soylentyellow__ has quit [Ping timeout: 265 seconds]
<awygle>
rqou: try to avoid sharing vias if you can
<awygle>
(not the end of the world if you can't though)
<rqou>
seems a bit tricky here
<awygle>
Otherwise looks fine to me (without looking at the schematic or the pinout)
<awygle>
I was mostly looking at the top left, seems room for a second ground via
<rqou>
ok, i'll think about it
<awygle>
beyond decoupling, a via for each gnd ball is preferred too
<awygle>
But ultimately nbd here I suspext
* awygle
zzz
_whitelogger has joined ##openfpga
<pie__>
hm
<pie__>
<feep> > BTW, if you are operating a data center, NVIDIA has recently updated their EULA to prohibit “desktop” grade GeForce and Titan GPUs from being used in data centers.
<pie__>
<feep> they're finally at a point where they can somewhat reliably clock cards to full
<pie__>
<feep> shouldn't google put a team on nouveau?
<pie__>
<feep> so
<pie__>
<feep> the eula only covers the driver
<pie__>
<feep> I wonder to what extent the use of driver lockout to make devices act as software for licensing purposes is illegal
<rqou>
i don't get why reclocking has been so hard to figure out
<pie__>
though now its got me wondering about gpu computation error rate
<azonenberg>
quadros are designed for rendering, teslas are pure compute
<azonenberg>
most of them dont even have video out connectors
<rqou>
o/ azonenberg
<rqou>
why are you so intermittently responsive?
<pie__>
> family man doing family things
<azonenberg>
rqou: busy doing construction, then $sidegig
<rqou>
how done is construction?
<azonenberg>
Framed out the access door for the attic in the lab today
<azonenberg>
Framed the wall under the stairs
<azonenberg>
Insulated that wall
<azonenberg>
Framed out the wall around the water main in the lab
<azonenberg>
Insulated that wall
<azonenberg>
Insulated all but about the last 4 feet of the lab ceiling
<azonenberg>
and a few other odds and ends
<rqou>
how come your "misc list" seems to be unending
<azonenberg>
Because new things keep being added to it :p
<azonenberg>
Tomorrow's goal is to finish insulating the last bit of the lab, then insulate the ceiling over the kitchen, dining room, and bathroom upstairs
<azonenberg>
Stretch goal is to do some of the bedrooms too
<azonenberg>
that may have to wait until monday over lunchtime / after work
<azonenberg>
rqou: but most of these items had been on the todo for a long time
<azonenberg>
The wall around the water main actually got rebuilt today in a slimmer fashion, my original design stuck out too much and the electrical inspector didnt like it
<azonenberg>
so now i'm tighter to the concrete
<azonenberg>
all of the other stuff had been pending for weeks and i just now got around to it
<azonenberg>
Sheetrock arrives monday so my focus right now is doing easy things that use up bulk quantities of materials, like insulating the ceiling
<azonenberg>
to free up space on the floors for the sheetrock
<rqou>
when i was helping you a whole month ago, i was under the impression that you would be imminently done
<rqou>
due to getting kicked out
<cr1901_modern>
>azonenberg: Because new things keep being added to it :p
<cr1901_modern>
Story of my life. This is why I'm refusing to add any more entries until a few old ones are done. It's... actually been quite effective
<azonenberg>
i'm working as fast as i can
<azonenberg>
And no faster
<rqou>
and in reality you weren't even _close_?!
<azonenberg>
Nope
<azonenberg>
we had 2+ weeks of electrical
<azonenberg>
after that
<azonenberg>
then insulation came a week or two ago, at this point we've done the majority of the walls and somewhere between a third and a half of the ceilings
<azonenberg>
Ceilings go a lot quicker than walls, for the most part, because there's less obstacles in the way
<azonenberg>
A lot of the wall cavities are partial height... the second floor is a few inches shorter than 8 feet because that was the style in the 70s
<azonenberg>
the first floor has a lot of split areas around the foundation where you have a top and bottom piece
<azonenberg>
so almost every piece had to be cut to length
<azonenberg>
a lot of studs are doubled or spaced around windows so you have to adjust width too
<azonenberg>
then you need cutouts for electrical boxes etc
<azonenberg>
meanwhile the ceiling cavities, for the most part, are exactly 2 feet wide everywhere and have relatively few electrical boxes to dodge
<rqou>
at the pace you're going i see no way you'll be done by the end of this year, let alone by your mandatory deadline
<azonenberg>
Then the second layer of insulation in the attic is just laid loosely on top
<azonenberg>
oh and we've been moving a carload of boxes every day we go over to do work
<azonenberg>
and just staging them in random corners of rooms that are insulated already
<azonenberg>
so we have less to move at EOM
<azonenberg>
And well, at this point there isnt a whole ton left
<azonenberg>
I have to frame dummy walls in two closets downstairs to hold insulation (maybe 2-foot long walls, should be like 3 studs each)
<azonenberg>
put one or two extra ceiling joists in a closet around a HVAC duct
<azonenberg>
put a few 2x4s around the HVAC in the office
<azonenberg>
Install a handful of insulation bits downstairs, but the first floor is almost done
<azonenberg>
Hang a bunch of insulation in the walls upstairs, almost all of these pieces are precut by ally while i was at work and just have to be hung
<rqou>
this list is absolutely no shorter than the "just" list from a month ago
<azonenberg>
Insulate the remainder of the second floor ceiling
<azonenberg>
Then hang sheetrock and do finish electrical
<azonenberg>
and well, as i get closer to various items they get split
<azonenberg>
"hang insulation" was one item a month ago
<azonenberg>
now it's like 30 tiny items
<azonenberg>
Because when something is a long way out it's hard to know all the steps
<azonenberg>
As soon as sheetrock arrives, i can start working on sheetrocking the interior (non-insulated) wall and ceiling areas downstairs
<azonenberg>
in parallel with the insulation
<azonenberg>
Since the inspector only needs to see the insulation in the walls that, well, have insulation in them
<azonenberg>
nothing is stopping me from sheetrocking the rest
<azonenberg>
i might try and pull some sar folks for a work party, thats a community i havent begged for help yet :p
<pie__>
amish style
sunxi_fan has joined ##openfpga
<azonenberg>
rqou: in any case, progress is being made and that's all i can say
<azonenberg>
I'm confident i will have insulation done in a couple more days, sheetrocking will most likely not be done on time but i can always hope
<azonenberg>
If we end up couch surfing for a bit, or sleeping on an air mattress on the deck or something, we'll deal with it
<rqou>
i'm going to predict that you won't be done even by december
<azonenberg>
Define "done"
<azonenberg>
insulation in, sheetrock up, power turned on?
<azonenberg>
there's a nontrivial chance that we'll slip the deadline by a week or two but certainly end of august i have no doubt
<azonenberg>
But that's when all of the other work starts
<azonenberg>
new roof (not much to do from our perspective there except sign the check), sanding and painting the walls, trim, flooring
<azonenberg>
ESD floor in the lab
<azonenberg>
buy and install the UPS
<azonenberg>
pull data lines through all the conduits
<azonenberg>
Measure the conduits and pull fiber
<azonenberg>
buying and installing cabinets and shelving
<azonenberg>
some new workbenches probably
<azonenberg>
hiring folks to put exhaust fans and HVAC in the lab
sunxi_fan has quit [Ping timeout: 260 seconds]
<azonenberg>
installing the racks properly, bolted to the floor etc
<rqou>
fine, i'm defining "done" as "able and willing to contribute to openfpga again"
<azonenberg>
Oh
<rqou>
i'm going to estimate this at january 2019 at the rate you're going
<azonenberg>
you do realize we are going to be remodeling the kitchen and both bathrooms too right?
<azonenberg>
this whole renovation we've been doing is only the initial stages
<azonenberg>
literally just enough to get the place livable
<azonenberg>
we have several years more work to do
<azonenberg>
that being said, as soon as we're moved in and i have a minimum viable lab put together in whatever corner of one room is somewhat clean
<azonenberg>
the pace of work will slow down drastically since we wont be under a time crunch
<rqou>
wtf
<azonenberg>
Which means the work will be taking much longer, but it won't be dominating my schedule like it does now
<rqou>
fine, my new revised estimate for done is now january 2025
<azonenberg>
i'll be spending a few hours here and there on it while doing other stuff too
<azonenberg>
Also, we were and continue to be budget limited in addition to time limited
<rqou>
your renovation has already been taking ludicrously long
<azonenberg>
So for example, when the new roof is going up
<azonenberg>
and during the weeks prior
<azonenberg>
i expect to be doing almost no other construction as we wont be able to afford any other materials
<azonenberg>
Ludicrously long? look at how many man-weeks it takes to build a house normally
<azonenberg>
now imagine about 1.2 people doing all that work
<azonenberg>
(i dont count ally as a full person because of her relative lack of experience, she's not nearly as fast as an experienced construction laborer)
<azonenberg>
i'm surprised its getting done as fast as it is
<cr1901_modern>
That's still mean lol
<azonenberg>
But yeah, from my perspective the immediate goal is to get the place to the point that it's legally occupiable
<azonenberg>
That means all insulation up and inspected, sheetrock hung with one coat of mud over the tape, and all electrical fixtures installed and inspected
<azonenberg>
goal is end of month, failing that as quickly as possible
<pie__>
#occupyazonenberg
<azonenberg>
After that i'm probably going to take a weekend to do... anything but construction
<azonenberg>
lol
<azonenberg>
after that, start working on all of the finish stuff
<azonenberg>
Which as i mentioned i full expect to run into next year
<azonenberg>
My lab will probably be largely offline for six or more months
<azonenberg>
this was the case with my last move too, and one of the reasons i hate moving
<azonenberg>
unpacking bit by bit, upgrading gear as i go
<azonenberg>
as part of the lab revamp for example i'm getting a curie point soldering iron
<azonenberg>
But not until i can afford one
<azonenberg>
i may not bother to set up the old aoyue POS
<rqou>
i'm still sticking to 2025 as my estimate for when you'll actually start working on openfpga again
<rqou>
given just how slow everything's been going
<rqou>
you should seriously just hire some illegal immigrants or whatever to help you
<azonenberg>
i expect to be spending a lot of time post-move on $sidegig stuff trying to bring in cash to fund both the renovations in general
<azonenberg>
and new lab equipment
<azonenberg>
before i settle in and start actually using said lab
<azonenberg>
i have probably $30-50K of upgrades i want to make to the lab alone once the construction proper is done
<azonenberg>
$8K of UPS, $5K or so of HVAC, $1K of soldering iron, a few $k of benches and shelving
<azonenberg>
$13K of 1 GHz DSO
<azonenberg>
this is my first time actually outfitting a lab for the long haul, rather than "hack something together to use until i move again"
<azonenberg>
So i'm doing it right
<azonenberg>
But i'm also not going to rush it just so i can work on some project right now
<rqou>
hey azonenberg, have you ever considered the fact that, despite not paying your bills, there are many people who would really love to see your open-source contributions?
bitd has joined ##openfpga
<azonenberg>
i'd love to have 40 hours a day to work on things
<azonenberg>
but i dont
<azonenberg>
and i have to prioritize
<azonenberg>
i'm in this for the long haul, investing in facilities that i'll be using for decades
<azonenberg>
Once i get back up and running, LATENTRED is the next priority so i can get a 10G LAN
<azonenberg>
STARSHIPRAIDER is now a priority at work, so i'll be doing that by day as time permits
<azonenberg>
they're funding me to work on jtaghal now too
<azonenberg>
i have scopehal/scopeclient building under cmake and will be adding a bunch of decoders for my $dayjob research
sunxi_fan has joined ##openfpga
<azonenberg>
So its not like i'm not doing open source work at all
<azonenberg>
But i have to focus on things that pay the bills
<azonenberg>
jtaghal is going to be getting a lot more arm coresight support over the coming weeks
<azonenberg>
and probably some stm32 stuff
<azonenberg>
scopehal is going to be getting a jtag protocol decoder to test all of this
<azonenberg>
(and to help with REing proprietary jtag tools)
<azonenberg>
among other things, as per twitter, i want to RE the chipscope jtag protocol
<azonenberg>
and make a scopehal backend to talk to chipscope cores
<azonenberg>
in addition to redtin
<rqou>
hmm, none of these are projects that i'm particularly interested in
<rqou>
i'd be much more interested in for example homecmos
<azonenberg>
yes, thats a long term project which requires substantial physical infrastructure
<azonenberg>
Like the $3k+ fume hood i dont yet have
<azonenberg>
or the glove box i might not even have a place to put
<azonenberg>
You see why i'm making infrastructure my priority right now?
<azonenberg>
i have to do the whole house to start so it's livable, but once we're moved in focus will shift towards outfitting the lab properly
<rqou>
or you can do it hackily
<rqou>
idk, i always find the specific way you prioritize things to be really weird
<gruetzkopf>
the upper level of the lower part of the building contains contains an object-oriented relay-based railway interlocking system, consisting of ~7500ish relays
<sorear>
How large of a geographical area is it responsible for?
<pie__>
i think im misunderstanding something here xD
<daveshah>
Chatting a few years ago with a friend who worked on railway systems, new signalling products were being created running something like Visual Basic 3 (way before .NET anyway) on Windows XP...
<daveshah>
It is probably one of the most conservative industries
<gruetzkopf>
nope, you're not
<gruetzkopf>
with this tech you can do 6.5 km in all directions
<sorear>
How many signals are there in that radius?
<gruetzkopf>
pretty much as many as you want. minimalistic config would be around 12, including distant signals, in our case its 50-ish, if you look at cologne central station it's a lot more
bitd has quit [Ping timeout: 245 seconds]
<gruetzkopf>
our setup is for development purposes, we're not actually controlling the railway line once controller from that building
<sorear>
That’s a big and expensive building for 12 signals?
<gruetzkopf>
tiny setups would have smaller buildings, of course
<gruetzkopf>
our setup is 6 rows of 5 frames of relays, ~50ish signals, 2 level crossings, 38 points, 4 lines on either end
<gruetzkopf>
the minimal case would be 4 frames for a simple crossing on a single track
<mithro>
I feel like the D/Q don't line up because of the q bit below the text line
<daveshah>
Think it will be OK once pins are added
<daveshah>
Personally I'd remove the pulse symbol from the latch, but I'm OK with it too
<mithro>
daveshah: I know without the pulse symbol it will confuse future Tim because he will mistake it for a FF
<daveshah>
mithro: sure
<mithro>
daveshah: I'm pondering if I should just put the word "latch" on it
<daveshah>
mithro: yeah, they're not really that common anyway
<daveshah>
I know 7 series has them, ice40 and ecp5 don't
<Adluc>
Hello guys, although obscure question, but does anyone have any beckhoff/ethercat stuff IP cores?
<mithro>
The ice40 has them in one place - the "iCE gate latch" which is how I ended up down this rat hole
<mithro>
daveshah: any idea why latches are so uncommon?
<daveshah>
mithro: there are many reasons why clocked logic is a better solution
<daveshah>
More predictable timing and no problem with glitches
<daveshah>
FPGA logic synthesis is not glitch free, which causes problems with anything asynchronous
<mithro>
daveshah: Also, Clifford mentioned that we should probably just be comparing the output of a simulator before and after pnr rather than trying to logic equivalence checking for things like blinky
<daveshah>
mithro: yeah, I think I mentioned that a while back
<daveshah>
For slow stuff like blinky simulation is definitely the correct approach
<daveshah>
For stuff which only needs a few clocks to do stuff, verification should also be good
<mithro>
daveshah: except that requires having an actual testbench, or do you think we can get away with the sim command in yosys?
<daveshah>
mithro: or use Verilator and dash off a quick C++ testbench
pie__ has quit [Ping timeout: 260 seconds]
Miyu has joined ##openfpga
<mithro>
daveshah: Run time isn't the issue - it's me writing verilog :-)
<daveshah>
mithro: exactly, that's why I suggested writing C++ instead
_whitelogger has joined ##openfpga
rohitksingh has quit [Quit: Leaving.]
pie_ has joined ##openfpga
<mithro>
How does one control the iverilog output vcd filename?
<daveshah>
I think that would normally be in your dumpvars statement
bitd has joined ##openfpga
<kc8apf>
Adluc: I looked a few months ago and couldn't find any. The specs are only available to businesses in the automation industry or for purchase through IEC.
<kc8apf>
Physical and link layer seems straightforward. I have no idea what the routing layer looks like other than slots in a process data packet.
user10032 has quit [Quit: Leaving]
uovo has joined ##openfpga
oeuf has quit [Ping timeout: 264 seconds]
<mithro>
daveshah: I'd like to pass it in on the command line... I thought I could do this with a -DVCDFILE="blah" but I'm getting "Unable to bind wire/reg/memory `VCDFILE' in `test'"
<mithro>
But my -D isn't being passed through to the preprocessor from what I can see
<mithro>
I'm obviously doing something wrong but I can't figure out what :-/
<reportingsjr>
esden: ping
<daveshah>
mithro: missing a backtick before VCDFILE in the Verilog?
<daveshah>
Classic mistake coming from C/C++ to Verilog