sb0 changed the topic of #m-labs to: ARTIQ, Migen, MiSoC, Mixxeo & other M-Labs projects :: fka #milkymist :: Logs http://irclog.whitequark.org/m-labs
<cyrozap>
What does it mean if I don't get any memtest errors during the ONEZERO and ZEROONE tests, but I get 262144 errors on the random data test and 4096 errors on the random write test?
<mithro>
sb0: so, flterm.py doesn't appear to work with my (probably broken) serial port, while flterm.c does
<mithro>
sb0: it just dies with "serial.serialutil.SerialException: device reports readiness to read but returned no data (device disconnected or multiple access on port?)"
<sb0>
mithro, haven't seen that ...
<sb0>
_florent_, ?
<mithro>
sb0: the serial port is probably broken, IE it is probably doing exactly what the above says - but it seems to work fine with flterm.c
<sb0>
whitequark, i just received a chinese angle valve. even though it has a label that says "RIP OFF NO RETURN", it looks very good and was rather inexpensive.
<sb0>
let me do a quick vac test...
stekern has quit [Ping timeout: 240 seconds]
<sb0>
yeah it has no gross problems
<cr1901_modern>
They couldn't have chosen a better void-warranty label statement. Unless it was in fact truth-in-advertising.
FabM has joined #m-labs
<sb0>
oh, actually the crappy ball valve can probably be used to vent the small volume between the forepump and the angle valve
aeris has quit [Read error: Connection reset by peer]
aeris has joined #m-labs
stekern has joined #m-labs
<_florent_>
cyrozap: in memtest, you can add a print in hexa of the expected value and the read value
<_florent_>
just keep the data test and reduce the size of the test
<_florent_>
then if you see that data is swapped, you can try to play with clock phase or rd_bitslip and wr_bitslip
<cr1901_modern>
Why would data be swapped if there was a bitslip? (I've never actually implemented a DRAM controller; it's on my todo list)
<cr1901_modern>
Also _florent_, I know the minispartan6+ made poor design choices, but what's the maximum HDMI resolution it can do?
<_florent_>
mithro: I never had that, it seems related to pyserial that is probably doing more checks than flterm
<cr1901_modern>
s/made/was implemented with/
<_florent_>
cr1901_modern: bitslip is on the iserdes and oserdes, but you have 16 serdes in //. You have to be sure the DRAM is sampling correcly your data: you adjust clk phase and wr_bitslip and that the FPGA is also sampling correctly the data returned from the DRAM: you adjust rd_bitslip
<_florent_>
cr1903_modenr: it depends what you want to do with the minispartan6+ but you will be limited by the DRAM's bandwidth
rohitksingh has joined #m-labs
<cr1901_modern>
What's the bandwidth for minispartan? 75MHz clock, right?
<cr1901_modern>
1920*1080*60 is about 125MHz, which is only an estimate
<mithro>
_florent_: yeah - I'm also getting a behavior with pyserial where data gets "stuck" in the incoming/outgoing buffers (don't really know how else to describe it)
<mithro>
running the C version of flterm seems to cause it to flush
rohitksingh has quit [Ping timeout: 264 seconds]
rohitksingh has joined #m-labs
wbx has quit [Ping timeout: 255 seconds]
rohitksingh has quit [Ping timeout: 265 seconds]
wbx has joined #m-labs
rohitksingh has joined #m-labs
cr1901_modern has left #m-labs [#m-labs]
cr1901_modern has joined #m-labs
<cr1901_modern>
You know what would be great? If MSYS's bash implementation didn't leak memory after its children died