<openfpga-github>
[Glasgow] whitequark commented on issue #27: On second thought having two entities in the system control I/O buffers doesn't sound like such an awesome idea after all, so probably not. And the polarity of ENVA/ENVB is different, too. https://github.com/whitequark/Glasgow/issues/27#issuecomment-386209529
rohitksingh has joined ##openfpga
<rqou>
lolol clifford's twitter
<rqou>
clifford needs to hire another intern just to answer dumb questions :P
<azonenberg>
rqou: funny, i keep using the cheapest leds on digikey
<azonenberg>
and they're never high efficiency
<rqou>
still higher than the "20mA for a dim red glow" LEDs of the 80s
bitd has joined ##openfpga
rohitksingh has quit [Quit: Leaving.]
azonenberg_work has joined ##openfpga
rohitksingh_work has joined ##openfpga
scrts has quit [Ping timeout: 240 seconds]
scrts has joined ##openfpga
wpwrak has quit [Read error: Connection reset by peer]
wpwrak has joined ##openfpga
scrts has quit [Ping timeout: 240 seconds]
<whitequark>
azonenberg: wow usb is a steaming pile of shit
<whitequark>
just how can a protocol be so useless
<whitequark>
" The best way to identify HDA is probably just to look for the Class code (4h) and subclass (3h), but the HDA specification doesn't tell you to do this."
<whitequark>
off to a good start
<whitequark>
"Setting things up at the near end is relatively easy, part of it being done by keyhole surgery through a couple of ordinary ports in the PCI device configuration space, while the rest is done through the memory mapped ports located at an address found at index 10-17h (though the lowest four bits should be taken as zeros)."
<whitequark>
wtf
<q3k>
scroll down to 'setting up the afg codec'
<q3k>
tl;dr you need to traverse graphs in your driver in order to play some audio
<whitequark>
"It may be wise to keep count of the number of links in the path just in case a codec sends you round in circles"
<whitequark>
yeah this has intel written all over it
<whitequark>
"A VMware document notes that "software does not have a reliable way to determine its frequency. Generally, the only way to determine the local APIC timer’s frequency is to measure it using the PIT or CMOS timer, which yields only an approximate result."
<whitequark>
what
<q3k>
oh, duh
<q3k>
you need calibration loops :P
<q3k>
that's the least annoying part
<whitequark>
no, it's not annoying
<whitequark>
it's just ridiculous
<whitequark>
the system knows how fast the timer is running!
<whitequark>
at least put it in ACPI or something
<q3k>
well then you have to parse acpi
<q3k>
fuuuuuuuck acpi
<whitequark>
h
<q3k>
fuuuuuuuuuck the dsdt in particular
<whitequark>
o. this is called HPET
<whitequark>
"the actual frequency is provided to the operating system by a hardware register giving the number of femtoseconds per period"
<whitequark>
4-LUT, DFF, CYC... is this an ice40 clone?
<whitequark>
on a smaller node and with more luts
<kc8apf>
that's my impression
<daveshah>
I like the fact that "multiple clock" is a feature
<whitequark>
lol
<daveshah>
I wonder if there has ever been an FPGA architecure that literally allowed only one clock
<whitequark>
some small CPLDs?
rohitksingh has quit [Ping timeout: 240 seconds]
<kc8apf>
my contact at Thales says they've already been discussing opening up the tools. Sounds like we might get access to official configuration docs and timing models.
<cr1901_modern>
I could use a node in complexity between GP4 and HX1K
<whitequark>
that's like
<whitequark>
lp384
<cr1901_modern>
oh wait, I can't read ._.
<cr1901_modern>
Oh right, forgot about that
<q3k>
>NX-eFPGA blocks are programmed with our in house mapping software (NanoXmap) performing all required steps to transform a synthesizable RTL description into a bit stream downloadable in the NX-eFPGA through various hardware interfaces. The mapping process is built on advanced proprietary algorithms optimized for the architecture and capable of handling complex routing structures as well as LUT utilization
<q3k>
greater than 90%.
<q3k>
the marketing wank is strong in this one
<q3k>
tl;dr 'we rolled our own p&r and it's good enough'
<kc8apf>
per my contact: "their tools are not very good"
<azonenberg>
kc8apf: ooooh
<q3k>
what a twist
<azonenberg>
that would be a game changer
<awygle>
lmao "advanced proprietary algorithms"
<azonenberg>
larger vendor releasing official docs
<awygle>
how "larger" is this vendor exactly?
<daveshah>
yeah, it would be very exciting
* awygle
has never heard of them before
<whitequark>
isn't it even fabless?
<whitequark>
i mean
<whitequark>
not just fabless
<whitequark>
it seems to provide ip on demand primarily
<whitequark>
do they have any actual products?
<q3k>
yeah, they give you a macro to fab the thing yourself in your design
<daveshah>
apparently the IDE is "Linux, Python-based"
<daveshah>
Looks like their chip viewer might not be too bad
<kc8apf>
as in, ESA will push for them to release the complete design free-for-use
<q3k>
that would be super cool
<kc8apf>
first discussion with Thales today. He's going to see if he can get me early access to docs.
<whitequark>
looks reasonably different from ice40 actually
rohitksingh has joined ##openfpga
<kc8apf>
EDN article says they've fabbed a few design that are available in dev kits
<kc8apf>
"Both contain 35k LUTs, 2.8 Mb of embedded RAM, one 430 Mbps SpaceWire CODEC and embedded DSP, and are available today with ESCC and QML qualification expected in Q2 and Q3 of this year respectively."
<q3k>
mmm, spacewire
<kc8apf>
just what i need: an FPGA w/ SpaceWire ¯\_(ツ)_/¯
rohitksingh1 has joined ##openfpga
<whitequark>
wtf is spacewire?
rohitksingh has quit [Client Quit]
<kc8apf>
whitequark: you really don't want to know
<q3k>
whitequark: it's a wire for space
<kc8apf>
satellite people have unique ideas for data buses
<azonenberg>
i thoguht people usually just used arinc 429 for that?
<q3k>
I only know of this because I regularly have beers with one very jaded satellite engineer.
<Bike>
project happy, or project blood, depending on your etymological decisions
<awygle>
i love the word Sanguine
<qu1j0t3>
it's OKAY
<qu1j0t3>
I use it from time to time
<qu1j0t3>
and of course nobody knows what the heck i meant
<Bike>
use what
<Bike>
"sanguine"?
<Bike>
i thought you meant, like, wyoming
<whitequark>
"The Clam Lake facility, which served as the test site and was originally called the Wisconsin Test Facility (WTF)"
<whitequark>
appropriate
<qu1j0t3>
the word 'sanguine', which i do not think means "happy"
<awygle>
it's one of those things, like using the other meanings of "affect" and "effect", that is good amateur grammarian snipe bait
<qu1j0t3>
yes
<Bike>
"Anticipating the best; optimistic; not despondent; confident; full of hope."
<Bike>
close enough
<qu1j0t3>
obscure senses FTL/FTW depending on pov
<qu1j0t3>
Bike: "not despondent" would be the closest to how i would have defined it
<Bike>
"irresponsible mirth; indulgent in pleasure to the exclusion of important matters" oh that's more like it
<qu1j0t3>
which dictionary is this? :)
<Bike>
wiktionary, so if you want to impeach it go ahead
<qu1j0t3>
those are senses i wasn't familiar with
* qu1j0t3
checks the closest dictionary
<Bike>
it quotes shakespeare tho
<q3k>
Bike: sanguine's 'happy' etymology actually stems from 'blood' anyway
<qu1j0t3>
interesting.
<Bike>
i've never actually heard it in relation to blood, and i only realized there was a connection when i heard "consanguinity" in anthropology class
<qu1j0t3>
i didn't realise it was so POSITIVE
<Bike>
q3k: yeah, like "melancholy"
<awygle>
lol "ensanguinate"
<Bike>
or bilious, but that's kind of more obvious
<Bike>
imo, exsanguinate.
<q3k>
Bike: the blood etymology in it is fairly obvious if you know romance languages
* qu1j0t3
toasts everyone with sangria
<qu1j0t3>
q3k: Yes :)
<whitequark>
Bike: i was thinking of exsanguination
<Bike>
q3k: well, you see.... i don't
<Ultrasauce>
this is definitely a subject that requires a sense of humour
<qu1j0t3>
Bike: haha, bilious... great segue
<awygle>
yeah but the connection to the four humors thing is less obvious
<q3k>
Bike: fair enough, I'm a bit of a language nerd
<awygle>
it's basically a yandere word
<Bike>
and i'm almost positive i've never heard "choleric" to refer to an emotional state
<Bike>
or at all, really
<q3k>
Bike: 'choleric' is a thing in polish, interestingly (choleryk/cholerycznosc)
<Bike>
nice
<Bike>
meaning what, angry?
<q3k>
so is 'sangwinik', TIL
<q3k>
choleryk meaning agitated, vivid, ...
<Bike>
and it's in common use?
<q3k>
ambitious
<q3k>
well, not ambitious
<q3k>
can't really translate it well
<q3k>
yeah, it is
<Bike>
i love that, when some word is obscure in some language but a cognate is common
<q3k>
I mean, you might not use it, but you'll understand it / have heard it before
<Bike>
unlike "choleric"
<q3k>
yeah
<Bike>
kind of fell out of fashion along with, like, the british empire
<q3k>
there's also this other more common meaning of 'cholera' which means being pissed off / angry / about done with this fucking thing
<q3k>
all of this shit is basically untrasnslatable literally
<q3k>
which indicates that it's more phraseological than strictly factual
rohitksingh1 has quit [Quit: Leaving.]
<whitequark>
q3k: ha, cholera means that in polish as well as in russian then
<Bike>
oh, damn, "chole" is greekish for bile, so it's the same chole as in melancholy, and when you're the other kind of choleric you're expelling chole
<q3k>
i'm not surprised :P
<Bike>
learning on a wish
<qu1j0t3>
Bike | i love that, when some word is obscure in some language but a cognate is common // This happens VERY often between say English and Portuguese.
<Bike>
yeah i've noticed! but i can't remember any examples since i don't actually know portuguese
<azonenberg>
whitequark: so what current did you end up running your searchlight^hindicator LEDs at?
<whitequark>
1 mA for green ones
<whitequark>
there are three of them which makes things significantly worse
<azonenberg>
no i mean, did you rework them>
<azonenberg>
?
<whitequark>
no
<whitequark>
too lazy :p
<gruetzkopf>
spacewire!
<gruetzkopf>
i've always wanted to redo my half-assed implementation and implement RMAP
<gruetzkopf>
the ECSS-E-ST-50 series of standards is fairly interesting
<awygle>
CCSDS/ECSS standards are pretty good generally
ZipCPU has quit [Ping timeout: 264 seconds]
ZipCPU has joined ##openfpga
<whitequark>
hm
<whitequark>
what if I have a toplevel I/O port that has mixed function bits in it in verilog
<whitequark>
if I say input or output they all become that
<whitequark>
if I say inout yosys wants me to connect it to SB_IO
<cr1901_modern>
I would connect it to an SB_IO during finalization, but I could've sworn yosys is just fine w/ inout by itself?
<cr1901_modern>
oh wait you said Verilog, not migen
<whitequark>
i'm fixing a migen bug
<whitequark>
and no
<whitequark>
top.blif:138: fatal error: toplevel inout port 'io[0]' not connected to SB_IO PACKAGE_PIN
<whitequark>
oh hm
<whitequark>
blif
<azonenberg>
This might be an icestorm issue?
<whitequark>
yes
<azonenberg>
if you have an inout pin
<azonenberg>
you should be able to assign it to a output-only signal
<azonenberg>
and just ignore thei nput
<whitequark>
yes, it's an arachne-pnr issue
<whitequark>
wtf arachne-pnr uses the gnu brace indentation style
<whitequark>
at least it doesn't use gnu indentation, or i would have ragequit