guide](https://github.com/john-rocky/hf-to-litertlm/blob/main/docs/android-npu.md).
154
 
155
  GPU wiring: [GPU guide](https://github.com/john-rocky/hf-to-litertlm/blob/main/docs/android-gpu.md).
156
+
157
+ ## Raspberry Pi 5 (CPU)
158
+
159
+ Measured on a Raspberry Pi 5 Model B Rev 1.1 (8 GB, Raspberry Pi OS 64-bit) with the LiteRT [`benchmark_model`](https://ai.google.dev/edge/litert/models/measurement) tool from litert-cli-nightly 0.2.0.dev20260805: CPU inference (XNNPACK, 4 threads), 3 invocations per file of 10 warm-up plus 50 timed runs (the tool caps a phase at 150 s, so very slow graphs run fewer — the Runs column is the actual timed total). The latency is the median across invocations; the spread is the min–max over all timed runs. No thermal throttling occurred during these runs (`vcgencmd get_throttled` stayed `0x0`).
160
+
161
+ | File | Inference (median) | Spread (min–max) | Runs | Peak memory |
162
+ |---|---:|---:|---:|---:|
163
+ | `ram_reweight_fp16.tflite` | 499.3 ms | 494.5–502.4 ms | 150 | 1185 MB |
164
+ | `ram_stage3_tail_fp16.tflite` | 235.1 ms | 224.4–246.8 ms | 150 | 391 MB |
165
+ | `ram_swin_s012_fp16.tflite` | 3,231.9 ms | 3,191.7–3,350.1 ms | 141 | 1099 MB |
166
+ | `ram_taghead_fp16.tflite` | 1,935.1 ms | 1,721.5–2,204.8 ms | 150 | 358 MB |