*** tpb has joined #litex | 00:00 | |
*** HoloIRCUser2 has joined #litex | 00:23 | |
*** HoloIRCUser has quit IRC | 00:23 | |
*** Skip has quit IRC | 01:05 | |
*** HoloIRCUser has joined #litex | 01:53 | |
*** HoloIRCUser2 has quit IRC | 01:57 | |
*** Degi has quit IRC | 03:05 | |
*** Degi has joined #litex | 03:06 | |
*** CarlFK has joined #litex | 03:36 | |
*** HoloIRCUser2 has joined #litex | 06:38 | |
*** HoloIRCUser has quit IRC | 06:41 | |
*** HoloIRCUser has joined #litex | 06:48 | |
*** HoloIRCUser2 has quit IRC | 06:51 | |
*** HoloIRCUser has quit IRC | 07:12 | |
*** st-gourichon-fid has quit IRC | 09:52 | |
*** st-gourichon-fid has joined #litex | 09:52 | |
*** st-gourichon-fid has quit IRC | 10:00 | |
*** st-gourichon-fid has joined #litex | 10:02 | |
*** st-gourichon-fid has quit IRC | 10:09 | |
*** st-gourichon-fid has joined #litex | 10:13 | |
kbeckmann | 1440p60 would have been fun but it seems I can't go above 800 MHz on the PLL, at least not using the algorithm in ecppll. Maybe it's possible to generate a PLL config that does that anyway :). | 10:43 |
---|---|---|
zyp | the ecp5 pll is specced to a VCO freq between 400-800MHz, so a config generator would naturally try to keep within those bounds to generate a config valid to spec | 10:51 |
zyp | but if you pick your own factors, nothing prevents you from exceeding that | 10:53 |
zyp | things will just start acting weid once you get too far out of spec :) | 10:54 |
zyp | weird | 10:54 |
*** CarlFK has quit IRC | 12:39 | |
*** proteusguy has quit IRC | 13:04 | |
*** acathla has quit IRC | 13:06 | |
*** FFY00 has quit IRC | 13:15 | |
*** FFY00 has joined #litex | 13:16 | |
*** proteusguy has joined #litex | 13:17 | |
*** scanakci has quit IRC | 13:18 | |
*** CarlFK has joined #litex | 13:21 | |
*** CarlFK has quit IRC | 13:45 | |
*** CarlFK has joined #litex | 13:45 | |
*** kgugala_ has quit IRC | 13:56 | |
*** kgugala has joined #litex | 13:56 | |
kbeckmann | zyp: thanks, I have to try it out. This is already out of spec since the output buffers are rated at 800 MHz and I use them at 1500 (!) | 14:08 |
tnt | Mbps not MHz. | 14:09 |
zyp | which output buffers? aren't they rated for 400 MHz? | 14:09 |
tnt | yeah 800 Mbps i.e. 400 MHz. | 14:09 |
zyp | I thought top/bottom banks would do 200 and side banks would do 400 | 14:09 |
zyp | ah, because DDR, right | 14:10 |
*** acathla has joined #litex | 14:38 | |
*** acathla has quit IRC | 14:39 | |
*** acathla has joined #litex | 14:39 | |
*** darren099 has joined #litex | 14:49 | |
*** CarlFK has quit IRC | 15:19 | |
*** Stary- is now known as Stary | 16:28 | |
*** Skip has joined #litex | 17:23 | |
loxodes | I'm trying to stream samples from an ADC into HyperRAM memory that is attached to a litex soc, do y'all have any advice on how to approach this? | 17:25 |
loxodes | On a board with DRAM I've used LiteDRAMDMAWriter to haul samples from a module into memory, but that doesn't support HyperRAM. | 17:25 |
loxodes | Is there a more general library for streaming data into memory from a module, or should I build something from scrach myself? (I'm inexperienced and mostly work by combining examples, in this case I'd be looking at wishbonebridge for writing to wishbone and LiteDRAMDMAWriter?) | 17:25 |
zyp | not sure if there are already any stream to wishbone dma modules, but it shouldn't be too hard making one | 17:41 |
_florent_ | loxodes: we don't have a DMA <--> Wishbone currently, but it would indeed be useful for some use cases | 18:06 |
loxodes | okay, thanks! | 18:07 |
_florent_ | loxodes: creating one similar to LiteDRAMDMAWriter but targeting Wishbone is a good idea, if you want to create it and need help, i'm happy to help | 18:07 |
*** FFY00 has quit IRC | 18:16 | |
*** FFY00 has joined #litex | 18:17 | |
*** kgugala has quit IRC | 18:49 | |
*** kgugala has joined #litex | 18:49 | |
*** darren099 has quit IRC | 20:32 |
Generated by irclog2html.py 2.17.2 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!