Mail Archives: geda-user/2018/05/17/07:05:06
--2SVASOMWQIEFIQSIUFNFJnhgwp
Content-Transfer-Encoding: quoted-printable
Content-Type: text/plain; charset=UTF-8
Hello Igor, Dnia 17 maja 2018 09:18 michalwd1979 <gedau AT igor2 DOT repo DOT hu&g=
t; napisa=C5=82(a): Please turn off the padstack "layer" in the la=
yer selector so we don't render them, then do the usual fps test (big =
window, zoom to extent, benchmark) with two different setup: After loadi=
ng refdes.lht the program is completely unusable, so yes - this must be the=
issue. I took about 30s to open the board and type "benchmark()". =
Benchmark says that speed is 0.1, however it seems slower for me. Doing any=
thing with pcb-rnd (cursor, pan, zoom) keeps cpu at 100% for ~30s to ~1.5mi=
nute. 1. subcircuits "layer" =C2=A0off, but top silk on - this w=
ay we measure how much does it cost to render all the refdes text objects =
on silk Much faster, benchmark says 1.4, however still rather unusable. I=
t is possible however to zoom or pan (2-3s for a single step zoom). 2.=
subcircuits "layer" on, but top silk off - this way we render abou=
t the same thing, just by a different part of the code Turning subcircui=
ts ON stops everything for very long time (~30-40s) only after that it was =
possible to do benchmark. Result is 0.1, but as I said it seems slower. If =
the subcircuits are ON, turnig top-silk ON/OFF makes no perceptible differe=
nce. I crafted this file to have the dash freq set to -1 so we don't=
measure the dashed line rendering in (would be noise on the results). I=
f these results will show what I expect, I know where to poke the code. M=
eanwhile I did some more testing with sw render and Ade did some more test=
ing with opengl render. We both identified parts of the rendering code tha=
t could be optimized. We'll release pcb-rnd 2.0.0 in about 3 weeks, so =
we don't start messing with the rendering now, but after the release w=
e will speed some parts up a bit. Thanks for testing! No problem I'=
;m glad that I can help with development. The main factor that keeps me f=
rom pcb-rnd is the speed and transparency rendering, I would love to see th=
e project with the rendering speed matching opengl pcb. By the way, am I=
the only person that noticed speed/text problems? Is my PC really so much =
slower then current machines or was my board so complex :-)? Best Regards,=
Michael Widlok=0D
--2SVASOMWQIEFIQSIUFNFJnhgwp
Content-Transfer-Encoding: quoted-printable
Content-Type: text/html; charset=UTF-8
<div>Hello Igor,<br></div><div class=3D"nh_extra"><p>Dnia 17 maja 2018 09:1=
8 michalwd1979 <gedau AT igor2 DOT repo DOT hu> napisa=C5=82(a):<br></p><blockqu=
ote class=3D"nh_quote" style=3D"border-left: 2px solid #999; padding-left: =
8px; margin: 0;"><div id=3D"gwp94ced498"><div>Please turn off the padstack =
"layer" in the layer selector so we don't<br></div><div>render them, then d=
o the usual fps test (big window, zoom to extent,<br></div><div>benchmark) =
with two different setup:<br></div></div></blockquote></div><div><br></div>=
<div>After loading refdes.lht the program is completely unusable, so yes - =
this must be the issue. I took about 30s to open the board and type "benchm=
ark()". Benchmark says that speed is 0.1, however it seems slower for me. D=
oing anything with pcb-rnd (cursor, pan, zoom) keeps cpu at 100% for ~30s t=
o ~1.5minute. <br></div><div class=3D"nh_extra"><blockquote class=3D"nh_quo=
te" style=3D"border-left: 2px solid #999; padding-left: 8px; margin: 0;"><d=
iv id=3D"gwp94ced498"><div><br></div><div>1. subcircuits "layer" off,=
but top silk on - this way we measure how<br></div><div>much does it cost =
to render all the refdes text objects on silk<br></div></div></blockquote><=
/div><div><br></div><div>Much faster, benchmark says 1.4, however still rat=
her unusable. It is possible however to zoom or pan (2-3s for a single step=
zoom).<br></div><div> <br></div><div class=3D"nh_extra"><blockquote class=
=3D"nh_quote" style=3D"border-left: 2px solid #999; padding-left: 8px; marg=
in: 0;"><div id=3D"gwp94ced498"><div><br></div><div>2. subcircuits "layer" =
on, but top silk off - this way we render about the<br></div><div>same thin=
g, just by a different part of the code<br></div></div></blockquote></div><=
div><br></div><div>Turning subcircuits ON stops everything for very long ti=
me (~30-40s) only after that it was possible to do benchmark. Result is 0.1=
, but as I said it seems slower. If the subcircuits are ON, turnig top-silk=
ON/OFF makes no perceptible difference.<br></div><div><br></div><div class=
=3D"nh_extra"><blockquote class=3D"nh_quote" style=3D"border-left: 2px soli=
d #999; padding-left: 8px; margin: 0;"><div id=3D"gwp94ced498"><div><br></d=
iv><div>I crafted this file to have the dash freq set to -1 so we don't mea=
sure<br></div><div>the dashed line rendering in (would be noise on the resu=
lts).<br></div><div><br></div><div>If these results will show what I expect=
, I know where to poke the code.<br></div><div><br></div><div>Meanwhile I d=
id some more testing with sw render and Ade did some more<br></div><div>tes=
ting with opengl render. We both identified parts of the rendering code<br>=
</div><div>that could be optimized. We'll release pcb-rnd 2.0.0 in about 3 =
weeks, so<br></div><div>we don't start messing with the rendering now, but =
after the release we<br></div><div>will speed some parts up a bit.<br></div=
><div><br></div><div>Thanks for testing!<br></div></div></blockquote></div>=
<div><br></div><div>No problem I'm glad that I can help with development. <=
br></div><div>The main factor that keeps me from pcb-rnd is the speed and t=
ransparency rendering, I would love to see the project with the rendering s=
peed matching opengl pcb.<br></div><div><br></div><div> By the way, am I th=
e only person that noticed speed/text problems? Is my PC really so much slo=
wer then current machines or was my board so complex :-)?<br></div><div>Bes=
t Regards,<br></div><div>Michael Widlok<br></div>
--2SVASOMWQIEFIQSIUFNFJnhgwp--
- Raw text -