| 02:00 | mumptai | left the channel |
| 04:41 | Bertl | off to bed now ... have a good one everyone!
|
| 04:41 | Bertl | changed nick to: Bertl_zZ
|
| 06:16 | BAndiT1983|away | changed nick to: BAndiT1983
|
| 07:58 | fredy__ | joined the channel |
| 07:59 | fredy | left the channel |
| 08:41 | mumptai | joined the channel |
| 09:33 | schmoggie1 | left the channel |
| 11:16 | Dest123 | joined the channel |
| 11:55 | anuejn | some more experiments resulted in this plot: https://files.niemo.de/fifo_interface_idle.svg
|
| 11:56 | anuejn | interestingly the "bursts" the ft601 accepts are always 2k in size
|
| 12:08 | mumptai | left the channel |
| 12:09 | mumptai | joined the channel |
| 12:10 | mumptai | left the channel |
| 12:10 | anuejn | when we force the burst size to 1k we get pretty exactly half the bandwidth
|
| 12:10 | anuejn | (180Mbyte /s)
|
| 12:28 | se6ast1an | anuejn: so this all seems to hint at 360MB/s being some kind of hard limit I assume?
|
| 12:36 | vup | we think it has to do with the mode we are using, there are two modes: 245 and 600
|
| 12:37 | vup | currently we are using the 245 mode, and this indeed looks like it might be kind of a hard limit in this mode
|
| 12:39 | vup | however we think using the 600 mode it might be possible to get better performance, and one can then have multiple parallel channels, which means that even if the ft601 is busy sending the packets for one of the channels, it might accept new data on one of the other channels and thus decreasing the number of cycles we are not sending anything
|
| 12:41 | anuejn | but we are not shure if it is worth the labour to put into that performance optimization
|
| 12:41 | anuejn | even if we would get like 380 Mb/s that is not much more and probably quite some additional work
|
| 12:42 | se6ast1an | I see
|
| 12:42 | se6ast1an | sustainable 360MB/s is indeed a very good value for one plugin module
|
| 12:43 | se6ast1an | which we could double by using 2
|
| 12:45 | anuejn | ok so it would be sufficient for doing 16:9 4k 12bit 24fps (which is 339.7386 megabyte/s)
|
| 12:45 | anuejn | so i think it is not worth the additional time investment
|
| 13:00 | se6ast1an | agreed
|
| 13:01 | se6ast1an | what steps are required to actually get video data to/from the usb3 plugin module?
|
| 13:01 | se6ast1an | I assume it require zynq gateware changes as well
|
| 13:02 | mumptai | joined the channel |
| 13:41 | anuejn | probably
|
| 13:41 | anuejn | and a working usb3 plugin
|
| 13:41 | anuejn | and some gataware on the machxo
|
| 14:06 | Dest123 | left the channel |
| 14:35 | danieeel | changed nick to: danieel
|
| 14:38 | anuejn | ft60x-rs should be usable now :)
|
| 14:38 | anuejn | https://github.com/apertus-open-source-cinema/ft60x-rs
|
| 14:39 | se6ast1an | great!
|
| 14:45 | Bertl_zZ | changed nick to: Bertl
|
| 14:45 | Bertl | morning folks!
|
| 14:48 | se6ast1an | good day!
|
| 14:52 | anuejn | morning :)
|
| 15:02 | Dest123 | joined the channel |
| 15:36 | Dest123 | left the channel |
| 17:23 | BAndiT1983 | changed nick to: BAndiT1983|away
|
| 19:54 | Bertl | off for now ... bbl
|
| 19:54 | Bertl | changed nick to: Bertl_oO
|
| 20:29 | BAndiT1983|away | changed nick to: BAndiT1983
|
| 20:57 | illwieckz | left the channel |
| 20:58 | illwieckz | joined the channel |
| 22:32 | Dest123 | joined the channel |
| 22:49 | BAndiT1983 | changed nick to: BAndiT1983|away
|
| 23:32 | Dest123 | left the channel |
| 00:29 | mumptai | left the channel |