| *** tpb <[email protected]> has joined #litex | 00:00 | |
| *** Coldberg <[email protected]> has quit IRC (Ping timeout: 256 seconds) | 01:04 | |
| *** jediminer543_ <[email protected]> has quit IRC (Read error: Connection reset by peer) | 01:49 | |
| *** Degi_ <[email protected]> has joined #litex | 02:32 | |
| *** Degi <[email protected]> has quit IRC (Ping timeout: 265 seconds) | 02:33 | |
| *** Degi_ is now known as Degi | 02:33 | |
| *** aquijoule_ <[email protected]> has joined #litex | 03:18 | |
| *** aquijoule__ <[email protected]> has quit IRC (Ping timeout: 258 seconds) | 03:20 | |
| *** alanvgreen_ <[email protected]> has joined #litex | 03:44 | |
| alanvgreen_ | I'm using a ULX3S board, and the video frame buffer. The video signal seems to drop while doing intensive memory access. I'm guessing that's due to a buffer underrun. The SoC sys_clk_freq is limited to about 50MHz. Is there a convenient way to increase the SDRAM speed to 100MHz, while keeping sys_clk_freq at 50MHz? | 04:13 |
|---|---|---|
| *** Leon[m] <Leon[m]!~leons@2001:470:69fc:105::abc> has joined #litex | 06:42 | |
| *** alanvgreen_ <[email protected]> has quit IRC (Quit: Connection closed for inactivity) | 06:43 | |
| *** FabM <[email protected]> has joined #litex | 06:56 | |
| mupuf | _florent_: Nice stories on twitter :D | 07:42 |
| *** sajattack[m] <sajattack[m]!~sajattack@2001:470:69fc:105::1d9> has joined #litex | 07:58 | |
| *** aquijoule_ <[email protected]> has quit IRC (Quit: Leaving) | 08:10 | |
| *** aquijoule_ <[email protected]> has joined #litex | 08:10 | |
| *** TMM_ <[email protected]> has quit IRC (Quit: https://quassel-irc.org - Chat comfortably. Anywhere.) | 08:16 | |
| *** TMM_ <[email protected]> has joined #litex | 08:16 | |
| *** alanvgreen_ <[email protected]> has joined #litex | 08:17 | |
| *** vomoniyi[m] <vomoniyi[m]!~vomoniyig@2001:470:69fc:105::3023> has joined #litex | 08:22 | |
| alanvgreen_ | Answer to me: This is exactly what the sdram_rate parameter does. Use sdram_rate="1:2" to get a double speed clock, which is exactly the 100MHz I thought I wanted. BIOS reports approximately the same read and write speeds, but HDMI signal no longer drops. Hooray! | 08:39 |
| *** cr1901 <cr1901!~William@2601:8d:8600:911:966:a289:368:199d> has quit IRC (Ping timeout: 240 seconds) | 08:50 | |
| *** cr1901 <cr1901!~William@2601:8d:8600:911:c012:1f45:543:1608> has joined #litex | 08:52 | |
| *** Coldberg <[email protected]> has joined #litex | 09:30 | |
| gatecat | alanvgreen_: the read and write speeds in the BIOS are as measured by the CPU and depend a lot on CPU performance - they're not DMA speeds or anything | 09:35 |
| gatecat | so that they don't change isn't a problem here | 09:35 |
| _florent_ | mupuf: Thanks, I was procrastinating a bit :) | 09:51 |
| _florent_ | alanvgreen_: 1:2 SDRAM ratio will indeed provide you 2X SDRAM bandwidth, but as gatecat said, it's possible you won't see it from the BIOS Memtest since the CPU is already the bottleneck with the 1:1 ratio | 09:52 |
| _florent_ | if you call add_sdram with "with_bist=True", you'll have a sdram_bist command in the BIOS that you could use to measure the SDRAM bandwidth | 09:54 |
| _florent_ | https://github.com/enjoy-digital/litex/blob/master/litex/soc/integration/soc.py#L1225 | 09:54 |
| _florent_ | this will add a hardware generator/checker, useful to do initial tests on SDRAM, but it's using resources, so generally disabled on designs once DRAM is validated | 09:55 |
| _florent_ | Otherwise, for the VideoFramebuffer, you can also play with the fifo_depth to compensate temporary unavailability of the DRAM: https://github.com/enjoy-digital/litex/blob/master/litex/soc/cores/video.py#L609 | 09:56 |
| _florent_ | the default value should be a good compromise between resource usage/amount of buffering, but that's adjusting it can probably be interesting on some systems | 09:57 |
| *** promach[m] <promach[m]!~promach@2001:470:69fc:105::ca1> has quit IRC (Quit: Bridge terminating on SIGTERM) | 10:28 | |
| *** Leon[m] <Leon[m]!~leons@2001:470:69fc:105::abc> has quit IRC (Quit: Bridge terminating on SIGTERM) | 10:28 | |
| *** sajattack[m] <sajattack[m]!~sajattack@2001:470:69fc:105::1d9> has quit IRC (Quit: Bridge terminating on SIGTERM) | 10:28 | |
| *** vomoniyi[m] <vomoniyi[m]!~vomoniyig@2001:470:69fc:105::3023> has quit IRC (Quit: Bridge terminating on SIGTERM) | 10:28 | |
| *** jryans <jryans!~jryansmat@2001:470:69fc:105::1d> has joined #litex | 10:29 | |
| *** shoragan[m] <shoragan[m]!~shoraganm@2001:470:69fc:105::39> has joined #litex | 10:30 | |
| *** jevinskie[m] <jevinskie[m]!~jevinskie@2001:470:69fc:105::bb3> has joined #litex | 10:30 | |
| *** dcallagh <dcallagh!~dcallagh@2001:470:69fc:105::9c5> has joined #litex | 10:30 | |
| *** CarlosEDP <CarlosEDP!~carlosedp@2001:470:69fc:105::218e> has joined #litex | 10:30 | |
| *** promach[m] <promach[m]!~promach@2001:470:69fc:105::ca1> has joined #litex | 10:30 | |
| *** Leon[m] <Leon[m]!~leons@2001:470:69fc:105::abc> has joined #litex | 10:30 | |
| *** DerekKozel[m] <DerekKozel[m]!~dkozelgnu@2001:470:69fc:105::2f14> has joined #litex | 10:31 | |
| *** vomoniyi[m] <vomoniyi[m]!~vomoniyig@2001:470:69fc:105::3023> has joined #litex | 10:31 | |
| *** sajattack[m] <sajattack[m]!~sajattack@2001:470:69fc:105::1d9> has joined #litex | 10:31 | |
| *** david-sawatzke[m <david-sawatzke[m!~david-saw@2001:470:69fc:105::1634> has joined #litex | 10:31 | |
| *** alanvgreen_ <[email protected]> has quit IRC (Quit: Connection closed for inactivity) | 11:23 | |
| *** pftbest <[email protected]> has quit IRC (Remote host closed the connection) | 13:10 | |
| *** alanvgreen_ <[email protected]> has joined #litex | 13:17 | |
| alanvgreen_ | gatecat: Makes sense, thanks for the explanation | 13:20 |
| alanvgreen_ | _florent_: Thank you for the pointers to the bist function and fifo depth parameter! | 13:21 |
| *** pftbest <pftbest!~pftbest@2a01:4f8:c17:6afc::1:2> has joined #litex | 13:22 | |
| *** Coldberg <[email protected]> has quit IRC (Ping timeout: 268 seconds) | 13:39 | |
| *** lkcl <[email protected]> has quit IRC (Ping timeout: 265 seconds) | 14:08 | |
| *** lkcl <[email protected]> has joined #litex | 14:28 | |
| Melkhior | _florent_: is there some documentation explaining the sdram initialization process ? | 14:33 |
| Melkhior | Trying to understand it, but in 'sdram_write_latency_calibration' I have test 'if (_sdram_write_leveling_bitslips[module] < 0)', and that global array is never initialized in my case (SDRAM_PHY_WRITE_LEVELING_CAPABLE is not set) | 14:34 |
| Melkhior | So the array sould be all-O (it's a global), so the test is always false, so 'bitslip' is always set to the array content, i.e; to 0 ... making the test loop somewhat redundant | 14:36 |
| Melkhior | Must be missing something ... | 14:36 |
| *** awordnot <awordnot!~awordnot@user/awordnot> has quit IRC (Ping timeout: 265 seconds) | 14:41 | |
| *** awordnot <awordnot!~awordnot@user/awordnot> has joined #litex | 14:41 | |
| *** FabM <FabM!~FabM@armadeus/team/FabM> has quit IRC (Quit: Leaving) | 15:00 | |
| *** Coldberg <[email protected]> has joined #litex | 15:21 | |
| *** alanvgreen_ <[email protected]> has quit IRC (Quit: Connection closed for inactivity) | 15:36 | |
| *** OmkarBhilare[m] <OmkarBhilare[m]!~ombhilare@2001:470:69fc:105::4e51> has joined #litex | 19:21 | |
| OmkarBhilare[m] | Hello, | 19:28 |
| OmkarBhilare[m] | I wanted to use litedram core in my Google summer of code project under beagleboard org. | 19:28 |
| OmkarBhilare[m] | The project is BeagleWire Software. The BeagleWire is a FPGA cape for beaglebone black based on iCE40HX. The cape has 32 MB SDRAM. I wanted to produce standalone verilog code for SDRAM controller without the CPU. I have been exploring the litedram but had couple of doubts related to it. | 19:28 |
| OmkarBhilare[m] | Can I ask my doubts here? | 19:29 |
| *** aquijoule_ <[email protected]> has quit IRC (Quit: Leaving) | 20:31 | |
| *** aquijoule_ <[email protected]> has joined #litex | 20:32 | |
| *** aquijoule_ is now known as richbridger | 20:32 | |
| *** richbridger <[email protected]> has quit IRC (Remote host closed the connection) | 20:36 | |
| *** richbridger <[email protected]> has joined #litex | 20:37 | |
| jevinskie[m] | Yes please. Don’t need to ask to ask :) | 20:55 |
| OmkarBhilare[m] | Actually my rest of the Designs were in Verilog. So I wanted to produce standalone verilog code with CPU as none. | 21:00 |
| OmkarBhilare[m] | So wanted to ask does litedram can produce the core without the cpu, just exposing the wishbone will do. | 21:00 |
| OmkarBhilare[m] | https://github.com/enjoy-digital/litedram/blob/master/litedram/gen.py | 21:00 |
| OmkarBhilare[m] | For standalone I was looking into this. | 21:00 |
| OmkarBhilare[m] | https://github.com/enjoy-digital/litedram/tree/master/examples | 21:00 |
| OmkarBhilare[m] | I see some yaml examples here | 21:00 |
| *** pftbest <pftbest!~pftbest@2a01:4f8:c17:6afc::1:2> has quit IRC (Remote host closed the connection) | 21:06 | |
| *** pftbest <pftbest!~pftbest@2a01:4f8:c17:6afc::1:2> has joined #litex | 21:11 | |
| *** pftbest <pftbest!~pftbest@2a01:4f8:c17:6afc::1:2> has quit IRC (Remote host closed the connection) | 21:35 | |
| *** alanvgreen_ <[email protected]> has joined #litex | 22:05 | |
| *** TMM_ <[email protected]> has quit IRC (Quit: https://quassel-irc.org - Chat comfortably. Anywhere.) | 22:58 | |
| *** TMM_ <[email protected]> has joined #litex | 22:58 | |
Generated by irclog2html.py 2.17.2 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!