| *** tpb <[email protected]> has joined #litex | 00:00 | |
| *** Degi_ <[email protected]> has joined #litex | 00:41 | |
| *** Degi <[email protected]> has quit IRC (Ping timeout: 252 seconds) | 00:42 | |
| *** Degi_ is now known as Degi | 00:42 | |
| *** shenchen <[email protected]> has quit IRC (Remote host closed the connection) | 01:11 | |
| *** nickoe <[email protected]> has quit IRC (Quit: Client closed) | 02:24 | |
| *** geertu <[email protected]> has quit IRC (Ping timeout: 260 seconds) | 02:26 | |
| *** acathla <[email protected]> has quit IRC (Ping timeout: 252 seconds) | 03:50 | |
| *** geertu <[email protected]> has joined #litex | 04:18 | |
| *** acathla <[email protected]> has joined #litex | 05:33 | |
| *** FabM <FabM!~FabM@2a03:d604:103:600:2e60:8c7c:e8fb:7990> has joined #litex | 06:20 | |
| pepijndevos[m] | Maybe try poking gowin... | 06:37 |
|---|---|---|
| *** nickoe <[email protected]> has joined #litex | 07:06 | |
| nickoe | Is it possible to define pins wihout them being suffixed wiht the number as in: ("io_in", 0, Pins(8)), becomes "output reg [7:0] io_out0" | 07:10 |
| nickoe | It gives me an exception when I do ("io_in", None, Pins(8)), | 07:11 |
| *** Brinx <[email protected]> has quit IRC (Remote host closed the connection) | 07:48 | |
| *** Brinx <[email protected]> has joined #litex | 08:31 | |
| *** Brinx <[email protected]> has quit IRC (Ping timeout: 252 seconds) | 08:35 | |
| *** Brinx <[email protected]> has joined #litex | 08:41 | |
| *** nickoe <[email protected]> has quit IRC (Quit: Client closed) | 09:39 | |
| *** TMM_ <[email protected]> has quit IRC (Quit: https://quassel-irc.org - Chat comfortably. Anywhere.) | 10:17 | |
| *** TMM_ <[email protected]> has joined #litex | 10:17 | |
| *** indy <[email protected]> has quit IRC (Quit: ZNC 1.8.2 - https://znc.in) | 10:26 | |
| *** indy <[email protected]> has joined #litex | 10:37 | |
| *** acathla <[email protected]> has quit IRC (Ping timeout: 244 seconds) | 10:41 | |
| *** acathla <[email protected]> has joined #litex | 10:43 | |
| _florent_ | cr1901: for the TEC0117, the sipped SDRAM should be a MT48LC8M16A2 | 11:00 |
| _florent_ | (that's what I got from Gowin) | 11:00 |
| *** acathla <[email protected]> has quit IRC (Ping timeout: 268 seconds) | 11:02 | |
| *** indy <[email protected]> has quit IRC (Ping timeout: 252 seconds) | 11:10 | |
| *** indy <[email protected]> has joined #litex | 11:15 | |
| *** indy <[email protected]> has quit IRC (Ping timeout: 268 seconds) | 11:34 | |
| *** Brinx <[email protected]> has quit IRC (Remote host closed the connection) | 11:36 | |
| *** Brinx <[email protected]> has joined #litex | 11:41 | |
| *** Brinx <[email protected]> has quit IRC (Remote host closed the connection) | 11:44 | |
| *** Brinx <[email protected]> has joined #litex | 11:44 | |
| *** Brinx <[email protected]> has quit IRC (Remote host closed the connection) | 11:44 | |
| *** Brinx <[email protected]> has joined #litex | 11:45 | |
| *** Brinx <[email protected]> has quit IRC (Remote host closed the connection) | 12:01 | |
| *** Brinx <[email protected]> has joined #litex | 12:10 | |
| *** Brinx <[email protected]> has quit IRC (Remote host closed the connection) | 12:13 | |
| *** Brinx_ <[email protected]> has joined #litex | 12:13 | |
| *** shenchen <[email protected]> has joined #litex | 13:00 | |
| *** Brinx_ <[email protected]> has quit IRC (Remote host closed the connection) | 13:46 | |
| *** Brinx <[email protected]> has joined #litex | 14:17 | |
| *** Brinx <[email protected]> has quit IRC (Ping timeout: 260 seconds) | 14:21 | |
| *** Brinx <[email protected]> has joined #litex | 14:37 | |
| tnt | I'm a bit perplexed by some DMA performance results. | 14:47 |
| tnt | I have a block generating fake data, one word of the data path width every N cycles (with N being programmable). | 14:47 |
| tnt | That goes into a small fifo ( 2048 entries ) and then directly to the DMA sink. | 14:48 |
| tnt | Measured internal loopback performance is 33 Gbps/s. I set my block to generate 17.06 Gbps. | 14:49 |
| tnt | If I do a fwrite() of the buffer I receive to a file in RAMfs, I do measure 17.06G perf as expected (for as long as the mem isn't filled at least). | 14:49 |
| tnt | If I remove the fwrite (no other changes), ... bandwidht goes to 0.16 Gbps. | 14:50 |
| tnt | wtf ... | 14:50 |
| *** indy <[email protected]> has joined #litex | 15:39 | |
| *** Brinx <[email protected]> has quit IRC (Remote host closed the connection) | 15:47 | |
| cr1901 | pepijndevos[m]: I might have to poke gowin... what email address did you use? | 15:53 |
| cr1901 | I suspect the DRAM module is different for the GW1NR-4 | 15:53 |
| pepijndevos[m] | [email protected] I think but... I'm not the one who got the timing info so maybe that should count as a negative advice | 16:01 |
| cr1901 | cc: _florent_ I suspect the DRAM module is different for the GW1NR-4 b/c it has a 32-bit mode as well as a 16-bit mode, and "plugging in" the MT48LC8M16 module only gets me a "half working" DRAM controller | 16:04 |
| cr1901 | (i.e. some of the memory stores/read succeed, some don't) | 16:05 |
| tnt | You got a free RNG :) | 16:08 |
| cr1901 | Which would be great if I was designing an RNG, but I'm not :( | 16:09 |
| *** FabM <FabM!~FabM@armadeus/team/FabM> has quit IRC (Quit: Leaving) | 16:24 | |
| *** shenchen <[email protected]> has quit IRC (Remote host closed the connection) | 16:24 | |
| *** shenchen <[email protected]> has joined #litex | 16:24 | |
| *** Brinx <[email protected]> has joined #litex | 16:34 | |
| *** Brinx <[email protected]> has quit IRC (Remote host closed the connection) | 16:35 | |
| *** Brinx <[email protected]> has joined #litex | 16:35 | |
| *** Johnsel <[email protected]> has joined #litex | 17:25 | |
| tnt | fifo.sink is where I push data to right ? | 18:26 |
| _florent_ | tnt: yes (regarding fifo.sink) | 19:49 |
| _florent_ | cr1901: so 16-bit is working and 32-bit is not? | 19:49 |
| _florent_ | cr1901: you can also eventually try to use the PHYPadsReducer to only compile with some of the byte groups | 19:50 |
| _florent_ | https://github.com/litex-hub/litex-boards/search?q=phypadsreducer | 19:51 |
| cr1901 | _florent_: 16-bit isn't working properly either... using that gist you gave me a few days ago... (1 sec) | 19:53 |
| cr1901 | http://gopher.wdj-consulting.com:70/paste/1d58ba8b-997a-4656-b599-ceeeb8919912.txt <-- I made this python script | 19:54 |
| cr1901 | http://gopher.wdj-consulting.com:70/paste/2458da78-1f7e-49ef-9065-5f942639a47f.txt And this is the result of a write... it's like only 8 bits are being written and read back for each 32-bit word | 19:55 |
| _florent_ | cr1901: BTW, be sure to disable the L2 cache when testing this, it will ease understanding the behavior :) | 19:57 |
| _florent_ | like this: https://github.com/litex-hub/litex-boards/commit/6c7157f7996b58fdf8b10fd6e22881c3fe252df9 | 19:58 |
| *** nickoe <[email protected]> has joined #litex | 19:58 | |
| nickoe | I will try again, not sure if I missed a reply. | 19:58 |
| cr1901 | _florent_: Ty for the tip, cache is disabled already :D | 19:58 |
| nickoe | Is it possible to define pins wihout them being suffixed wiht the number as in: ("io_in", 0, Pins(8)), becomes "output reg [7:0] io_out0" | 19:59 |
| nickoe | It gives me an exception when I do ("io_in", None, Pins(8)), | 19:59 |
| _florent_ | nickoe: If you only have one "io_out" in your design, not sure the numbered suffix will be added (I'm not on a dev machine so could check tomorrow) | 20:03 |
| nickoe | yes, I have only one | 20:06 |
| nickoe | even if I browse the simulated output | 20:06 |
| nickoe | in gtkwave. | 20:06 |
| nickoe | But for ("sys_rst", 0, Pins(1)), it does not get added. | 20:07 |
| nickoe | I am not sure why. | 20:07 |
| nickoe | There is this that appears to handle it for the request | 20:08 |
| nickoe | https://github.com/enjoy-digital/litex/blob/ea8ba57eabc0d47fe98e8b45227c61491702d7a3/litex/build/generic_platform.py#L213-L216 | 20:08 |
| nickoe | But ... now I see that I use request_all for those that get a suffix. | 20:08 |
| nickoe | It goes into the exception here, https://github.com/enjoy-digital/litex/blob/ea8ba57eabc0d47fe98e8b45227c61491702d7a3/litex/build/generic_platform.py#L243 whre it breks and skips the ValueError | 20:11 |
| _florent_ | can you check if you also have this with just "request" | 20:15 |
| _florent_ | ? | 20:15 |
| nickoe | mm, not sure what to break on there. | 20:16 |
| nickoe | in request_all, r will be zero in len if I have the tuple as ("io_out", None, Pins(8)) | 20:17 |
| _florent_ | None is probably not supported. What do you get with 0 and platform.request? | 20:18 |
| nickoe | It goes into | 20:19 |
| nickoe | if number is None: | 20:19 |
| nickoe | resource_name = name | 20:19 |
| nickoe | Ah, I see, r.append(self.request(name, len(r)))... that will sorta alwayus set a "non None" "number" in request | 20:20 |
| nickoe | I guess this is not really pythonic, but I if patch it like so, it appears to do what I would have expected. https://dpaste.com/8N89G464Z | 20:23 |
| tpb | Title: dpaste: 8N89G464Z (at dpaste.com) | 20:23 |
| nickoe | or really, in the second part under "was" it was len(r) for the number variable. | 20:24 |
| tnt | Anyone familiar with linux perf tuning maybe ? I get DMA bw dip every second on the dot ... (causing overflows) | 20:34 |
| nickoe | I am not familiar with perf, but that is "the dot"? | 20:34 |
| nickoe | ah, you mean _exactly_ | 20:35 |
| nickoe | .. | 20:35 |
| tnt | yeah :) | 20:35 |
| nickoe | When it has None, it does not appear to be in the self.matched list. | 20:38 |
| *** acathla <[email protected]> has joined #litex | 21:21 | |
| *** nickoe <[email protected]> has quit IRC (Quit: Client closed) | 22:13 | |
| *** nickoe <[email protected]> has joined #litex | 22:14 | |
| *** Johnsel <[email protected]> has quit IRC (Ping timeout: 268 seconds) | 22:32 | |
| *** Johnsel <[email protected]> has joined #litex | 22:33 | |
| *** TMM_ <[email protected]> has quit IRC (Quit: https://quassel-irc.org - Chat comfortably. Anywhere.) | 23:05 | |
| *** TMM_ <[email protected]> has joined #litex | 23:05 | |
Generated by irclog2html.py 2.17.2 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!