Home / GPU guide / RTX 5090 32GB
| Size (16:9) | 5s | 8s | 10s | 15s |
|---|---|---|---|---|
| 416 × 256 0.1MP · flicker-prone | 46 s VRAM 23G | 51 s VRAM 24G | 55 s VRAM 24G | 1 min 5 s VRAM 24G |
| 864 × 480 0.4MP | 1 min 15 s VRAM 25G | 1 min 45 s VRAM 26G | 2 min 11 s VRAM 27G | 3 min 36 s VRAM 29G |
| 960 × 544 0.5MP | 1 min 27 s VRAM 25G | 2 min 12 s VRAM 27G | 2 min 51 s VRAM 28G | 5 min 2 s VRAM 30G |
| 1152 × 640 0.7MP | 1 min 57 s VRAM 26G | 3 min 21 s VRAM 29G | 4 min 36 s VRAM 30G | 9 min 1 s VRAM 31G |
| 1376 × 768 1.0MP · above native | 2 min 56 s VRAM 28G | 5 min 43 s VRAM 31G | 8 min 21 s VRAM 31G | 18 min 11 s VRAM 31G |
| 1920 × 1088 2.0MP · above native | 8 min 21 s VRAM 31G | 20 min 35 s VRAM 31G | 33 min 17 s VRAM 31G | 1.4 h VRAM 31G |
Times are per whole clip (model load and text encoding included); the first run is slower due to kernel compilation. ⚠ = weight offloading / noticeably slower; ✕ = not recommended.
--fast-disk1344x768 full length 5-15s; 4-8 steps; ~80s end-to-end for a 3s clip (measured on our 5090)
Yes. Estimated with 32GB RAM + fast-disk: 1920×1088·15s at ~1.4 h per clip. Configs that exceed VRAM fall back to weight offloading (slower, but they still render).
About 1.4 h per 1920×1088·15s clip (including model load and encoding). Longer or denser clips scale quadratically in the attention term.
At least 64GB of system RAM recommended (minimum 48GB plus a 16GB pagefile). With 32GB, always enable --fast-disk or weights spill into the pagefile.
Stack step distillation (e.g. LightX2V Turbo 8-step v1.0), enable SageAttention / Kitchen attention; for long clips try SLA / VSA sparse attention. All of it is switchable in step 2 of the calculator.
→ Run the numbers for your RAM and budget in the calculator (RTX 5090 32GB pre-filled)
Same tier: RTX 5090 D 32GB · RTX 5090 D V2 24GB
Related: Hardware guide · Acceleration compared · 8GB VRAM setups