Current Server Time: 15:17 (Central Europe)

#apertus IRC Channel Logs

2015/03/11

Timezone: UTC


00:30
Bertl
off to bed now ... have a good one everyone!
00:30
Bertl
changed nick to: Bertl_zZ
00:34
MatthiasF
left the channel
01:10
intracube
left the channel
01:22
g3gg0
joined the channel
02:05
slikdigit
left the channel
02:22
g3gg0
left the channel
02:22
g3gg0
joined the channel
05:06
ItsMeLenny
joined the channel
07:15
ItsMeLenny
left the channel
07:16
ItsMeLenny
joined the channel
07:25
cbohnens|away
changed nick to: cbohnens
07:39
masplund|away
changed nick to: masplund
07:40
masplund
changed nick to: masplund|away
08:04
g3gg0
left the channel
08:10
MatthiasF
joined the channel
08:12
se6astian|away
changed nick to: se6astian
08:12
se6astian
good morning
08:16
Bertl_zZ
changed nick to: Bertl
08:16
Bertl
morning folks!
08:40
jucar
joined the channel
09:01
ItsMeLenny
left the channel
09:07
Jin|away
changed nick to: Jin^eLD
09:11
jucar
left the channel
09:25
danieel
left the channel
09:57
Francky
joined the channel
09:58
Francky
hi all
10:13
se6astian
hey Francky
10:14
Francky
how was the studio time ?
10:16
se6astian
good good
10:16
se6astian
I think it will result in another great team talk episode
10:17
se6astian
unfotunately its raining today so we cannot film unloading the CNC mill...
10:17
se6astian
hopefully tomorrow morning though
10:20
danieel
joined the channel
10:29
Francky
what is "CNC mill" ?
10:29
Bertl
https://wiki.apertus.org/index.php?title=Compact_CNC_Mill_Evaluation#Tormach
10:31
Francky
ok i didn't know the english term
10:32
Francky
you plan to produce yourself the camera in production ?
10:33
Bertl
if possible, we plan to build the case or at least certain parts ourselves
10:33
Bertl
but the main purpose is to have something for quick prototyping
10:34
Francky
ok
10:36
Francky
did you find a compromise about all the board interfraces finally ?
10:37
Bertl
I think we almost had it last time, so we should make two 20-24 pin interfaces on each side
10:37
Bertl
I still have to check if we go for an iCE40 or maybe MachXO2
10:38
Bertl
i.e. I have to make a feature comparison, unless somebody volunteers to do that :)
10:38
g3gg0
joined the channel
10:39
Francky
somebody who have skills in fpga, so... you ! :)
10:42
Bertl
yeah, well, it's more digging through datasheets
10:45
davidak
joined the channel
11:03
fsteinel
joined the channel
11:18
se6astian
brb reboot
11:19
se6astian
changed nick to: se6astian|away
11:20
se6astian|away
changed nick to: se6astian
12:23
intracube
joined the channel
12:46
se6astian
bbl
12:46
se6astian
changed nick to: se6astian|away
13:23
Francky
Bertl, I have general question about axiom software (on zyng PL)
13:23
Francky
on the wiki you said that the pl take images from sensor and store them into memory
13:24
Bertl
correct
13:24
Francky
so you use the external DDR memory, and the PL access to it throught AXI right ?
13:24
Bertl
the memory manager, yes
13:24
Francky
with something like a rouded buffer
13:24
Bertl
kind of, it is more a frame buffer actually
13:24
Francky
the memory manager is a part of the PS ?
13:25
Francky
or a software on PS ?
13:25
Bertl
it is part of the zynq hardened peripherials
13:25
Bertl
it has connections to PS and PL
13:27
Bertl
http://blog-imgs-51-origin.fc2.com/m/a/r/marsee101/Zynq_test_1_120715.png
13:27
Francky
is it DMA controller or another thing ?
13:27
Francky
thanks for the picture
13:27
Bertl
it manages the DDR memory independant from PL or PS
13:28
Bertl
you can run DMA (on the PS) to the DDR memory
13:28
Bertl
or you cann access it via AXI ports on the PL
13:30
Francky
ok so the chain is sensor -> PL -> AXI -> memory controller -> DDR
13:30
Francky
the PS is never used
13:30
Bertl
correct
13:30
Francky
the PS is used to configure the sensor, the hdmi or to take a snapshot from the memory
13:30
Francky
right ?
13:31
Bertl
yep
13:31
Francky
thanks
13:42
Francky
in fact i am reading "The Zynq book" and i am trying to make a parrallela between your design and how is build the zynq to understand the system
13:42
Francky
when you say it is a frame buffer, do you mean that only the last video image is storen into memory ?
13:46
Bertl
it supports quadrupple buffering in the current setup
13:46
Bertl
but the design allows arbitrary frame numbers (as long as you have enough memory)
13:48
Francky
this is why you need a adress generator IP ?
13:48
Bertl
correcy
13:48
Bertl
*correct
13:48
Francky
and this IP just calculate the adress of the start of the frame or it need to calculate every adress of every pixel ? ie : is there burst mode (you give the start adress and the data in burst mode) ?
13:48
Bertl
yes, it calculates the burst addresses
13:52
cbohnens
changed nick to: cbohnens|away
13:52
Francky
ok. I am trying to build the cmv_io2 project but the timing report says that the timing constraints are not meet. But how does it know what are the constraints to meet ?
13:56
Bertl
complicated calculations :)
13:56
Bertl
you might need to rerun certain parts to get timing closure
13:56
Bertl
i.e. the tools not always manage to get the routing right
13:57
Francky
so from the same project source, the tool might build different routing ?
13:57
Bertl
correct
13:58
Francky
beurk....
13:59
Francky
so you compile your project time after time as far as you get a correct routing ?
13:59
Bertl
kind of, i.e. you rerun the routing parts with different options
14:00
Francky
but I use your project with your tcl script, doesn't it integrate the "good" options ?
14:00
Bertl
but in general you can create designs and constraints which work every time, it just gets harder the more complicated the design is
14:00
Bertl
as I said, the tools are non-deterministic
14:01
Bertl
i.e. they often seem to make random decisions and sometimes they are good and at other times not so much
14:02
Bertl
I got rather annoyed by this fact at some point, and the many pblocks help to guide the tools somewhat, but it's not perfect
14:02
niemand
joined the channel
14:08
Francky
it is disturbing because you cannot garanty that the built program will have the same performance in each builder
14:11
niemand
left the channel
14:11
intracube
afternoon Bertl, Francky :)
14:12
Francky
hi intracube
14:13
Bertl
yeah, I have a bunch of ideas how to improve that for the Beta
14:13
Bertl
starting from better designs, over automated placing, to an incremental build system which retries until it succeeds
14:14
Francky
designs = vhdl code or placement and routing of blocks ?
14:15
Bertl
better way of handling interconnects
14:15
Francky
so you plan to make manually the job of the placer and router ?
14:16
Bertl
at least to keep the tools on a short leash :)
14:19
Bertl
I can understand that the tools have to make somewhat random decisions, because the search space for routing becomes very large with recent FPGAs
14:19
Bertl
and if you take the typical PCB autorouter and their success rate ...
14:24
Francky
autorouter makes me ill
14:28
Francky
but i thought that xilinx tools should be very strong
14:28
Francky
but in fact it is quite undestandable because it IS routing in fact
14:33
Bertl
yes
14:33
Bertl
off for now ... bbl
14:33
Bertl
changed nick to: Bertl_oO
14:35
Andrej74
left the channel
15:10
fadro
left the channel
15:28
se6astian|away
changed nick to: se6astian
15:30
se6astian
back
15:32
se6astian
I just visited the event location for the AXIOM kickoff
15:32
se6astian
will be very nice!
16:14
dmj726
left the channel
16:17
dmj726
joined the channel
16:22
niemand
joined the channel
16:57
intracube
se6astian: where will it be held?
16:59
se6astian
at the angewandte innovation lab, central vienna
16:59
se6astian
announcement will follow soon
17:00
mars_
can't wait!
17:02
intracube
se6astian: nice :)
17:07
Francky
left the channel
17:47
jucar
joined the channel
17:53
slikdigit
joined the channel
17:53
slikdigit
left the channel
17:53
slikdigit
joined the channel
18:03
danieel
left the channel
18:09
intracube_
joined the channel
18:12
intracube
left the channel
18:46
danieel
joined the channel
19:13
niemand
left the channel
19:26
niemand
joined the channel
19:35
niemand
left the channel
19:48
niemand
joined the channel
20:20
Jin^eLD
changed nick to: Jin|away
20:31
intracube_
changed nick to: intracube_caffei
20:32
intracube_caffei
changed nick to: intracube
20:33
troy_s
se6astian: I re-added that photo. It disappeared. Was it a URL link?
20:34
se6astian
I think I didnt add the actual image, just suggested it with URL in the comments
20:34
se6astian
but its fine that way as well now
20:35
se6astian
I think the first half of the article with the two images is great now
20:36
se6astian
what images could we add to spice up the second half?
20:38
se6astian
about the 1931 color experiments
20:39
se6astian
instead of the result curves
20:39
se6astian
what did the actual test look like?
20:39
se6astian
did they show people color swatches?
20:42
intracube
left the channel
20:43
se6astian
seems they showed a white screen with two separated colors
20:43
se6astian
http://fabricgraphicsmag.com/repository/5/10345/full_0311_f2side1_4.jpg
20:43
intracube
joined the channel
20:52
lab-bot
sebastian created T311: create HDMI module holder to 3d print. http://lab.apertus.org/T311
21:14
se6astian
we got sponsored by a flower delivery company
21:14
se6astian
so if you need any flowers deliverd
21:14
se6astian
or chocolate
21:14
se6astian
plants...
21:14
se6astian
you know where to go now
21:14
se6astian
http://www.serenataflowers.com/
21:15
se6astian
thats what 100⬠of camera development support gets you :)
21:16
intracube
lol
21:28
Bertl_oO
well, that's big advertising I'd say ... after all the channel is logged :)
21:29
jucar
left the channel
21:33
slikdigit
left the channel
21:40
niemand
left the channel
21:55
jucar
joined the channel
22:17
se6astian
time for bed
22:17
se6astian
goood night
22:17
se6astian
changed nick to: se6astian|away
22:34
se6astian|away
changed nick to: se6astian
22:35
se6astian
changed nick to: se6astian|away
22:49
aombk2
joined the channel
22:50
aombk
left the channel
23:04
intracube_
joined the channel
23:04
jucar
left the channel
23:06
g3gg0
left the channel
23:07
intracube
left the channel
23:34
MatthiasF
left the channel