clifford changed the topic of #yosys to: Yosys Open SYnthesis Suite: http://www.clifford.at/yosys/ -- Channel Logs: https://irclog.whitequark.org/yosys
tpb has quit [Remote host closed the connection]
tpb has joined #yosys
citypw has joined #yosys
cr1901_modern has quit [Quit: Leaving.]
Degi has quit [Ping timeout: 258 seconds]
Degi has joined #yosys
cr1901_modern has joined #yosys
m4ssi has joined #yosys
_whitelogger has joined #yosys
m4ssi has quit [Remote host closed the connection]
emeb_mac has quit [Quit: Leaving.]
craigo has quit [Ping timeout: 240 seconds]
m4ssi has joined #yosys
m4ssi has quit [Remote host closed the connection]
jakobwenzel1 has joined #yosys
xtro has quit [Ping timeout: 258 seconds]
Asu has joined #yosys
proteusguy has quit [Ping timeout: 256 seconds]
proteusguy has joined #yosys
proteusguy has quit [Ping timeout: 256 seconds]
kraiskil has joined #yosys
kraiskil has quit [Ping timeout: 272 seconds]
<pepijndevos> Yosys doesn't synthesize floats, or does it? I know it's bad, but for a thing where I don't really care about speed or area, it beats fiddling with fixed point.
<whitequark> pepijndevos: doesn't
<pepijndevos> ok
<whitequark> also i feel like if yosys dropped some combinatorial ieee754 ops into your design you'd suddenly start caring about speed and area
<pepijndevos> haha maybe
<whitequark> right now floats are secretly strings
<whitequark> which... is a pretty awful representation if you ask me
<pepijndevos> w-what?
<whitequark> floats in yosys are just strings in the decimal notation
<whitequark> like a float RTLIL::Const is a byte vector containing ASCII containing the float in decimal
<pepijndevos> So... what happens if you synth a design with floats?
<whitequark> no idea
<whitequark> i'm sure it breaks but i don't know how exactly
<pepijndevos> lol ok
<whitequark> 1364.1 is clear about floats not being synthesizable tho
<pepijndevos> 100% sure quartus does it though. Back when I assisted FPGA labs in uni, I've seen some unfortunate souls try.
<whitequark> huh
<whitequark> i wonder if it's quartus itself or just synplify
<pepijndevos> Combinatorial float FFT did not fit on the FPGA though
<pepijndevos> Maybe I'll mess around and see if the Gowin or Latice tools can do it. Gowin usus Synplify IIRC
<pepijndevos> Gowin using Symplify pro: Synthesis of real vin_val is not supported
jakobwenzel1 has quit [Quit: jakobwenzel1]
jakobwenzel1 has joined #yosys
<pepijndevos> Seems quartus support is also rather limited actually
craigo has joined #yosys
<daveshah> It is probably mostly for the Stratixes with hard float DSPs these days
<Lofty> Actually the Cyclone 10GX supports some measure of floating point arithmetic
<Lofty> But the hardware *can* do it
<daveshah> Oh interesting I thought it was only the Stratix 10
<daveshah> And their newer agilex or whatever it's called
<Lofty> It's probably further enhanced on those architectures, but it's still got support on C10GX
<daveshah> Mmm
<daveshah> Xilinx are doing it too with Versal
<Lofty> The S10 (and presumably Agilex) both changed the ALM a bit, so I'm gonna have to investigate that. If anybody decides to care/pay me
u0m3 has quit [Write error: Broken pipe]
az0re has quit [Remote host closed the connection]
citypw has quit [Remote host closed the connection]
kraiskil has joined #yosys
u0m3 has joined #yosys
citypw has joined #yosys
kraiskil has quit [Read error: Connection reset by peer]
jakobwenzel1 has quit [Quit: jakobwenzel1]
jakobwenzel1 has joined #yosys
lambda has quit [Ping timeout: 256 seconds]
lambda has joined #yosys
emeb has joined #yosys
craigo has quit [Ping timeout: 258 seconds]
X-Scale has quit [*.net *.split]
qu1j0t3 has quit [*.net *.split]
tlwoerner has quit [*.net *.split]
ZipCPU has quit [*.net *.split]
somlo has quit [*.net *.split]
oldtopman has quit [*.net *.split]
tmiw has quit [*.net *.split]
tmbinc has quit [*.net *.split]
acertain has quit [*.net *.split]
qu1j0t3 has joined #yosys
ZipCPU has joined #yosys
acertain has joined #yosys
tmbinc has joined #yosys
somlo has joined #yosys
tlwoerner has joined #yosys
oldtopman has joined #yosys
tmiw has joined #yosys
X-Scale has joined #yosys
maartenBE has quit [Ping timeout: 246 seconds]
X-Scale has quit [Ping timeout: 272 seconds]
X-Scale has joined #yosys
maartenBE has joined #yosys
sjkelly1 has joined #yosys
evil_chuck has joined #yosys
kraiskil has joined #yosys
ross_s has quit [Ping timeout: 256 seconds]
az0re has joined #yosys
citypw has quit [Ping timeout: 240 seconds]
xtro has joined #yosys
ross_s has joined #yosys
jakobwenzel1 has quit [Quit: jakobwenzel1]
m4ssi has joined #yosys
kraiskil has quit [Read error: Connection reset by peer]
kraiskil has joined #yosys
m4ssi has quit [Remote host closed the connection]
Asu has quit [Read error: Connection reset by peer]
Asu has joined #yosys
emeb_mac has joined #yosys
evil_chuck has quit [Remote host closed the connection]
evil_chuck has joined #yosys
kraiskil has quit [Read error: Connection reset by peer]
Asuu has joined #yosys
Asu has quit [Ping timeout: 240 seconds]
evil_chuck has quit [Quit: Leaving.]
thardin has joined #yosys
<thardin> there we go, thought I forgot a channel
<thardin> got my little lock-in amplifier working. with filters tunable by potentiometer
<thardin> realizing I can do multiplication as bits arrive from the ADC was a life-saver
<emeb> Oooo - lockin amps are fun. Yours is an all digital version?
<thardin> except for the ADCs and DACs yes
<thardin> it's very bare bones at the moment
<thardin> an octopus of pmods coming off the icestick, with a hand-soldered board for the pots for good measure
<emeb> Is it dual-phase?
<thardin> not yet
<thardin> might need a larger device for that
<thardin> in fact I haven't added adjustable phase to the mixer yet
<emeb> Yeah - the HX1k on icestick is a bit small.
<thardin> I was looking at HX8K board a few weeks ago
<emeb> The DSP cores in UP5k might be useful for that.
<thardin> I'm only doing addition and look-ups at the moment
<thardin> but if I were to bump up the speed into the tens of MHz then being able to multiply faster would be useful
<emeb> Interesting. How do you do the mixing operation w/o multiplication?
<thardin> it does serial multiplication
<emeb> aha
<thardin> as in, it remembers the last LO sample and for each non-zero ADC bit it adds it shifted to an accumulator
<emeb> yep. longhand
<emeb> so your sample rate is << clock rate
<thardin> yeah it's 1/32
<thardin> I only need 100 kHz bandwidth for now
<emeb> how do you generate sinusoids? LUT?
<thardin> yeah
<thardin> plus symmetry
<emeb> what's the phase resolution?
<thardin> so it stores 0-90°
<thardin> uh.. the table is 4096 entries so
<thardin> let's see
<emeb> 2pi/16k
<thardin> 0.02°
<thardin> 0.4 mrad
<emeb> not bad
<thardin> should be more than enough
<emeb> Yeah.
<thardin> plan is to squeeze the design into a eurorack module
<emeb> heh
<thardin> they're not just for musicians!
* emeb does DSP for eurorack
<thardin> cool
<emeb> hard to imagine doing "real science" with that stuff. Presumably you just use the rack, supplies and physical dimensions for some custom equipment?
<thardin> yeah there's modules aimed at academia
<emeb> Interesting. Got a link so such stuff?
<emeb> link to?
<thardin> nope, but if you search around for laser stuff you'll probably find some
<thardin> they come with a hefty price tag
<emeb> I'm sure
<emeb> I know some of our customers use the synth modules for controlling galvos in laser shows, but that's about as close as I come to laser stuff.
<thardin> pew pew
<thardin> a laser is just a radio running at much higher frequency
<emeb> true dat
<thardin> you have non-linear optics that behave much like RF mixers
<emeb> indeed - hence green laser pointers
<thardin> so you can beam two lasers into it and get the difference in frequency out as RF
<emeb> how does one get that difference freq into a circuit though? Stick an antenna on the crystal?
<thardin> they probably have some kind of antenna there yes
<thardin> a transition of some sort
<thardin> you can go the other way with pockels cells and similar devices
<thardin> modulate a laser with RF
<thardin> you also have piezo devices that do aucusto-optical modulation
<emeb> yep - that's what I've seen before.
craigo has joined #yosys
<emeb> and of course simple AM of laser emitters by modulating supply voltages.
<thardin> yes
<emeb> worked on a microwave RF delay based on that long ago. 2GHz -> IR laser diode -> long coil of fiber optic -> detector.
<thardin> nice
<emeb> Hard to get a good RF match on a commercial IR laser diode.
kraiskil has joined #yosys
kraiskil has quit [Read error: Connection reset by peer]
kraiskil has joined #yosys
Asuu has quit [Quit: Konversation terminated!]
kraiskil has quit [Read error: Connection reset by peer]
emeb_mac has quit [*.net *.split]
ross_s has quit [*.net *.split]
tlwoerner has quit [*.net *.split]
qu1j0t3 has quit [*.net *.split]
ZipCPU has quit [*.net *.split]
somlo has quit [*.net *.split]
oldtopman has quit [*.net *.split]
tmiw has quit [*.net *.split]
tmbinc has quit [*.net *.split]
acertain has quit [*.net *.split]
oldtopman has joined #yosys
emeb_mac has joined #yosys
acertain has joined #yosys
qu1j0t3 has joined #yosys
somlo has joined #yosys
ZipCPU has joined #yosys
tlwoerner has joined #yosys
ross_s has joined #yosys
tmbinc has joined #yosys
tmiw has joined #yosys