On this page, we will display the required memory and execution time under different combinations of memory optimization parameters. We would also like to thank the community for their support in helping us enrich the table data.
Please note that due to variations in the running environment, the content in the table may differ from actual conditions and is for reference only.
Results from bmgjet
- Environments
| CPU | Ram | Disk | GPU |
|---|---|---|---|
| 7950X | 64GB 6200mhz, 30,32,32 | NVME 990 Pro | 3090 (Water Cooled 500W PL, +1000mhz Vram) Sits about 2.1ghz on core |
Memory usage reported from GPUz logs.
Torch was not compiled with flash attention.
Windows 11 Using Python Environment from Forge (webui_forge_cu124_torch24.7z) Feb 5 2024.
- Resolution of 512
| Num frames | Parameters | GPU memory | Time |
|---|---|---|---|
| 24 frames | normal 0 | 18117mb | 1:32s - 3.80s/it |
| 48 frames | normal 0 | 19995mb | 4:06s - 9.98s/it |
| 72 frames | normal 0 | 22133mb | 9:57s - 23.88s/it |
| 96 frames | normal 10 | 22200mb | 12:51s - 30.84s/it |
| 120 frames | normal 7 | 23195mb | 15:57s - 38.26s/it |
- Resolution of 768
| Num frames | Parameters | GPU memory | Time |
|---|---|---|---|
| 24 frames | normal 0 | 22105mb | 4:07s - 9.94s/it |
| 48 frames | normal 7 | 22195mb | 14:34s - 34.98s/it |
| 72 frames | lowmem 0 | 17998mb | 38:29s - 91.99s/it |
| 96 frames | lowmem 0 | 23000mb | 1:04:16s - 153.74s/it |
| 120 frames | lowmem 0 | 28195mb | 5:29:54s - 797.28s/it (froze system during vram overflows for a few sec each pass) |
The GPU memory reported below is the
max_memory_allocated()value. The values read from nvidia-smi may be higher than the reported values because CUDA occupies some GPU memory (usually between 500 - 800 MiB), and PyTorch's caching mechanism also requests additional GPU memory.
- Resolution of 384
| Num frames |
normal_mode + 0 steps | normal_mode + 10 steps | normal_mode + 7 steps | normal_mode + 5 steps | normal_mode + 1 steps |
|---|---|---|---|---|---|
| 24 frames |
15387MiB 00:42s |
15052MiB 00:50s |
14941MiB 00:50s |
14867MiB 00:57s |
14718MiB 00:58s |
| 48 frames |
16686MiB 01:15s |
16018MiB 01:33s |
15795MiB 01:35s |
15647MiB 01:38s |
15350MiB 01:40s |
| 72 frames |
17996MiB 01:51s |
16993MiB 02:12s |
16659MiB 02:21s |
16436MiB 02:22s |
15990MiB 02:31s |
| 96 frames |
19297MiB 02:31s |
17959MiB 02:58s |
17514MiB 03:09s |
17216MiB 03:13s |
16621MiB 03:27s |
| 120 frames |
20599MiB 03:17s |
18928MiB 03:50s |
18371MiB 04:09s |
17999MiB 04:14s |
17256MiB 04:29s |
| Num frames |
low_gpu_mode + 0 steps | low_gpu_mode + 10 steps | low_gpu_mode + 7 steps | low_gpu_mode + 5 steps | low_gpu_mode + 1 steps |
|---|---|---|---|---|---|
| 24 frames |
4096MiB 02:51s |
4096MiB 03:25s |
4096MiB 03:35s |
4096MiB 02:45s |
4095MiB 03:13s |
| 48 frames |
4148MiB 03:37s |
4148MiB 04:01s |
4148MiB 04:19s |
4148MiB 03:43s |
4147MiB 04:52s |
| 72 frames |
4200MiB 05:01s |
4200MiB 05:40s |
4200MiB 07:51s |
4200MiB 05:01s |
4200MiB 06:18s |
| 96 frames |
5248MiB 05:59s |
4249MiB 06:49s |
4249MiB 08:43s |
4249MiB 06:03s |
4249MiB 07:09s |
| 120 frames |
6549MiB 07:24s |
4876MiB 08:12s |
4321MiB 10:45s |
4299MiB 08:09s |
4299MiB 08:32s |
- Resolution of 512
| Num frames |
normal_mode + 0 steps | normal_mode + 10 steps | normal_mode + 7 steps | normal_mode + 5 steps | normal_mode + 1 steps |
|---|---|---|---|---|---|
| 24 frames |
16366MiB 01:18s |
15805MiB 01:26s |
15607MiB 01:37s |
15475MiB 01:36s |
15211MiB 01:39s |
| 48 frames |
18720MiB 02:21s |
17532MiB 02:49s |
17136MiB 02:55s |
16872MiB 02:58s |
16344MiB 03:01s |
| 72 frames |
21036MiB 03:41s |
19254MiB 04:25s |
18660MiB 04:34s |
18264MiB 04:36s |
17472MiB 04:51s |
| 96 frames |
-----MiB --:--s |
20972MiB 06:18s |
20180MiB 06:24s |
19652MiB 06:36s |
18596MiB 06:56s |
| 120 frames |
-----MiB --:--s |
-----MiB --:--s |
21704MiB 08:50s |
21044MiB 08:53s |
19724MiB 09:08s |
| Num frames |
low_gpu_mode + 0 steps | low_gpu_mode + 10 steps | low_gpu_mode + 7 steps | low_gpu_mode + 5 steps | low_gpu_mode + 1 steps |
|---|---|---|---|---|---|
| 24 frames |
4212MiB 04:31s |
4212MiB 04:44s |
4211MiB 05:12s |
4211MiB 04:51s |
4212MiB 04:14s |
| 48 frames |
4666MiB 05:17s |
4401MiB 06:11s |
4400MiB 06:33s |
4399MiB 08:29s |
4401MiB 06:18s |
| 72 frames |
6981MiB 07:27s |
5199MiB 08:00s |
4605MiB 08:57s |
4598MiB 11:02s |
4598MiB 08:46s |
| 96 frames |
9298MiB 09:33s |
6922MiB 10:13s |
6130MiB 11:31s |
5602MiB 13:46s |
4794MiB 11:29s |
| 120 frames |
11613MiB 11:33s |
8643MiB 13:11s |
7653MiB 14:18s |
6993MiB 14:23s |
5673MiB 14:16s |
- Resolution of 640
| Num frames |
normal_mode + 0 steps | normal_mode + 10 steps | normal_mode + 7 steps | normal_mode + 5 steps | normal_mode + 1 steps |
|---|---|---|---|---|---|
| 24 frames |
17671MiB 01:47s |
16739MiB 02:09s |
16429MiB 02:17s |
16234MiB 02:23s |
15850MiB 02:28s |
| 48 frames |
21324MiB 03:49s |
19468MiB 04:23s |
18849MiB 04:55s |
18437MiB 04:59s |
17612MiB 05:14s |
| 72 frames |
-----MiB --:--s |
-----MiB --:--s |
21230MiB 07:57s |
20611MiB 08:13s |
19373MiB 08:25s |
| 96 frames |
-----MiB --:--s |
-----MiB --:--s |
-----MiB --:--s |
-----MiB --:--s |
21134MiB 11:57s |
| 120 frames |
-----MiB --:--s |
-----MiB --:--s |
-----MiB --:--s |
-----MiB --:--s |
-----MiB --:--s |
| Num frames |
low_gpu_mode + 0 steps | low_gpu_mode + 10 steps | low_gpu_mode + 7 steps | low_gpu_mode + 5 steps | low_gpu_mode + 1 steps |
|---|---|---|---|---|---|
| 24 frames |
4301MiB 03:49s |
4301MiB 04:31s |
4301MiB 04:55s |
4301MiB 06:33s |
4301MiB 04:52s |
| 48 frames |
7271MiB 07:05s |
5416MiB 08:04s |
4798MiB 08:07s |
4589MiB 09:42s |
4589MiB 09:18s |
| 72 frames |
10889MiB 10:38s |
8106MiB 11:54s |
7179MiB 11:58s |
6560MiB 12:52s |
5322MiB 13:13s |
| 96 frames |
14509MiB 14:13s |
10795MiB 15:50s |
9557MiB 16:10s |
8732MiB 17:08s |
7082MiB 17:39s |
| 120 frames |
18128MiB 18:26s |
13487MiB 19:57s |
11942MiB 21:12s |
10910MiB 21:55s |
8847MiB 26:20s |
- Resolution of 768
| Num frames |
normal_mode + 0 steps | normal_mode + 10 steps | normal_mode + 7 steps | normal_mode + 5 steps | normal_mode + 1 steps |
|---|---|---|---|---|---|
| 24 frames |
19223MiB 02:38s |
17900MiB 03:06s |
17448MiB 03:18s |
17153MiB 03:23s |
16624MiB 03:34s |
| 48 frames |
-----MiB --:--s |
-----MiB --:--s |
20946MiB 07:28s |
20352MiB 07:35s |
19164MiB 08:04s |
| 72 frames |
-----MiB --:--s |
-----MiB --:--s |
-----MiB --:--s |
-----MiB --:--s |
-----MiB --:--s |
| 96 frames |
-----MiB --:--s |
-----MiB --:--s |
-----MiB --:--s |
-----MiB --:--s |
-----MiB --:--s |
| 120 frames |
-----MiB --:--s |
-----MiB --:--s |
-----MiB --:--s |
-----MiB --:--s |
-----MiB --:--s |
| Num frames |
low_gpu_mode + 0 steps | low_gpu_mode + 10 steps | low_gpu_mode + 7 steps | low_gpu_mode + 5 steps | low_gpu_mode + 1 steps |
|---|---|---|---|---|---|
| 24 frames |
5251MiB 05:19s |
4398MiB 05:55s |
4398MiB 06:52s |
4398MiB 07:24s |
4399MiB 07:35s |
| 48 frames |
10457MiB 09:38s |
7786MiB 11:09s |
6896MiB 12:43s |
6304MiB 12:52s |
5114MiB 13:20s |
| 72 frames |
15671MiB 15:14s |
11661MiB 17:50s |
10325MiB 18:22s |
9433MiB 19:32s |
7652MiB 19:19s |
| 96 frames |
-----MiB --:--s |
15534MiB 24:30s |
13752MiB 24:28s |
12564MiB 25:54s |
10188MiB 29:19s |
| 120 frames |
-----MiB --:--s |
19406MiB 33:04s |
17179MiB 33:14s |
15694MiB 35:01s |
12724MiB 39:11s |
- Resolution of 896
| Num frames |
normal_mode + 0 steps | normal_mode + 10 steps | normal_mode + 7 steps | normal_mode + 5 steps | normal_mode + 1 steps |
|---|---|---|---|---|---|
| 24 frames |
21085MiB 03:46s |
19265MiB 04:25s |
18659MiB 04:39s |
18248MiB 04:43s |
17543MiB 04:59s |
| 48 frames |
-----MiB --:--s |
-----MiB --:--s |
-----MiB --:--s |
-----MiB --:--s |
20994MiB 11:34s |
| 72 frames |
-----MiB --:--s |
-----MiB --:--s |
-----MiB --:--s |
-----MiB --:--s |
-----MiB --:--s |
| 96 frames |
-----MiB --:--s |
-----MiB --:--s |
-----MiB --:--s |
-----MiB --:--s |
-----MiB --:--s |
| 120 frames |
-----MiB --:--s |
-----MiB --:--s |
-----MiB --:--s |
-----MiB --:--s |
-----MiB --:--s |
| Num frames |
low_gpu_mode + 0 steps | low_gpu_mode + 10 steps | low_gpu_mode + 7 steps | low_gpu_mode + 5 steps | low_gpu_mode + 1 steps |
|---|---|---|---|---|---|
| 24 frames |
7128MiB 07:26s |
5309MiB 07:45s |
4703MiB 07:50s |
4517MiB 08:29s |
4517MiB 09:31s |
| 48 frames |
14220MiB 14:05s |
10582MiB 15:01s |
9370MiB 15:43s |
8562MiB 16:34s |
6946MiB 17:36s |
| 72 frames |
-----MiB --:--s |
15857MiB 24:53s |
14038MiB 25:06s |
12826MiB 26:49s |
10401MiB 27:30s |
| 96 frames |
-----MiB --:--s |
-----MiB --:--s |
-----MiB --:--s |
17087MiB 38:52s |
13853MiB 45:19s |
| 120 frames |
-----MiB --:--s |
-----MiB --:--s |
-----MiB --:--s |
-----MiB --:--s |
17310MiB 55:50s |
The GPU memory reported below is the
max_memory_allocated()value. The values read from nvidia-smi may be higher than the reported values because CUDA occupies some GPU memory (usually between 500 - 800 MiB), and PyTorch's caching mechanism also requests additional GPU memory.
- Resolution of 384
| Num frames |
normal_mode + 0 steps | normal_mode + 10 steps | normal_mode + 7 steps | normal_mode + 5 steps | normal_mode + 1 steps |
|---|---|---|---|---|---|
| 24 frames |
15387MiB 00:36s |
15052MiB 00:43s |
14941MiB 00:50s |
14867MiB 00:50s |
14717MiB 00:56s |
| 48 frames |
16686MiB 00:57s |
16018MiB 01:10s |
15795MiB 01:28s |
15647MiB 01:16s |
15350MiB 01:21s |
| 72 frames |
17996MiB 01:29s |
16993MiB 01:43s |
16659MiB 01:53s |
16436MiB 01:53s |
15990MiB 01:57s |
| 96 frames |
19297MiB 02:01s |
17959MiB 02:21s |
17514MiB 02:29s |
17216MiB 02:33s |
16621MiB 02:54s |
| 120 frames |
20599MiB 02:33s |
18928MiB 03:04s |
18371MiB 03:12s |
17999MiB 03:16s |
17256MiB 03:25s |
| Num frames |
low_gpu_mode + 0 steps | low_gpu_mode + 10 steps | low_gpu_mode + 7 steps | low_gpu_mode + 5 steps | low_gpu_mode + 1 steps |
|---|---|---|---|---|---|
| 24 frames |
4096MiB 03:56s |
4096MiB 04:24s |
4096MiB 04:16s |
4096MiB 04:14s |
4095MiB 04:41s |
| 48 frames |
4148MiB 04:45s |
4148MiB 05:27s |
4148MiB 05:29s |
4148MiB 05:32s |
4147MiB 06:06s |
| 72 frames |
4200MiB 06:10s |
4200MiB 07:10s |
4200MiB 07:13s |
4200MiB 07:19s |
4200MiB 07:24s |
| 96 frames |
5248MiB 07:53s |
4249MiB 08:16s |
4249MiB 08:30s |
4249MiB 08:28s |
4249MiB 09:05s |
| 120 frames |
6547MiB 08:43s |
4876MiB 09:54s |
4321MiB 10:16s |
4299MiB 10:20s |
4299MiB 11:26s |
- Resolution of 512
| Num frames |
normal_mode + 0 steps | normal_mode + 10 steps | normal_mode + 7 steps | normal_mode + 5 steps | normal_mode + 1 steps |
|---|---|---|---|---|---|
| 24 frames |
16399MiB 00:59s |
15805MiB 01:07s |
15607MiB 01:13s |
15475MiB 01:18s |
15211MiB 01:16s |
| 48 frames |
18720MiB 01:47s |
17532MiB 02:10s |
17136MiB 02:25s |
16872MiB 02:22s |
16344MiB 02:24s |
| 72 frames |
21036MiB 02:55s |
19254MiB 03:27s |
18660MiB 03:36s |
18264MiB 03:41s |
17472MiB 03:51s |
| 96 frames |
23348MiB 04:08s |
20972MiB 04:46s |
20180MiB 05:05s |
19652MiB 05:12s |
18596MiB 05:38s |
| 120 frames |
25664MiB 05:42s |
22694MiB 06:28s |
21704MiB 06:47s |
21044MiB 06:50s |
19724MiB 07:06s |
| Num frames |
low_gpu_mode + 0 steps | low_gpu_mode + 10 steps | low_gpu_mode + 7 steps | low_gpu_mode + 5 steps | low_gpu_mode + 1 steps |
|---|---|---|---|---|---|
| 24 frames |
4212MiB 05:03s |
4213MiB 05:31s |
4212MiB 05:23s |
4212MiB 05:50s |
4212MiB 05:51s |
| 48 frames |
4666MiB 06:47s |
4401MiB 07:13s |
4401MiB 07:34s |
4400MiB 08:21s |
4400MiB 08:19s |
| 72 frames |
6981MiB 08:41s |
5199MiB 09:44s |
4605MiB 10:10s |
4598MiB 11:13s |
4598MiB 12:00s |
| 96 frames |
9298MiB 10:59s |
6922MiB 12:10s |
6130MiB 12:16s |
5602MiB 13:27s |
4794MiB 15:17s |
| 120 frames |
11613MiB 13:27s |
8643MiB 14:55s |
7653MiB 15:22s |
6993MiB 16:04s |
5673MiB 18:04s |
- Resolution of 640
| Num frames |
normal_mode + 0 steps | normal_mode + 10 steps | normal_mode + 7 steps | normal_mode + 5 steps | normal_mode + 1 steps |
|---|---|---|---|---|---|
| 24 frames |
17707MiB 01:26s |
16778MiB 01:45s |
16470MiB 01:48s |
16263MiB 01:51s |
15850MiB 01:53s |
| 48 frames |
21324MiB 03:09s |
19468MiB 03:36s |
18849MiB 03:53s |
18437MiB 03:53s |
17612MiB 04:03s |
| 72 frames |
24943MiB 05:06s |
22158MiB 05:54s |
21230MiB 06:12s |
20612MiB 06:18s |
19374MiB 06:35s |
| 96 frames |
28562MiB 07:35s |
24848MiB 08:36s |
23611MiB 09:04s |
22785MiB 09:07s |
21134MiB 09:35s |
| 120 frames |
32182MiB 10:32s |
27541MiB 11:54s |
25994MiB 12:20s |
24962MiB 12:37s |
22899MiB 13:05s |
| Num frames |
low_gpu_mode + 0 steps | low_gpu_mode + 10 steps | low_gpu_mode + 7 steps | low_gpu_mode + 5 steps | low_gpu_mode + 1 steps |
|---|---|---|---|---|---|
| 24 frames |
4301MiB 05:40s |
4301MiB 06:07s |
4301MiB 06:27s |
4301MiB 06:33s |
4301MiB 06:57s |
| 48 frames |
7271MiB 09:18s |
5416MiB 09:51s |
4798MiB 10:26s |
4589MiB 10:36s |
4589MiB 11:52s |
| 72 frames |
10889MiB 13:03s |
8106MiB 14:07s |
7179MiB 14:35s |
6561MiB 17:30s |
5322MiB 16:33s |
| 96 frames |
14509MiB 16:57s |
10795MiB 18:12s |
9557MiB 18:29s |
8732MiB 18:55s |
7082MiB 21:10s |
| 120 frames |
18128MiB 22:05s |
13487MiB 22:49s |
11942MiB 23:47s |
10910MiB 23:50s |
8847MiB 32:46s |
- Resolution of 768
| Num frames |
normal_mode + 0 steps | normal_mode + 10 steps | normal_mode + 7 steps | normal_mode + 5 steps | normal_mode + 1 steps |
|---|---|---|---|---|---|
| 24 frames |
19300MiB 02:09s |
17962MiB 02:32s |
17517MiB 02:37s |
17219MiB 02:39s |
16624MiB 02:46s |
| 48 frames |
24510MiB 04:46s |
21837MiB 05:32s |
20946MiB 05:51s |
20352MiB 05:59s |
19164MiB 06:10s |
| 72 frames |
29725MiB 08:23s |
25715MiB 09:35s |
24378MiB 09:56s |
23486MiB 10:05s |
21703MiB 10:40s |
| 96 frames |
34925MiB 12:52s |
29579MiB 14:33s |
27797MiB 15:01s |
26609MiB 15:13s |
24233MiB 15:53s |
| 120 frames |
40134MiB 18:29s |
33451MiB 20:39s |
31223MiB 21:31s |
29737MiB 21:42s |
26766MiB 22:51s |
| Num frames |
low_gpu_mode + 0 steps | low_gpu_mode + 10 steps | low_gpu_mode + 7 steps | low_gpu_mode + 5 steps | low_gpu_mode + 1 steps |
|---|---|---|---|---|---|
| 24 frames |
5251MiB 08:04s |
4398MiB 08:15s |
4398MiB 08:21s |
4398MiB 08:48s |
4399MiB 09:14s |
| 48 frames |
10457MiB 12:42s |
7786MiB 13:54s |
6896MiB 14:17s |
6304MiB 14:37s |
5114MiB 16:21s |
| 72 frames |
15671MiB 18:18s |
11661MiB 19:44s |
10325MiB 20:53s |
9435MiB 20:47s |
7652MiB 23:42s |
| 96 frames |
20880MiB 24:55s |
15534MiB 27:07s |
13752MiB 28:13s |
12564MiB 28:44s |
10188MiB 30:30s |
| 120 frames |
26092MiB 31:21s |
19406MiB 34:09s |
17179MiB 35:13s |
15694MiB 35:15s |
12724MiB 32:55s |
- Resolution of 896
| Num frames |
normal_mode + 0 steps | normal_mode + 10 steps | normal_mode + 7 steps | normal_mode + 5 steps | normal_mode + 1 steps |
|---|---|---|---|---|---|
| 24 frames |
21181MiB 02:58s |
19362MiB 03:32s |
18755MiB 03:48s |
18351MiB 03:47s |
17543MiB 03:54s |
| 48 frames |
28271MiB 07:18s |
24633MiB 08:24s |
23420MiB 08:49s |
22611MiB 08:53s |
20994MiB 09:10s |
| 72 frames |
35360MiB 13:22s |
29902MiB 14:57s |
28083MiB 15:33s |
26870MiB 15:46s |
24445MiB 16:13s |
| 96 frames |
42447MiB 21:08s |
35169MiB 23:48s |
32744MiB 24:41s |
31126MiB 25:08s |
27891MiB 26:33s |
| 120 frames |
49538MiB 30:55s |
40442MiB 34:03s |
37410MiB 35:11s |
35388MiB 35:31s |
31345MiB 36:40s |
| Num frames |
low_gpu_mode + 0 steps | low_gpu_mode + 10 steps | low_gpu_mode + 7 steps | low_gpu_mode + 5 steps | low_gpu_mode + 1 steps |
|---|---|---|---|---|---|
| 24 frames |
7128MiB 08:52s |
5309MiB 09:30s |
4702MiB 10:37s |
4517MiB 10:55s |
4517MiB 12:09s |
| 48 frames |
14220MiB 15:56s |
10582MiB 17:04s |
9370MiB 18:49s |
8562MiB 19:08s |
6946MiB 21:50s |
| 72 frames |
21315MiB 24:58s |
15857MiB 26:54s |
14038MiB 29:10s |
12826MiB 29:42s |
10401MiB 30:58s |
| 96 frames |
28407MiB 31:16s |
21130MiB 34:22s |
18704MiB 36:45s |
17087MiB 36:34s |
13853MiB 43:48s |
| 120 frames |
35504MiB 45:00s |
26405MiB 48:05s |
23373MiB 49:27s |
21352MiB 50:34s |
17310MiB 59:38s |