rohitksingh has quit [Remote host closed the connection]
_whitelogger has joined ##openfpga
Laksen has joined ##openfpga
Patater has quit [Quit: Explodes into a thousand pieces]
Patater has joined ##openfpga
Patater has quit [Quit: Explodes into a thousand pieces]
Patater has joined ##openfpga
Miyu has joined ##openfpga
Patater has quit [Quit: Explodes into a thousand pieces]
Patater has joined ##openfpga
lexano has quit [Ping timeout: 264 seconds]
rohitksingh has joined ##openfpga
<tnt>
whitequark: just curious, can you run multiple applets at once on different pins ? (haven't really been able to run the code for real yet :p)
jevinskie has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
<sorear>
that's been discussed as a design requirement
<tnt>
Yeah, I did see that in the readme. But I'm not sure if it's implemented.
<tnt>
Also, a quick look at the code made me think it might be limited to 2 applets max ?
jevinskie has joined ##openfpga
rohitksingh has quit [Ping timeout: 240 seconds]
rohitksingh has joined ##openfpga
GenTooMan has joined ##openfpga
lovepon has joined ##openfpga
mumptai has joined ##openfpga
pie_ has joined ##openfpga
rohitksingh has quit [Ping timeout: 252 seconds]
<pie_>
apparently this is a thing https://deskthority.net/wiki/Holtite , which makes sense. but i dont know why they chose such an expensive model
<whitequark>
tnt: it is not implemented yet
<whitequark>
the 2 applet limitation is temporary
<whitequark>
it's mostly an UI issue
<pie_>
hm thats interesting, theyre solderless
<mithro>
whitequark: USB makes sense now!
<mithro>
whitequark: I was staring at the low level USB traces and looking at the J/K symbols and my mind kept seeing "ha jk" (as in joking) -- it's clear that USB was a practical joke that everyone mistook for a real protocol!
<whitequark>
lol
<whitequark>
yse
<mithro>
whitequark: Its right there in the low level protocol :-P
SpaceCoaster has quit [Ping timeout: 252 seconds]
SpaceCoaster has joined ##openfpga
<tnt>
daveshah: Is there some subtle connections between the fabric PLLOUT_B of a PLL and the HFOSC ?!?
<daveshah>
tnt: it's not something I've seen
<daveshah>
But I'm not sure I've tested that either
<tnt>
I have a PLL with both output connected (through fabric, not global). I also have global promotion disabled. Things works fine, I can blink 2 leds. I add a HFOSC (with global output and the required bits to make that work added) to blink a third led. And then ... the led linked to the PLL second output stops working.
<tnt>
arachneworks fine. But I don't see anything I'd consider relevant in the 2 bitstream difference.
<daveshah>
Very odd
<daveshah>
Can you check if this also affects different seeds in nextpnr?
<tnt>
daveshah: huh ... trying random seeds, I now have a bitstream where the PLL output A stopped working.
<tnt>
(but output B works)
<tnt>
Mmm, it might just be tht clock distribution through the fabric just incurs too much skew between different flip flops and the logic fails miserably.
<daveshah>
Yes that is quite possible
<daveshah>
Might be worth testing just by scoping the raw clock outputs rather than with the counter
<tnt>
Well, feeding the fabric output to a SB_GB always works AFAICT.
<tnt>
(scope is ... far away :p)
<tnt>
Anyway, at least it's not an arachne regression, so I can look at that specific issue later.
<tnt>
and continue cleaning up HF/LF global out support ...
<daveshah>
Sure sounds sensible
<daveshah>
Thanks for your work!
mumptai has quit [Quit: Verlassend]
<_whitenotifier>
[whitequark/Glasgow] whitequark pushed 1 commit to master [+4/-0/±1] https://git.io/fpcHC
<_whitenotifier>
[whitequark/Glasgow] whitequark a2df0aa - applet.jtag.arc: new applet.
<mithro>
tnt: you seem to be doing a lot of PLL stuff in nextpnr?
<tnt>
mithro: Yeah. Working on that now :p Trying to clean stuff up and adding global out support for HFOSC/LFOSC which is the latest piece missing AFAICT.
<mithro>
tnt: want to add support for calculating timing constraints on the output of PLLs based in the input timing constraints and settings? :-P
<tnt>
Oh wow, I haven't even looked at how all the timing stuff works ...
<_whitenotifier>
[whitequark/Glasgow] whitequark pushed 3 commits to master [+2/-0/±5] https://git.io/fpc7f