GNUmoon has quit [Remote host closed the connection]
ur5us has quit [Ping timeout: 255 seconds]
Junxter has quit [Remote host closed the connection]
Junxter has joined #openscad
GNUmoon has joined #openscad
<zingos>
Hi, designing a rack for mounting ssds outside of a computer, how much space would it be wise to leave between them and between the first ssd and the table surface, for proper air flow? Is something like 3mm enough?
<teepee>
if that rack has fans, that sounds reasonable
<zingos>
Nah it won't have fans but it will be open, it won't be actually housing the entire ssds, just the sides and even that just the holes mostly
<othx>
zingos linked to "Double 2.5" SSD Mounting Bracket by jiaweichen" on thingiverse => 1 IRC mentions
GNUmoon has quit [Remote host closed the connection]
GNUmoon has joined #openscad
teepee has quit [Remote host closed the connection]
teepee has joined #openscad
teepee_ has joined #openscad
teepee has quit [Ping timeout: 240 seconds]
teepee_ is now known as teepee
petaflot has joined #openscad
<petaflot>
hello! I am making a simple anim (of a quite complex model), and since it takes a long time to render openscad does not write all the frames, on average only one out of 2. this is obviously a problem, is there a workaround? setting FPS to <1 writes no frames at all
linext has joined #openscad
linext_ has joined #openscad
linext has quit [Ping timeout: 258 seconds]
<teepee>
petaflot: never heard about that problem so far, what version is it? and on what platform?
<petaflot>
teepee: v2021.01
<petaflot>
on gentoo linux
<petaflot>
I notice that openscad keeps track of the frames written, so it very slowly writes the missing frames. but it is really, really slow : openscad will waste time rendering frames that have already been written
<petaflot>
IMHO when the "write frames" option is set, openscad should ignore the FPS setting and process all frames in order, starting from the specified $t value.
<teepee>
that's what it does
<petaflot>
it's really not the behaviour I can observe : if it really were the case, I would have no missing frames in that list http://dpaste.com/69CLH6QHA
<petaflot>
also, openscad is still running the animation loop and the "dump pictures" checkbox is still checked
<teepee>
could you try with the dev builds? there were quite a number of changes to the animation stuff
<teepee>
although more for the GUI so it's not hugely likely to behave differently for dumping the images
<teepee>
is that model online somewhere?
<petaflot>
I hope there's a button or keyboard shortcut to start/stop the anim, a way to reload the file without having to stop+start the anim every time. and also a button to reset $t to 0 :-]
<petaflot>
teepee: model is not online, it's for a design project which I am not allowed to share
* petaflot
looking for the source tree...
teepee has quit [Ping timeout: 240 seconds]
teepee has joined #openscad
<teepee>
hmm, extra unstable connection today :(
<teepee>
you can try the AppImage too
<petaflot>
I'd rather compile it... I've never had any problems (not true fro some other apps such as FreeCAD) and this way I can write my own patches
<teepee>
that's good too :)
<teepee>
I don't think it needs any special flags, the experimental features are not enabled than, but the animation changes should be active
<petaflot>
\me <3 OSS
<teepee>
\o/
<petaflot>
doesn't require one to systematically beg for some changes that are not suited to the majority of users...
<petaflot>
crap... gentoo only has CGAL < 5.2.2
<teepee>
true, you can still ask though, sometimes there's no problem adding those ;-)
<teepee>
that should still work
<teepee>
not recommended, but it should build down to 5.0.x which ships in ubuntu
<teepee>
a bit surprising, as it's now header-only it really easy to upgrade for the most part
<petaflot>
well I got the appimage dev build
<petaflot>
woohooo, nice UI controls for the animation :-)
<petaflot>
.. and perspective parameters :-)
<teepee>
yep, done by someone doing some animations for wikipedia with openscad (I believe)
<J22>
petaflot you can set $t=0; or with a switch $t=switch?$t:0; i use that in my animations
<petaflot>
J22: setting $t=0; works but I find it quite anoying
<J22>
i use a switch in the customizer so i can check the box when i need it (further it sets to 0 when rendering)
<J22>
and with alt arrow you can toggle 0 and 1 on the keyboard
<J22>
but more importand i use it to set it to certain values to tweak animations at t = .25 or .5 or ⅓ etc
<petaflot>
well... the dev snapshot has been trying to open my model for close to 10 minutes now, has been running at 100% CPU all this time. this seem very wrong, it only takes about 10 seconds with 2021.01
<teepee>
are there any of the experimental features enabled? that should not be the case by default
<teepee>
the fast-csg stuff is usually much faster then normal calculation, but there's still some cases where it can go crazy
<petaflot>
no features at all are enabled
<petaflot>
oh, ow it opened
<petaflot>
uh. I guess I'm gonna have to send you the model... there is some quite weird behaviour, ie. one frame took really long to process - and it's an empty image! (just rotating my object)
<petaflot>
at least the frames are being processed in sequence
<teepee>
maybe gpu gets angry at that frame?
<teepee>
is there anything in the console like "stopping rendering at 1000000" or so?
<petaflot>
nothing on the console. out of 150 frames it processed 1-3,5-50 properly and reasonnably fast ; now frame 51 is getting the CPU crazy and it seems to be stuck there
<petaflot>
FIY GPU is Radeon PRO WX 3200
<petaflot>
radeontop seems to show almost no use of the GPU
<teepee>
yeah, there's not a huge amount of GPU usage
<teepee>
you can try to enable the vertex-object stuff in preferences->features
<teepee>
this batches display calls, maybe that helps
rawgreaze_ has joined #openscad
rawgreaze has quit [Ping timeout: 240 seconds]
<petaflot>
there's a slghtly simplified version of my model http://dpaste.com/2NLB2E4CY, I have not been able to run through a full anim loop so far
<petaflot>
is there a way to specify initial viewport settings in a scad file?
<petaflot>
.. crashed again with vertex-* options enabled
<J22>
$vpt $vpr $vpd $vpf
<petaflot>
thanks J22
<J22>
the first two are vector3 the last two numbers
<petaflot>
yup I figured it out :-)
<petaflot>
this is very sweet
snaked has quit [Ping timeout: 260 seconds]
Guest94 has joined #openscad
Guest94 has quit [Quit: Client closed]
Virindi has quit [Quit: ZNC 1.8.2+deb2+b1 - https://znc.in]
lostapathy has quit [Quit: WeeChat 3.5]
lostapathy has joined #openscad
Virindi has joined #openscad
Junxter has quit [Read error: Connection reset by peer]
joseph_ has quit [Quit: Quit!]
<InPhase>
petaflot: So what this will need is a binary search process to find the part that is causing the huge delay. What you do, is comment out half of the design and see if that huge performance fail happens. If it no longer has the performance fail, comment out the other half instead. Continue with this progression until you've reduced it to the minimal set of top level code that produces the failure, and
<InPhase>
then start commenting out the include/use dependencies that you can get rid of. Then when you get to the end, copy/paste the minimal set of content from the remaining dependency that causes the performance fail to make a small self-contained failing testcase, and try once more to shrink it down if you can. This will probably give you a very small simple design that massively underperforms, and then
<InPhase>
it's possible to fix.
<InPhase>
Basically, we want to find the minimal failing testcase, because that will highlight the exact phenomenon which trips it up.
<InPhase>
It ultimately doesn't need to look anything like the original design, but just reproduce the bad behavior.
<J22>
we need a debug mode that is giving timecodes with code lines Ü
<J22>
(also "mesh not closed" needs line numbers .. )
Guest28 has joined #openscad
<Guest28>
Please tell me how to get the sum of the function values? I've already asked this question, but I still don't understand how to do it. The variable i takes values from 0 to 5 for example.
<Guest28>
function tube_d(i) = [2*(i*l_sec+diametr_first_inner_tube/2+th*(i-1))];
Virindi has quit [Quit: ZNC 1.8.2+deb2+b1 - https://znc.in]
Virindi has joined #openscad
<InPhase>
Guest28: Do you mean summing the values in a list?
<Guest28>
I need the sum of the values of the presented function.
<J22>
Guest28 that function gives you only one value .. so you create another variable (array) that stores all 5 values .. then you can sum them
<J22>
an array can be sum up either with recursion or as t-paul showed by multiply with an same sized array of 1s
<J22>
array=[for(i=[1:5])tube_d(i)];
<Guest28>
Interested in how to sum an array using recursion.
Virindi has quit [Quit: ZNC 1.8.2+deb2+b1 - https://znc.in]
<J22>
or you can do it manually result = array[0]+array[1] …+ array[4]
<Guest28>
The manual method is not suitable, because "i" takes different values.
<J22>
or better function sumV(v=array,sum=0,i=0)=i<len(v)?sumV(v=v,sum=sum+v[i],i=i+1):sum;
Virindi has joined #openscad
<J22>
so as long i is smaller the length of your input array it will call itself (recursion) and Add the current array position (i) to the sum value (sum=sum+v[i])
<J22>
Guest28 if all numbers are summed up it will return sum i=len(v) ↦ :sum
<J22>
Guest that is the other methode this is a vector multiplication - not recursive
teepee_ has joined #openscad
<Guest28>
echo ([for(i=[1:end]) 1] * [for (i=[1:end]) tube_d(i)]); An acceptable option, but not very beautiful. Yes, I want to do it through recursion, but it's not getting there yet.
<J22>
i have posted you how
<J22>
function sumV(v=array,sum=0,i=0)=i<len(v)?sumV(v=v,sum=sum+v[i],i=i+1):sum;
<J22>
here again
<J22>
put your array into v and it will sum the parts
<Guest28>
sorry, history not save
teepee has quit [Ping timeout: 240 seconds]
teepee_ is now known as teepee
<J22>
could you follow what the function does or what part is unclear?
<J22>
function sumV(v=array,sum=0,i=0)=i<len(v)?sumV(v=v,sum=sum+v[i],i=i+1):
<J22>
sum;
<J22>
you can condense this further without the sum variable but that way it may be easier to grasp
<J22>
have you replaced "array" with an actual array or define array =[ 1, 2, 3 ];
<J22>
array=[for(i=[1:5])tube_d(i)];
<Guest28>
error have if i add --- echo(sumV());
<Guest28>
array=[for(i=[1:5])tube_d(i)]; - this is written
Guest28 has quit [Quit: Client closed]
<J22>
hmm
Sauvin has quit [Ping timeout: 244 seconds]
Sauvin has joined #openscad
Chillum has joined #openscad
Chillum2 has quit [Ping timeout: 240 seconds]
j1mu5 has joined #openscad
j1mu5 has quit [Client Quit]
snaked has joined #openscad
GNUmoon has quit [Remote host closed the connection]
ur5us has joined #openscad
GNUmoon has joined #openscad
teepee has quit [Remote host closed the connection]
teepee has joined #openscad
joseph_ has joined #openscad
muesli4 has joined #openscad
muesli has quit [Quit: Ping timeout (120 seconds)]
muesli4 is now known as muesli
<joseph_>
InPhase teepee: Before GSoC coding starts I think I'll try setting up an IRC bouncer so that I can get push notifications on my phone when I'm mentioned. I haven't done this before but there are some GH repos I can likely set up on my own. Just wanted to check with you first to see if you have cautions about this specific IRC server's config, etc.
<teepee>
not really, I'm using ZNC with SASL authentication, that's probably the best bet for IRC
<teepee>
for mobile it might be easier to go via Matrix though
snaked has quit [Read error: Connection reset by peer]
snaked has joined #openscad
teepee_ has joined #openscad
teepee has quit [Ping timeout: 240 seconds]
teepee_ is now known as teepee
noonien has quit [Ping timeout: 240 seconds]
<InPhase>
joseph_: Yeah, that's a pretty reasonable choice. I personally go with irssi inside GNU screen with a VPS, which I then access via the termius app when I want to do by phone.
<InPhase>
But the bouncers like ZNC are a very reasonable choice.
snaked has quit [Read error: Connection reset by peer]