Support multi GPU

  • A single encoding process can only be done by one GPU. What would be possible: If you have multiple encodings like one in Premiere and one in Media Encoder, that you assign each encoding one specific GPU.

    Maybe it could be possible that Voukoder supports parallel encoding for Media Encoder in the future, but at this point Vouk has a lot more knowledge.

  • Spoiler anzeigen

    [15:30:31] ---------------------------------------------

    [15:30:31] Export started

    [15:30:31] ---------------------------------------------

    [15:30:31] Frame size: 1920x1080

    [15:30:31] Pixel aspect: 1:1

    [15:30:31] Frame rate: 29.97

    [15:30:31] Interlaced: No

    [15:30:31] Color range: tv

    [15:30:31] Color space: bt709

    [15:30:31] Color primaries: bt709

    [15:30:31] Color TRC: bt709

    [15:30:31] Sample rate: 48000

    [15:30:31] Audio channels: 2

    [15:30:31] ---------------------------------------------

    [15:30:31] Opening codec: h264_nvenc with options: gpu=0

    [15:30:31] Opening codec: aac with options: b=256000|profile=aac_low

    [15:30:31] Requesting pixel format: yuv420p

    [15:30:32] Frame #0: vRender: 1158382 us, vProcess: 9 us, vEncoding: 277623 us, Latency: 1436075 us

    [15:30:33] Frame #1: vRender: 311 us, vProcess: 7 us, vEncoding: 1472 us, aRender: 45 us, aEncoding: 318 us, Latency: 1175782 us

    [15:30:33] Frame #2: vRender: 639 us, vProcess: 5 us, vEncoding: 1870 us, aRender: 33 us, aEncoding: 494 us, Latency: 3629 us

    [15:30:33] Frame #3: vRender: 20 us, vProcess: 4 us, vEncoding: 1735 us, aRender: 85 us, aEncoding: 2657 us, Latency: 4545 us

    [15:30:34] Frame #4: vRender: 15 us, vProcess: 4 us, vEncoding: 99658 us, aRender: 69 us, aEncoding: 6630 us, Latency: 108024 us

    [15:30:34] Frame #5: vRender: 20 us, vProcess: 5 us, vEncoding: 22509 us, aRender: 114 us, aEncoding: 5738 us, Latency: 28460 us

    [15:30:34] Frame #6: vRender: 30 us, vProcess: 7 us, vEncoding: 231282 us, aRender: 158 us, aEncoding: 6922 us, Latency: 246833 us

    [15:30:34] Frame #7: vRender: 481 us, vProcess: 10 us, vEncoding: 3144 us, aRender: 157 us, aEncoding: 6583 us, Latency: 10442 us

    [15:30:34] Frame #8: vRender: 10987 us, vProcess: 9 us, vEncoding: 242662 us, aRender: 127 us, aEncoding: 2111 us, Latency: 258352 us

    [15:30:34] Frame #9: vRender: 21042 us, vProcess: 9 us, vEncoding: 1628 us, aRender: 53 us, aEncoding: 1321 us, Latency: 25496 us

    [15:30:34] Frame #10: vRender: 267766 us, vProcess: 8 us, vEncoding: 2212 us, aRender: 121 us, aEncoding: 1191 us, Latency: 271358 us

    [15:30:35] Frame #11: vRender: 273780 us, vProcess: 9 us, vEncoding: 1750 us, aRender: 40 us, aEncoding: 804 us, Latency: 277450 us

    [15:30:35] Frame #12: vRender: 268446 us, vProcess: 10 us, vEncoding: 2113 us, aRender: 122 us, aEncoding: 1118 us, Latency: 271869 us

    [15:30:35] Frame #13: vRender: 483630 us, vProcess: 8 us, vEncoding: 1748 us, aRender: 35 us, aEncoding: 2182 us, Latency: 488700 us

    [15:30:36] Frame #14: vRender: 247467 us, vProcess: 8 us, vEncoding: 17414 us, aRender: 110 us, aEncoding: 2082 us, Latency: 267139 us

    [15:30:36] Frame #15: vRender: 257441 us, vProcess: 9 us, vEncoding: 50824 us, aRender: 35 us, aEncoding: 990 us, Latency: 310608 us

    [15:30:36] Frame #16: vRender: 210475 us, vProcess: 9 us, vEncoding: 53661 us, aRender: 34 us, aEncoding: 1070 us, Latency: 266592 us

    [15:30:37] Frame #17: vRender: 763536 us, vProcess: 8 us, vEncoding: 10424 us, aRender: 100 us, aEncoding: 2103 us, Latency: 776230 us

    [15:30:37] Frame #18: vRender: 14607 us, vProcess: 9 us, vEncoding: 2323 us, aRender: 37 us, aEncoding: 836 us, Latency: 18955 us

    [15:30:37] Frame #19: vRender: 18779 us, vProcess: 8 us, vEncoding: 254939 us, aRender: 104 us, aEncoding: 857 us, Latency: 274730 us

    [15:30:38] Frame #20: vRender: 224331 us, vProcess: 11 us, vEncoding: 2317 us, aRender: 40 us, aEncoding: 885 us, Latency: 228768 us

    [15:30:38] Frame #21: vRender: 11750 us, vProcess: 6 us, vEncoding: 1575 us, aRender: 90 us, aEncoding: 1293 us, Latency: 14760 us

    [15:30:40] Frame #22: vRender: 2450333 us, vProcess: 7 us, vEncoding: 2206 us, aRender: 89 us, aEncoding: 954 us, Latency: 2454658 us

    [15:30:40] Frame #23: vRender: 305 us, vProcess: 7 us, vEncoding: 2575 us, aRender: 131 us, aEncoding: 1131 us, Latency: 4201 us

    [15:30:40] Frame #24: vRender: 40 us, vProcess: 7 us, vEncoding: 6525 us, aRender: 108 us, aEncoding: 1173 us, Latency: 9270 us

    [15:30:40] Frame #25: vRender: 8838 us, vProcess: 6 us, vEncoding: 3351 us, aRender: 95 us, aEncoding: 1059 us, Latency: 14786 us

    [15:30:40] Frame #26: vRender: 107345 us, vProcess: 12 us, vEncoding: 188086 us, aRender: 108 us, aEncoding: 998 us, Latency: 296594 us

    [15:30:40] Frame #27: vRender: 29 us, vProcess: 7 us, vEncoding: 11323 us, aRender: 36 us, aEncoding: 713 us, Latency: 13084 us

    [15:30:41] Frame #28: vRender: 30 us, vProcess: 7 us, vEncoding: 274486 us, aRender: 97 us, aEncoding: 731 us, Latency: 275391 us

    [15:30:41] Frame #29: vRender: 31 us, vProcess: 6 us, vEncoding: 11417 us, aRender: 34 us, aEncoding: 719 us, Latency: 13141 us

    [15:30:41] Frame #30: vRender: 418 us, vProcess: 6 us, vEncoding: 36321 us, aRender: 90 us, aEncoding: 803 us, Latency: 37680 us

    [15:30:41] Frame #31: vRender: 21 us, vProcess: 6 us, vEncoding: 264145 us, aRender: 34 us, aEncoding: 715 us, Latency: 265835 us

    [15:30:41] Frame #32: vRender: 324 us, vProcess: 6 us, vEncoding: 10341 us, aRender: 40 us, aEncoding: 690 us, Latency: 12386 us

    [15:30:41] Frame #33: vRender: 15389 us, vProcess: 8 us, vEncoding: 70236 us, aRender: 106 us, aEncoding: 826 us, Latency: 86607 us

    [15:30:41] Frame #34: vRender: 217215 us, vProcess: 7 us, vEncoding: 1314 us, aRender: 37 us, aEncoding: 658 us, Latency: 220149 us

    [15:30:42] Frame #35: vRender: 323485 us, vProcess: 9 us, vEncoding: 1551 us, aRender: 103 us, aEncoding: 708 us, Latency: 325914 us

    [15:30:42] Frame #36: vRender: 281654 us, vProcess: 8 us, vEncoding: 1563 us, aRender: 33 us, aEncoding: 643 us, Latency: 284772 us

    [15:30:42] Frame #37: vRender: 311395 us, vProcess: 8 us, vEncoding: 1496 us, aRender: 101 us, aEncoding: 1051 us, Latency: 314117 us

    [15:30:42] Frame #38: vRender: 286730 us, vProcess: 8 us, vEncoding: 1432 us, aRender: 46 us, aEncoding: 742 us, Latency: 289875 us

    [15:30:43] Frame #39: vRender: 320286 us, vProcess: 9 us, vEncoding: 1702 us, aRender: 106 us, aEncoding: 1121 us, Latency: 323290 us

    [15:30:43] Frame #40: vRender: 293885 us, vProcess: 8 us, vEncoding: 1550 us, aRender: 37 us, aEncoding: 703 us, Latency: 297083 us

    [15:30:44] Frame #41: vRender: 267870 us, vProcess: 8 us, vEncoding: 1867 us, aRender: 163178 us, aEncoding: 711 us, Latency: 434715 us

    [15:30:44] Frame #42: vRender: 91490 us, vProcess: 8 us, vEncoding: 1377 us, aRender: 87 us, aEncoding: 697 us, Latency: 93715 us

    [15:30:44] Frame #43: vRender: 307168 us, vProcess: 8 us, vEncoding: 1469 us, aRender: 38 us, aEncoding: 714 us, Latency: 310346 us

    [15:30:44] Frame #44: vRender: 295467 us, vProcess: 7 us, vEncoding: 1323 us, aRender: 80 us, aEncoding: 750 us, Latency: 297684 us

    [15:30:45] Frame #45: vRender: 310153 us, vProcess: 8 us, vEncoding: 1602 us, aRender: 47 us, aEncoding: 780 us, Latency: 313541 us

    [15:30:45] Frame #46: vRender: 268894 us, vProcess: 8 us, vEncoding: 1348 us, aRender: 83 us, aEncoding: 839 us, Latency: 271228 us

    [15:30:45] Frame #47: vRender: 316624 us, vProcess: 8 us, vEncoding: 1594 us, aRender: 38 us, aEncoding: 1025 us, Latency: 320273 us

    [15:30:45] Frame #48: vRender: 292254 us, vProcess: 8 us, vEncoding: 1555 us, aRender: 39 us, aEncoding: 721 us, Latency: 295545 us

    [15:30:46] Frame #49: vRender: 304562 us, vProcess: 8 us, vEncoding: 1506 us, aRender: 95 us, aEncoding: 819 us, Latency: 307050 us

    [15:30:46] Frame #50: vRender: 264212 us, vProcess: 9 us, vEncoding: 1804 us, aRender: 36 us, aEncoding: 679 us, Latency: 267814 us

    [15:30:46] Frame #51: vRender: 282213 us, vProcess: 8 us, vEncoding: 1480 us, aRender: 87 us, aEncoding: 768 us, Latency: 284613 us

    [15:30:47] Frame #52: vRender: 296431 us, vProcess: 8 us, vEncoding: 1368 us, aRender: 35 us, aEncoding: 770 us, Latency: 299523 us

    [15:30:47] Frame #53: vRender: 306640 us, vProcess: 8 us, vEncoding: 1406 us, aRender: 84 us, aEncoding: 826 us, Latency: 309024 us

    [15:30:47] Frame #54: vRender: 284094 us, vProcess: 9 us, vEncoding: 2254 us, aRender: 47 us, aEncoding: 793 us, Latency: 288227 us

    [15:30:47] Frame #55: vRender: 266128 us, vProcess: 9 us, vEncoding: 1712 us, aRender: 37 us, aEncoding: 739 us, Latency: 269607 us

    [15:30:48] Frame #56: vRender: 292414 us, vProcess: 8 us, vEncoding: 1308 us, aRender: 82 us, aEncoding: 825 us, Latency: 294690 us

    [15:30:48] Frame #57: vRender: 306988 us, vProcess: 8 us, vEncoding: 2051 us, aRender: 35 us, aEncoding: 761 us, Latency: 310921 us

    [15:30:48] Frame #58: vRender: 295473 us, vProcess: 10 us, vEncoding: 1601 us, aRender: 97 us, aEncoding: 798 us, Latency: 298050 us

    [15:30:49] Frame #59: vRender: 273355 us, vProcess: 8 us, vEncoding: 1482 us, aRender: 35 us, aEncoding: 662 us, Latency: 276481 us

    [15:30:49] Frame #60: vRender: 253383 us, vProcess: 8 us, vEncoding: 1339 us, aRender: 85 us, aEncoding: 738 us, Latency: 255609 us

    [15:30:49] Frame #61: vRender: 319213 us, vProcess: 9 us, vEncoding: 1658 us, aRender: 48 us, aEncoding: 1134 us, Latency: 323270 us

    [15:30:50] Frame #62: vRender: 285027 us, vProcess: 8 us, vEncoding: 1369 us, aRender: 87 us, aEncoding: 850 us, Latency: 287395 us

    [15:30:50] Frame #63: vRender: 315407 us, vProcess: 9 us, vEncoding: 1595 us, aRender: 49 us, aEncoding: 1060 us, Latency: 319410 us

    [15:30:50] Frame #64: vRender: 253384 us, vProcess: 8 us, vEncoding: 1669 us, aRender: 37 us, aEncoding: 617 us, Latency: 256635 us

    [15:30:50] Frame #65: vRender: 308628 us, vProcess: 8 us, vEncoding: 1414 us, aRender: 88 us, aEncoding: 766 us, Latency: 310963 us

    [15:30:51] Frame #66: vRender: 292440 us, vProcess: 8 us, vEncoding: 1465 us, aRender: 33 us, aEncoding: 675 us, Latency: 295615 us

    [15:30:51] Frame #67: vRender: 307551 us, vProcess: 8 us, vEncoding: 1570 us, aRender: 90 us, aEncoding: 682 us, Latency: 309956 us

    [15:30:51] Frame #68: vRender: 266196 us, vProcess: 8 us, vEncoding: 1566 us, aRender: 36 us, aEncoding: 815 us, Latency: 269521 us

    [15:30:52] Frame #69: vRender: 275330 us, vProcess: 8 us, vEncoding: 1583 us, aRender: 89 us, aEncoding: 742 us, Latency: 277812 us

    Einmal editiert, zuletzt von Svav (19. August 2019 um 07:54)

  • Vouk 18. August 2019 um 17:33

    Hat das Label Abgewiesen hinzugefügt.
  • I am not an expert in this program. Maybe this is because I used an SD card for reading and writing.

    My friend asked me how to speed up the rendering of a video. And I thought, why you can’t use a multiGPU system to render video. I know for sure that the CPU does not load 100%. When one video card is loaded at ~ 98%.

    I want to try rendering on my workstation. Where are used. 4 * Nvidia Tesla K10 8GB on 2 CPU X5680 and 48Gb ram.

    I think that in the present and in the future there will not be a big problem to use multiCPU systems for rendering. And it’s worth developing your program so that it can support multiGPU.

    • Offizieller Beitrag

    You can test this. Download and install "Vouk's Render Performance Test". It will tell you the maximum FPS that Premiere can deliver with exporting your project. Just Pure Premier performance, no voukoder, no disk i/o. It just renders the frames with all effects, filters and so on.

    Depending on that number of FPS we can think about choosing an encoding solution. Todays GPUs are pretty fast. A single RTX 2080 TI / RTX 6000 can encode like 400-500 FPS of 2K AVC video (depending on the settings). If you want more you should think about special hardware encoder solutions.

  • Vouk 19. August 2019 um 10:55

    Hat das Label von Abgewiesen auf Nicht möglich geändert.