Lofty changed the topic of #prjmistral to: Project Mistral: Yosys (and hopefully nextpnr) on Cyclone FPGAs - https://github.com/ZirconiumX/mistral - logs: https://freenode.irclog.whitequark.org/prjmistral
<jordigw> Lofty: Which board to use to try to contribute ?
<bwidawsk> jordigw: de10-nano is a pretty good choice
<jordigw> bwidawsk: it's a cycloneV Soc ?
<jordigw> If so I have de0nanoSoc & de1Soc -> I suppose these boards are usable too
<Lofty> jordigw: we're focusing primarily on the Cyclone V. The DE0-Nano is Cyclone IV, but the DE1-SoC should be okay
<Lofty> The board myself and Sarayan are using is the DE10-Nano though
<jordigw> Lofty: not de0nano but de0nanoSoc <- with soc
<Lofty> Ah, okay
<jordigw> a bit cheaper :)
trabucayre has joined #prjmistral
<Lofty> Eh, not by much
<jordigw> "a bit" :)
<Lofty> The whole thing that sparked project Mistral was the MiSTer project emulators
<trabucayre> Hi
<Lofty> And those are designed for the DE10-Nano
<Lofty> Hello
<trabucayre> I try to find some informations about sof file format and about JTAG instructions.
<Lofty> SOF is basically RBF with some metadata
<jordigw> Okay. It seems it's not the same model :-(
<Lofty> No, but there are only 5 dies in the Cyclone V family, if I remember correctly
<Lofty> So you might end up with a pretty serious upgrade
<jordigw> three boards three different version :-/
<Lofty> Again: not really
<Lofty> I'm pretty sure when we properly RE this that the chips will all be the same: the sx150f die
<jordigw> I will try.
<Lofty> As for JTAG; that's presently unknown information, sorry
<trabucayre> I have analyzed some part of the sof file but I'm unable to do the link between this file and svf
<trabucayre> :(
<daveshah> It's possible they split out the commands somehow before creating the svf?
<trabucayre> maybe
<trabucayre> I need, maybe, to dump jtag traffic with wireshark (as I done for gowin FPGA)...
<trabucayre> (and anlogic)
<Lofty> Oh, now I recognise the username
<Lofty> You wrote openFPGALoader, right?
<trabucayre> Lofty: true :)
<Lofty> pepijndevos mentioned you I think
<trabucayre> me I don't know but openFPGALoader yes :)
<Sarayan> Lofty 7 dies, not 5
<Lofty> ...I swear it was 5, but okay
<Sarayan> And I'm tackling all 7, so that I generalize correctly anyway
<Sarayan> so as long as it's cyclone v, it's good
<Sarayan> I may have to do cyclone10 at some point, but later :-)
<Sarayan> after cv is done and buried
<trabucayre> Sarayan: I'm interested by cyclone10 :)
<trabucayre> cycloneV too in fact
<daveshah> Do you know if there are differences in the internal data formats?
<daveshah> or would the differences be more on the Yosys/nextpnr side (at least in the case of 10lp)?
<Sarayan> I haven't looked at 10 at all
<Sarayan> but it's used in the future analog pocket, hence my interest
<Sarayan> Lofty: look at the 7 subdirectories in quartus/common/devinfo/cyclonev/cyclonev_*, these are the dies
<daveshah> Oh, if anyone wants to play with big devices, it looks like Quartus Pro is available on Azure (similar to the Xilinx deal with AWS for Vivado)
<Sarayan> I can't afford a big device :-)
<Lofty> daveshah: 10 LP is quite different to 10 GX
<daveshah> 10 GX is closer to Cyclone V, right?
<Lofty> Correct
<Lofty> 10 LP is more like Cyclone IV
<daveshah> yeah
<daveshah> I think Cyc IV and III are the same RTL with different standard cells, as their JTAG IDCODE are the same (I remember having the pick between the two in Quartus programmer when using Cyc III 5 years ago)
<trabucayre> daveshah: true. Same idcode ...
<daveshah> It would be odd to make substantial design changes but keep the IDCODE the same
<daveshah> Unless someone just forgot to change it
<trabucayre> I've this issue for openFPGALoader. Not possible to determine which fpga
<daveshah> Quartus has a dialog box...
<Sarayan> anyone knows of a cyclonev idcode table?
<Lofty> Sarayan: knowing you, one will magically appear somehow :P
<daveshah> Is there one in one of the datasheets? Xilinx and Lattice both have a "configuration user guide" type document with them
<Lofty> daveshah: so there is
<Lofty> Table 9-1
<Lofty> trabucayre: ^
<trabucayre> ah yes!
<trabucayre> but some IR used in svf file is not mentionned in this ds ... :-/
<daveshah> Lattice use the 4 version bits to distinguish between different part numbers for the same die (with/without SERDES and the fake 12k device)
<daveshah> sadly seems that Intel haven't made it that obvious
<daveshah> trabucayre: that's common sadaly
<daveshah> *sadly
<trabucayre> daveshah: yep
<daveshah> Lattice put comments in the SVF which helps a lot
<trabucayre> anlogic too
<Lofty> Caution: never invoke the following instruction codes
<Lofty> Hmm...
<trabucayre> Lofty: this mean : "it's the most important one"
<Lofty> If it comes up in the SVF, make a note, I think :P
<trabucayre> I need to check this yes
<kc8apf> I have an Arria II GX dev kit if someone wants to poke at one
<kc8apf> It's second hand so included 1-year quartus license is probably expired
<Lofty> Arria II is actually supported by Quartus Lite, so
<kc8apf> Only one specific device is supported in Lite. All other Arria II needs standard
<Lofty> Ah
QDX45 has joined #prjmistral
_whitelogger has joined #prjmistral