<sb0>
could be that different RNGs cause the P&R to fail randomly...
<rjo>
sb0: gnu 3.81
<rjo>
it takes weirdly long time here as well.
<rjo>
it doesn't do the #include at all. the output is the input. but there is no error about not finding the include either.
<rjo>
regarding this and misoc tooling in general: once misoc is a module, all this can go away and the bitstream/bios/etc can be built in the artiq build dir. then the $target.py can build/do whatever it likes and also just expose some cli options like artiq_pipistrello.py -s AMP --csr-addr CSR_KERNEL_CPU_BASE
<rjo>
but it does not seem to work that easily for vivado. might require more tweaking.
<sb0>
GNU Make 4.1 here
<rjo>
where does CC_normal come from?
<sb0>
misoc software/common.mak
<sb0>
once misoc is a module, where to put the bios sources?
<rjo>
python packages have these package "data" mechanisms. maybe that works well enough.
<rjo>
but libcrt etc and the verilog trees might be more inconvenient.
<rjo>
otoh, you can just demand that people install misoc editably (pip --user -e .) and then everything can stay where it is, just need to point the .o and generated files to where the target wants them.
<rjo>
gotta go. see you.
<GitHub85>
[artiq] sbourdeauducq pushed 1 new commit to master: http://git.io/vebhj
<mithro>
sb0: we are using travis to kick of a build on a more beefy dedicated machine for the P&R
nengel has quit [Remote host closed the connection]
nengel has joined #m-labs
FabM has joined #m-labs
<sb0>
whitequark, seems those plasma windows are stable, but you need something like 8kW per centimeter of diameter (of a circular window) to sustain them
<whitequark>
yeah they eat a lot of power
<sb0>
that figure sound a bit dubious, i'd expect that to scale with the square of the diameter. but yes, they are power hungry
<whitequark>
I still think EB sintering could be even practical with these
<whitequark>
adding material becomes a non-problem
<sb0>
maybe focusing the electron beam through a microscopic plasma window would work
<whitequark>
no pumping down the chamber... just treat it nearly as you would a nozzle
<sb0>
yes
<whitequark>
it might not be viable after all because of all the power it consumes, though
<sb0>
and much easier to troubleshoot mech problems with adding material
<sb0>
make it a 0.1mm window...
<sb0>
or less
<whitequark>
indeed. I should look at how it scales down, someday
<sb0>
though the divergence of the electron beam might be a problem
<sb0>
ideally, you'd want it to be focused at both the workpiece and the plasma window
<whitequark>
well, the paper mentions the window pinching the beam with lorentz force
<whitequark>
it sounds like if you get a slightly defocused beam at the window, you could get back one focused in proximity of the window
<sb0>
and bah, 8kW is a reasonable amount of power ;)
<whitequark>
I mean... at some point you're just hitting ROEI (as in return on energy invested) ceiling
<whitequark>
say against a traditional EB sintering process
<sb0>
but you want a focused beam at the window as well, so that you can make the window small
<whitequark>
hence slightly defocused. balance the size of the window against nozzle-workpiece distance
<sb0>
I wonder what sort of electron energy we are talking about, too
<sb0>
that thing would probably produce a lot of x-rays and what not
<sb0>
also when the electrons are hitting the atmosphere
sj_mackenzie has quit [Remote host closed the connection]
SturmFlut has quit [Ping timeout: 250 seconds]
SturmFlut has joined #m-labs
SturmFlut has quit [Ping timeout: 250 seconds]
SturmFlut has joined #m-labs
<rjo>
sb0: that phase_half counter could potentially be dropped (use phase_sel) but i got distracted before i could convince myself that it would work.