Standardize measured performance comparison

README.md

@@ -52,13 +52,22 @@ download only the full input-major checkpoint.
 On a 32 GiB M1 Pro, using the same prompt, seed, 512x512 canvas, eight passes,
 50 blocks, and output settings:
 
-| path | transformer | total generation | gain vs regular |
+| comparison | regular / baseline | optimized path | resulting gain |
 |---|---:|---:|---:|
-| regular output-major | 253.9 s | 278.7 s | baseline |
-| input-major | 231.4 s | 256.5 s | **8.9% transformer; 8.0% total (1.09x)** |
+| transformer layout, matched eight-pass run | 253.9 s output-major | 231.4 s input-major | **8.9% (1.10x)** |
+| complete generation, matched layout A/B | 278.7 s | 256.5 s | **8.0% (1.09x)** |
+| regular H3 INT8 vs 8-step Turbo INT8, published RTX 4070 mean | 272.97 s / 20 passes | 130.66 s / 8 passes | **52.1% (2.09x)** |
+| regular H3 INT8 vs Turbo INT8 denoising work | 20 passes / 1,000 block evaluations | 8 passes / 400 block evaluations | **60% less work; 2.5x fewer passes** |
 
 The generated images were identical. Actual gains vary with canvas size,
-memory pressure, and Mac.
+memory pressure, and Mac. The external wall-time row comes from a reproducible
+[three-pattern RTX 4070 benchmark](https://github.com/sepiablue-ai/minimax-h3-turbo-lora-benchmark/tree/main/202608_larry_vs_lightx2v)
+that used the regular `minimax_h3_fl2va_pruned_int8_convrot.safetensors` as its
+272.97-second baseline. Its two published eight-step Turbo means were 130.30
+and 131.02 seconds; their 130.66-second mean is shown above. That is external
+NVIDIA/ComfyUI evidence, not a projected M1 time. The work-count row is exact,
+and the Turbo checkpoint is intended to reach roughly the regular model's
+20-pass fidelity in 6–8 passes.
 
 ## Why this merge was made and republished