-
Notifications
You must be signed in to change notification settings - Fork 0
Performance
Headline numbers for v0.11.0. Timed SKUs: Strix Halo (gfx1151), MI210 (gfx90a), MI300X (gfx942). Build with an explicit HIP_ARCH (Getting started). Wall time moves with page-cache state; treat E2E as a band. Denoise GPU is the stable GPU figure.
In-tree copy: docs/PERFORMANCE.md.
| Preset | Strix Halo (gfx1151) | MI210 (gfx90a) | MI300X (gfx942) |
|---|---|---|---|
| fox-s2 E2E | ~85–90 s (I/O) | ~10.8 s | ~16 s |
| fox-fast E2E | ~2 min (I/O) | ~18 s | ~12 s |
| 15 s cinematic E2E | 40 min 46 s | 12 min 11 s | 3 min 46 s |
Strix Halo v0.9.0 fox-s2 md5 1731f95c4aa582597cf83d57f46b8f9e. On v0.11.0 default 512 px VAE tiles, fox-s2 is 34507f072c5cabbde6592b3f70b8fa35.
Halo 15 s quality path: denoise 36 min 38 s, VAE 174 s (2×1 @ 480 px), DiT peak 27.9 GiB. Opt-in all-opts 15 s is 27 min 3 s (TR + INT8 VAE; GPU sampler does not move Halo 15 s). Details: Long video.
# fox-s2 — HIP A/B smoke (gfx1151 md5 gate)
./h3 --profile -d /path/to/MiniMax-H3 \
-p "A red fox walks through fresh snow." \
--width 512 --height 512 --frames 22 --steps 2 --layers 35 --reuse 1 \
-o outputs/fox-s2.mp4
# fox-fast — upstream tutorial “first fast video”, without --show
./h3 --profile -d /path/to/MiniMax-H3 \
-p "A red fox walks through fresh snow in a pine forest. Medium tracking shot, natural winter light, realistic fur, soft footsteps and wind." \
--width 512 --height 512 --frames 22 --steps 20 --layers 45 --reuse 2 \
-o outputs/fox-fast.mp4--reuse 2 on 20 steps yields 11 DiT evaluations. fox-s2 / fox-fast are complete ~0.9 s MP4s, not truncated previews. Showcase fox uses --layers 50 --reuse 1 and is not the fox-fast scoreboard.
Long T2VA: Long video. MI210 (gfx90a) DiT is BF16 by default (H3_INT8_MLP=1 restores INT8).
Optional --token-reduction (off by default; same CLI as h3-spark.c) pairs middle-block video tokens. Do not replace the tagged quality-path row (40 min 46 s / 12 min 11 s) with these numbers.
| quality path | speed path | |
|---|---|---|
| Strix Halo (gfx1151) 15 s E2E | 40 min 46 s (v0.11.0) | 27 min 3 s (TR + INT8 VAE); GPU sampler does not move Halo 15 s |
| MI210 (gfx90a) 15 s E2E | 12 min 11 s |
8 min 21 s (--token-reduction) |
| MI300X (gfx942) 15 s E2E | 3 min 46 s | ~2.4 min (all-opts) |
| When | fox-s2 E2E | fox-fast E2E | What landed |
|---|---|---|---|
| 2026-08-18 | ~257 s | — | First HIP profile |
| 2026-08-19 | ~117 s | — | INT8 DiT, faster load I/O |
| 2026-08-22 | — | ~213 s | First fox-fast measure |
| v0.9.0 | 83–87 s | 95 s | WMMA kernels; weights in VRAM carveout |
| v0.10.0 | same on gfx1151 | ~18 s on gfx90a | Dual ISA; MI210 MFMA flash |
| v0.11.0 | ~85–90 s I/O | ~2 min I/O | gfx942; 512 px VAE tiles; 15 s 40 min 46 s |