<azonenberg>
rqou: me? my lab is actually surprisingly free of carcinogens
<azonenberg>
compared to others i've seen
<rqou>
lol
<azonenberg>
i use rohs solder, have decent ventilation
<rqou>
but your presence itself can cause cancer :P :P
<azonenberg>
no that would be dig and his monster xray head
<azonenberg>
:p
<azonenberg>
Also, that was in part a reference to an incident that occurred circa 2010
<azonenberg>
when me and dig lived together
<azonenberg>
we did family-style cooking for a while so somebody was making noms for all of us
<azonenberg>
(5 people in the apartment)
<azonenberg>
i forget who it was, but he was taking forever
<azonenberg>
So one of us suggested maybe we should kill and eat the chef instead
<azonenberg>
this of course led to a discussion about which of the five of us would make the best meal
<azonenberg>
We had me - skinny, not much meat, and in decent shape so hard to catch
<azonenberg>
There were two guys high on the list - lots of meat and not very athletic
<azonenberg>
one other guy was kinda in the middle
<azonenberg>
then dig was firmly in last place
<azonenberg>
Plenty of stuff to eat, not a great runner, but if you actually tried to eat him you'd get ten different kinds of cancer
<azonenberg>
from all the stuff he had been exposed to :p
<pie__>
haaahahaha
<lain>
XD
<whitequark>
lol
* whitequark
writes down: "digshadow is a chemical pufferfish"
<digshadow>
azonenberg: haha I now know my spirit animal
<azonenberg>
lool
<digshadow>
the zombies will bite me and then die off
<digshadow>
i'll cure the apocalypse
<azonenberg>
lol
<whitequark>
what if they absorb the toxins instead
<whitequark>
now you've got poisonous zombies.
<azonenberg>
grrreat
<azonenberg>
but they'll glow in the dark
<azonenberg>
easier to aim at :D
* rqou
flips a table at abc
<rqou>
i can't figure out how to hack in "xor on top of SOP"
<digshadow>
rqou: use glue
FabM_Cave has joined ##openfpga
FabM_Cave is now known as FabM
<pie__>
* sniff glue
* pie__
sublimates
<pie__>
^ dont breathe this
<azonenberg>
pie__: smoke, don't breathe this
<azonenberg>
speaking of which i just got epoxy all over my artiq shirt :(
<azonenberg>
picked up a bottle that had gotten all cloudy and solidified, planning to throw it out
<azonenberg>
the whole outside was covered in goop
<azonenberg>
Gonna have a fun time deconning the shirt tomorrow, didnt have time to do it tonight
<azonenberg>
it's uncured resin so should clean off
user10032 has joined ##openfpga
pie_ has joined ##openfpga
pie__ has quit [Read error: Connection reset by peer]
azonenberg_work has joined ##openfpga
<azonenberg>
awygle: btw, re ethernet switch
<azonenberg>
digikey has the ksz9031rnx (RGMII) for $2.48 each @ qtyt 25
<azonenberg>
and the DP83867 (SGMII) for $5.84 ea @ qty 25
<azonenberg>
Which comes out to $62.00 / $148.50 for 25 PHYs (24 ports + one extra to hit price break)
<azonenberg>
Original plan: RGMII ports only (way cheaper) but this fails if i want 24 + 4 b/c i dont have enough pins on the fpga
<azonenberg>
Second plan: SGMII ports only, costs over twice as much but this may not be worth optimizing given the cost of the remainder of the system
<azonenberg>
Most PHYs i've seen recommend keeping RGMII lanes quite short, 6" max / 2" ideal, and i dont think that's possible within the geometry of a 24-port switch with three line cards plus a main board
<azonenberg>
SGMII seems to be more forgiving with layout since it's self-timed differential, and i can even put lvds buffers on the backplane for the last line card if i really have to
ondrej2 has joined ##openfpga
bitd has joined ##openfpga
<gruetzkopf>
hmm, four-pair SGMII
eduardo__ has joined ##openfpga
eduardo_ has quit [Ping timeout: 246 seconds]
noobineer has joined ##openfpga
noobineer has quit [Remote host closed the connection]
rohitksingh_work has quit [Read error: Connection reset by peer]
rohitksingh has joined ##openfpga
Bike has joined ##openfpga
ondrej2 has quit [Read error: Connection reset by peer]
ondrej2 has joined ##openfpga
kmehall has quit [Remote host closed the connection]
rohitksingh has quit [Quit: Leaving.]
kmehall has joined ##openfpga
<azonenberg>
gruetzkopf: I'm thinking 3-pair
<azonenberg>
the PHY i have in mind does clock recovery and doesn't require TX clock
<azonenberg>
But the FPGA i have can't do clock recovery on the RX clock
<azonenberg>
i may be able to do oversampling, TBD
<azonenberg>
have to look into that more
<azonenberg>
But for the time being i wanted to make sure i had RXC just in case i needed it
azonenberg_work has quit [Ping timeout: 252 seconds]
kmehall has quit [Read error: Connection reset by peer]
kmehall has joined ##openfpga
pie_ has quit [Ping timeout: 260 seconds]
fouric has quit [Ping timeout: 240 seconds]
azonenberg_work has joined ##openfpga
azonenberg_work has quit [Ping timeout: 256 seconds]
azonenberg_work has joined ##openfpga
digshadow has quit [Ping timeout: 260 seconds]
m_w has quit [Quit: leaving]
mumptai has joined ##openfpga
genii has joined ##openfpga
digshadow has joined ##openfpga
<mithro>
morning!
<mithro>
daveshah: ping?
<daveshah>
mithro: morning!
lexano has quit [Remote host closed the connection]
<kc8apf>
azonenberg: RGMII (and RMII for that matter) definitely fail over 6". I've run into problems with that trying to run NCSI from one side of a chassis to another.
<kc8apf>
SGMII has been run over 10ft of copper for years at Google
<kc8apf>
XAUI is also very forgiving
<azonenberg_work>
kc8apf: i was looking at using sgmii because of rgmii not working at long range
<azonenberg_work>
my concern was sgmii *using the xilinx lvds gpio*
<azonenberg_work>
rather than a GTX
<azonenberg_work>
there's no equalization or preemphasis
<azonenberg_work>
So your usable eye is a lot smaller
<kc8apf>
RGMII has a timing issue over 6"
<azonenberg_work>
well parallel busses in general fall flat after a bunch of pcb
<kc8apf>
SGMII over a backplane is pretty normal for large switch chassis
<kc8apf>
same with XAUI
<azonenberg_work>
yeah and i bet those use real serdes
<azonenberg_work>
:p
<kc8apf>
not always
<azonenberg_work>
xilinx's sgmii ip specifically warns that if you don't use a serdes it's limited to very short range chip to chip links on a single pcb
<kc8apf>
huh. I don't remember it being _that_ sensitive
<azonenberg_work>
i think this is more a general issue of using 1.25 Gbps Fmax LVDS IOs on a kintex7
<azonenberg_work>
you dont have much margin
<kc8apf>
I guess the receive side would have a hard time meeting LVDS after enough uncontrolled impedance
<azonenberg_work>
That too
<azonenberg_work>
So i'm thinking, lvds buffers on the backplane
<kc8apf>
makes sense to me
lexano has joined ##openfpga
sgstair has quit [Read error: Connection reset by peer]
<rqou>
(drama, offtopic) does anybody know what the heck is going on between @RealSexyCyborg and Vice?
sgstair has joined ##openfpga
kmehall has quit [Ping timeout: 276 seconds]
<whitequark>
you don't want to know
<whitequark>
it's one of those topics where the more you dig in the more insane it becomes
kmehall has joined ##openfpga
<rqou>
is it a repeat of what happened with Make or is it even worse?
<q3k>
rqou: iiuc: 1) vice interviews naomi, promises to not dig into her pesonal life 2) vice proceeds to dig into her personal life 3) naomi is pissed