<
sorear>
are gecko feet mechanical?
Bicyclidine is now known as Bike
<
swetland>
oh that was a duuuuuumb bug
<
swetland>
program memory is multibanked and synchronous, so part of ins_rd_addr is decoded into chip select... which is great to initiate the read
<
swetland>
but on the following cycle when the right data needs to be routed into ins_rd_data, well.... not so good
pie__ has joined ##openfpga
pie_ has quit [Remote host closed the connection]
<
SolraBizna>
I can't wait until my core gets used with external memory and I have to worry about things like that D:
<
swetland>
oh this was just on-chip bram!
<
swetland>
for some reason I tossed in multiple banks up front just because
azonenberg_work has quit [Ping timeout: 250 seconds]
<
SolraBizna>
I avoided that so far by having the memory and IO devices clocked on the rising edge, and the logic clocked on the falling edge
<
SolraBizna>
like on the 65xx bus
Bob_Dole has quit [Ping timeout: 260 seconds]
Bob_Dole has joined ##openfpga
Flea86 has joined ##openfpga
luvpon has joined ##openfpga
<
_whitenotifier>
[whitequark/Glasgow] q3k bad3e59 - revC: length match LVDS pairs (#82)
pie__ has quit [Ping timeout: 268 seconds]
<
_whitenotifier>
[whitequark/Glasgow] marcan 5288648 - revC: swap U2/U3, fix DRC issue
<
_whitenotifier>
[whitequark/Glasgow] marcan 422462c - revC: pull +5V a bit further away from GND
<
_whitenotifier>
[whitequark/Glasgow] marcan aa64dad - revC: add ATECC508A
<
_whitenotifier>
[whitequark/Glasgow] marcan 5a3ec2e - revC: update wave soldering keepouts
unixb0y has quit [Ping timeout: 250 seconds]
<
_whitenotifier>
[whitequark/Glasgow] marcan 1473145 - revC: update wave soldering keepouts
<
_whitenotifier>
[Glasgow] whitequark commented on issue #68: Expose on-board I2C for accessories -
https://git.io/fpgAr
unixb0y has joined ##openfpga
<
_whitenotifier>
[Glasgow] whitequark commented on issue #68: Expose on-board I2C for accessories -
https://git.io/fpgxv
<
TD-Linux>
exposing the i2c is a bit obnoxious as it skips applets, and also putting the wrong thing on those pins screws up the whole of glasgow
<
whitequark>
that's another reason i don't do it
Bike has quit [Quit: Lost terminal]
rohitksingh_work has joined ##openfpga
azonenberg_work has joined ##openfpga
<
_whitenotifier>
[whitequark/Glasgow] marcan b50bd69 - revC: fix up diffpairs
genii has quit [Remote host closed the connection]
<
_whitenotifier>
[whitequark/Glasgow] marcan ec5f4da - revC: fix some more diffpair issues
Miyu has quit [Ping timeout: 246 seconds]
mmicko has quit [Quit: leaving]
bcoppens has quit [Remote host closed the connection]
wolfspraul has quit [Ping timeout: 252 seconds]
mmicko has joined ##openfpga
Bike has joined ##openfpga
catplant is now known as catplant[it]
catplant[it] has left ##openfpga ["WeeChat 2.2"]
jevinski_ has joined ##openfpga
jevinskie has quit [Ping timeout: 268 seconds]
jevinskie has joined ##openfpga
jevinski_ has quit [Ping timeout: 268 seconds]
GuzTech has joined ##openfpga
luvpon has quit [Ping timeout: 268 seconds]
futarisIRCcloud has quit [Quit: Connection closed for inactivity]
<
tnt>
this definitely turned out a bit more complex that I had originally imagined :/
<
daveshah>
ice40 PLLs are really horrible tbh
<
daveshah>
Thanks for your work on this!
<
tnt>
Heh np, happy to help :)
Miyu has joined ##openfpga
Miyu has quit [Ping timeout: 250 seconds]
rohitksingh_work has quit [Read error: Connection reset by peer]
pie__ has joined ##openfpga
Flea86 has quit [Quit: Goodbye and thanks for all the dirty sand ;-)]
rohitksingh has joined ##openfpga
<
tnt>
I swear everytime I submit I PR, I manage to screw up something somewhere in the process ...
<
tnt>
So many steps to extract one commit from my work tree and rebase and ... and I'm bound to fuck up one of them.
Miyu has joined ##openfpga
<
tnt>
daveshah: tx for the quick review ;)
rofl_ has joined ##openfpga
jcarpenter2 has quit [Read error: Connection reset by peer]
GuzTech has quit [Quit: Leaving]
pie_ has joined ##openfpga
<
pie_>
not that i really want to start reading some huge ESD standard
* pie_
googles around some more
Bike is now known as Bicyclidine
[X-Scale] has joined ##openfpga
X-Scale has quit [Ping timeout: 246 seconds]
[X-Scale] is now known as X-Scale
<
pie_>
im leafing around in our copy of the art of electronics and its really hard to decide where to start reading :D
<
qu1j0t3>
pie_: Don't forget Kuphaldt.
<
pie_>
qu1j0t3: hm thats an ida
<
qu1j0t3>
i read just about all of his books on my ereader, e.g. while taking the bus...
<
qu1j0t3>
yeah, that's true, some are incomplete. but the earlier, tutorial books are complete
<
qu1j0t3>
and he's a good writer.
rofl_ has quit [Ping timeout: 268 seconds]
jcarpenter2 has joined ##openfpga
Adluc has quit [Ping timeout: 252 seconds]
Adluc has joined ##openfpga
<
pie_>
azonenberg_work: see above
genii has joined ##openfpga
Bob_Dole has quit [Remote host closed the connection]
Bob_Dole has joined ##openfpga
rohitksingh has quit [Ping timeout: 246 seconds]
jevinskie has quit [Ping timeout: 268 seconds]
jevinskie has joined ##openfpga
Bob_Dole has quit [Remote host closed the connection]
Bob_Dole has joined ##openfpga
<
_whitenotifier>
[whitequark/Glasgow] whitequark e0c6f86 - cli: build: fix a confusing message.
<
_whitenotifier>
[whitequark/Glasgow] whitequark d4fedf0 - gateware.fx2: implement reset for async FIFOs.
<
_whitenotifier>
[whitequark/Glasgow] whitequark a91aa09 - arch.lpc: new definitions.
Bob_Dole has quit [Ping timeout: 268 seconds]
<
SolraBizna>
it should be possible to search for the first time a specific signal has a specific value in GTKWave, but I can't figure it out
azonenberg_work has quit [Ping timeout: 268 seconds]
azonenberg_work has joined ##openfpga
Zorix has quit [Ping timeout: 264 seconds]
<
whitequark>
daveshah: Info: Max frequency for clock 'sys_clk': 41.31 MHz (PASS at 30.00 MHz)
<
whitequark>
Info: Max frequency for clock 'lpcmonitorapplet_lclk_i_$glb_clk': 35.31 MHz (PASS at 33.33 MHz)
<
whitequark>
with the global promotion fix I get
*significantly* better timing
<
whitequark>
in almost all applets
<
daveshah>
Very nice
Zorix has joined ##openfpga
<
sorear>
41 MHz on the up5k?
<
sorear>
istr that was considered out of reach for a long time, nice
<
whitequark>
well, nextpnr is pretty good, and also my gateware is pretty good, too :P
<
daveshah>
Well it depends what you are doing
<
whitequark>
i used to barely clear 30 MHz iirc
<
daveshah>
From memory a simple UART controlling LEDs that is one of our icebreaker demos hits a bit over 50MHz
<
zkms>
my smol gmsk modulator works up to like 127MHz with nextpnr on ice40 8k
<
whitequark>
8k is much faster
<
whitequark>
daveshah: yeah but an fx2 arbiter is quite a bit more complex
<
daveshah>
Yeah, it must be well designed to manage that
<
daveshah>
hx8k is round about 3x faster than up5k
<
whitequark>
daveshah: i had to use registers in pads
<
whitequark>
registered tristates...
<
whitequark>
while meeting the weird fx2 timings
<
whitequark>
it was not trivial
<
daveshah>
I can imagine
<
TD-Linux>
any tips on converting a systemverilog design to yosys compatible verilog?
<
daveshah>
Improve the SV support in Yosys /s
<
daveshah>
I think in general the two things you'd need to remove are structs and enums
<
TD-Linux>
unironically one option (well also improve either verilator or icarus)
<
daveshah>
Yosys supports interfaces to some extent already
<
swetland>
verilator supports structs, enums, typedefs, interfaces, modports
<
TD-Linux>
oh cool. I'll run this though verilator first and then find the set that verilator supports and yosys doesn't
<
swetland>
I think enums, typedefs, and structs are the significant features missing from yosys
<
daveshah>
yes, that sounds about right
<
swetland>
there's at least two issues opened where folks have started hacking on support for them but not gotten to a submitable state
<
TD-Linux>
does always_* just get turned into always?
<
daveshah>
I'm not sure if this was fixed but for a long time logic was just converted to reg too
<
swetland>
both verilator and yosys recognize 'em but yeah I don't think they do any checking or enforcement
<
daveshah>
And Yosys doesn't really care for reg/wire interchange
<
whitequark>
daveshah: any chance you could add another global promotion option? for CE
<
whitequark>
same problem as logic
<
whitequark>
daveshah: also, do you think you could add a log message that would indicate the fanout of an inferred SB_GB
<
zkms>
is there a way you can ask nextpnr what the SB_GBs are assigned to?
<
whitequark>
it prints that in the log
<
whitequark>
but... it's usually $abc$screwyou
pie__ has quit [Remote host closed the connection]
pie___ has joined ##openfpga
<
whitequark>
daveshah: probably gonna do it myself, hm