**** BEGIN LOGGING AT Thu Jun 25 02:59:57 2020
Jun 25 03:04:41 <MattB0ne> zmatt: to configure the UART I need to define everything for the __fields_ structure for pruss.uart?
Jun 25 03:06:18 <zmatt> MattB0ne: beware that I don't think I ever got around to _testing_ the icss.uart struct in py-uio
Jun 25 03:06:38 <MattB0ne> well i can be the guinea pig
Jun 25 03:07:18 <zmatt> at a minimum you need to configure lcr and the divider (div_lsb/div_msb)
Jun 25 03:07:39 <zmatt> and then enable the receiver and transmitter in sysconfig
Jun 25 03:09:05 <zmatt> oh and fcr I think but it looks like its definition is missing
Jun 25 03:14:53 <MattB0ne> with these comments regarding bits do I pass in an int whos binary would have the bits enabled that I want
Jun 25 03:14:57 <MattB0ne> so for example
Jun 25 03:15:35 <MattB0ne> sysconfig
Jun 25 03:15:51 <MattB0ne> i would want bit 1 as 1 always so I would only pick even integers
Jun 25 03:16:47 <MattB0ne> and I I want bit 13 and 14 enabled so I need to have 2^13 and 2^14 in my number set to 1 by picking 8192
Jun 25 03:17:18 <MattB0ne> that is probably wrong
Jun 25 03:17:23 <MattB0ne> I think I will over flow uint
Jun 25 03:17:38 <MattB0ne> doing that since 2^14 would be 16384
Jun 25 03:18:05 <MattB0ne> so I would use 24576 which would be force the 13th and 14th bits to be 1
Jun 25 03:18:06 <ayjay_t> yeah zmatt the width and height is powers of 2 for sure
Jun 25 03:21:26 <ayjay_t> i'm going to bed
Jun 25 03:21:29 <ayjay_t> i sent the bad news
Jun 25 03:24:42 <MattB0ne> ok I know I need to pass ints like pruss.uart(int,int,int)
Jun 25 03:24:51 <MattB0ne> now I just need to get the values
Jun 25 03:43:08 <MattB0ne> scratch what I wrote
Jun 25 03:43:12 <MattB0ne> I need to use bit operators
Jun 25 03:43:17 <MattB0ne> need to read how you do it
Jun 25 03:43:25 <MattB0ne> never used << or >> before
Jun 25 03:45:58 <zmatt> MattB0ne: I just added a .initialize method to the uart
Jun 25 03:46:04 <zmatt> (pushed to git)
Jun 25 03:46:27 <MattB0ne> ok
Jun 25 03:46:41 <zmatt> at minimum it needs just the divisor
Jun 25 03:47:17 <zmatt> the baudrate will be UART_CLK / oversampling / divisor   where oversampling is either 16 or 13 (defaults to 16)
Jun 25 03:47:26 <zmatt> I think the UART_CLK is 192 MHz but I'd need to check
Jun 25 03:47:54 <zmatt> yep, 192 MHz
Jun 25 03:48:15 <zmatt> so if you leave oversampling alone the baudrate will be 12 MHz / divisor
Jun 25 03:51:28 <zmatt> e.g. for a baudrate of 460800 you'd use divisor=26
Jun 25 03:51:53 <MattB0ne> ok cool
Jun 25 03:52:56 <MattB0ne> going to use 1250 and stick with good oled 9600bps
Jun 25 03:52:57 <MattB0ne> for now
Jun 25 03:53:18 <zmatt> for initial testing that makes sense
Jun 25 03:55:37 <zmatt> hmm I think every pru-icss has a 192 MHz uart clock, so I could just take baudrate as argument
Jun 25 03:57:41 <MattB0ne> is this like the ecap such that once it is initialized it is active and functioning?
Jun 25 04:00:00 <MattB0ne> also div_lsb why do you need to right shift the divisor 8 spots
Jun 25 04:00:37 <MattB0ne> i will take a look in the TRM
Jun 25 04:00:42 <MattB0ne> later
Jun 25 04:38:00 <zmatt> god I hate the 8250-like uart register interfaces
Jun 25 04:40:51 <MattB0ne> i cannot appreciate the statement due to a lack of knowledge
Jun 25 04:40:59 <MattB0ne> i see you are pushing some updates
Jun 25 04:41:03 <MattB0ne> what are you tinkering with
Jun 25 04:42:36 <zmatt> initialize now takes baudrate, and for testing I made a write method but I already know that's not the best idea, I should make it a combined write+read method
Jun 25 04:46:19 <MattB0ne> ok i will repull
Jun 25 04:47:05 <zmatt> but I don't really feel like writing that combined write+read method right now
Jun 25 04:50:50 <MattB0ne> ok
Jun 25 04:51:56 <MattB0ne> I see the write method
Jun 25 04:52:36 <MattB0ne> what do I read with
Jun 25 04:52:44 <MattB0ne> i can try messing with it before hitting the sack
Jun 25 08:10:01 <RC9190> I am having so much trouble trying to get the "examples" to work. All the documentation on the site talks about the bonescript interface but none mention the cloud9 ide. and yet thats Iall I can access on the BBBW I have
Jun 25 08:10:32 <RC9190> can I get the image I have to work as its described on the bone101 web pages?
Jun 25 09:07:14 <zmatt> I know he's gone but, https://beagleboard.org/Support/bone101 literally describes using the cloud9 ide, so I wonder what he's talking about
Jun 25 10:06:06 <Parduz> Why i have this error in Journal, even if the file is there: slim[209]: /usr/bin/X11/xauth:  file /home/debian/.Xauthority does not exist
Jun 25 10:09:02 <Parduz> my journal: https://pastebin.com/bevzbMKk
Jun 25 12:56:14 * jkridner[m] needs to remove references to the the bonescript RPC sooner than later. :-/
Jun 25 12:56:38 <jkridner[m]> unless there is really enough demand to actually fix it.
Jun 25 12:59:50 <zmatt> what's wrong with it still?
Jun 25 13:00:14 <zmatt> has the socket activation thing been fixed yet?
Jun 25 13:03:01 <zmatt> I'm also puzzled by why it didn't break sooner, given that it depends on the rather broken "systemd" module which does not support node 0.11.0 or later
Jun 25 13:04:48 <zmatt> I don't strongly care either way obviously since I don't use it
Jun 25 13:37:28 <MattB0ne> zmatt: after I run the script and init the UART how do I read
Jun 25 13:37:43 <MattB0ne> do I need a special method for that or is it something generic
Jun 25 13:37:51 <MattB0ne> I see you have a write procedure
Jun 25 13:37:57 <MattB0ne> do I not need a read ?
Jun 25 13:38:55 <jkridner[m]> the website doesn't know to look for it on port the right port.
Jun 25 14:34:32 <ayjay_t> how hard is my life going to be trying to put this onto a board jkridner[m] https://www.ti.com/product/TCI6638K2K
Jun 25 14:54:47 <Letothe2nd> ayjay_t: depends on the means you have at your disposal
Jun 25 14:58:06 <ayjay_t> Let's consider the worst case scenario
Jun 25 14:58:19 <Letothe2nd> ayjay_t: undoable.
Jun 25 14:58:21 <ayjay_t> I have KiCAD, an oscilloscope, a reflow station, and like $10k
Jun 25 14:58:30 <ayjay_t> okay not the worst worst case scenario
Jun 25 14:58:44 <Letothe2nd> semi-worst case scenario: undoable
Jun 25 14:59:02 <ayjay_t> Haha this is unhelpful
Jun 25 14:59:09 <ayjay_t> what are the minimum requirements for laying this board down
Jun 25 14:59:23 <ayjay_t> and what type of costs are involved in working with people who've already done a lot of the work
Jun 25 15:00:42 <ayjay_t> Letothe2nd: I'm trying to build a 5 Gigapixel microscope
Jun 25 15:01:18 <ayjay_t> I was raised on the Beaglebone Black, I would love to use a TI part
Jun 25 15:01:39 <Letothe2nd> well its 1500something finest pitch BGA, require high speed design, probably like a 20 layer board....
Jun 25 15:01:51 <Letothe2nd> for the 10k you won't even get the naked PCBs probably
Jun 25 15:02:02 <Letothe2nd> well maybe the PCBs.
Jun 25 15:02:08 <ayjay_t> well the dev board costs $1500
Jun 25 15:02:17 <ayjay_t> Do people who use this actually spin their own boards then?
Jun 25 15:02:38 <ayjay_t> I mean maybe one or two has, but there must be a supply line
Jun 25 15:03:06 <Letothe2nd> buying a dev board that is fully (or mostly) debugged plus run through a well working manufacturing line is in no way comparable to spinning a custom board.
Jun 25 15:03:21 <ayjay_t> yeah no i know that
Jun 25 15:03:37 <Letothe2nd> yeah sure there are folks who can do that, but i'd estimate the costs in multiples of 100k$
Jun 25 15:03:38 <ayjay_t> Letthe2nd I'm not trying to come across like an idiot even if I do seem casual
Jun 25 15:03:49 <ayjay_t> the business need here is real, the funding is real, the engineering is real
Jun 25 15:04:04 <Letothe2nd> ayjay_t: 10k$ is no real funding for such a board.
Jun 25 15:04:21 <ayjay_t> no the fudning is real
Jun 25 15:04:25 <Letothe2nd> its enough for a PoC on the dev kit at most.
Jun 25 15:04:45 <ayjay_t> $10k is what i'd spend out of my own wallet before i talk to anyone at the company
Jun 25 15:04:48 <ayjay_t> that's why i threw that out there
Jun 25 15:05:25 <Letothe2nd> like i said, 10k gets you the dev board plus time to work on it. designing a custom board: multiples of 100k$
Jun 25 15:05:28 <Letothe2nd> and yes, i'm certain.
Jun 25 15:05:35 <ayjay_t> i'm not questioning you
Jun 25 15:05:45 <ayjay_t> well i'm literally questioning you but i'm not sceptical
Jun 25 15:05:47 <ayjay_t> skeptical
Jun 25 15:05:58 <Letothe2nd> go ahead if you don't believe  me :)
Jun 25 15:06:09 <ayjay_t> my guess is that there are companies that have broken the chip out and work with other companies to integrate it onto their design
Jun 25 15:06:24 <ayjay_t> or there _should_ be
Jun 25 15:06:43 <Letothe2nd> if you find a som that properly incorporates it, things might change. but this is pretty specific tech, so i doubt it.
Jun 25 15:07:44 <Letothe2nd> for 10k you can get a prototype run of a custom board on something like the am3xxx, 10pcs if you find somebody interested, i'd guess.
Jun 25 15:07:45 <ayjay_t> Yeah something as nice as a SOM i don't expect
Jun 25 15:08:17 <ayjay_t> but even a board would be nice
Jun 25 15:08:23 <ayjay_t> well thats probably what you mean
Jun 25 15:08:27 <ayjay_t> when you said SOM i was thinking octavo
Jun 25 15:08:29 <ayjay_t> i guess thats a SIP
Jun 25 15:08:38 <Letothe2nd> ayjay_t: given the speeds and number of interconnects, anything other than a som doesn't make any sense as you cannot just route all that stuff through pins.
Jun 25 15:09:01 <Letothe2nd> but again, my $.02. make of it what you think.
Jun 25 15:09:33 <ayjay_t> no i appreciate the discussion
Jun 25 15:10:21 <ayjay_t> it's just hard to believe that the only two options on the market are to spend $500k spinning out a board and setting up manufacturing or buying a dev board
Jun 25 15:11:05 <ayjay_t> the guys who make the dev board, i imagine, would be forming partnerships- probably in defense, i wonder if our CEO could form a partnernship
Jun 25 15:11:29 <Letothe2nd> ayjay_t: well theres a reason why the original beagleboard with its 149$ price tag was considered disruptive
Jun 25 15:12:09 <Letothe2nd> once you're into runs that are not counted in 10ks of units, prices change quite a bit.
Jun 25 15:13:08 <Letothe2nd> and even for those runs, you have substantial ramp-up. especially if its a first design in that level of complexity.
Jun 25 15:13:58 <ayjay_t> yeah i can only imagine
Jun 25 15:14:12 <ayjay_t> the testing itself is a couple people
Jun 25 15:14:28 <ayjay_t> a SOM or SOM-like situation would be the best
Jun 25 15:14:40 <ayjay_t> driving down the cost of the dev-board if buying at quantity might be a stop gap
Jun 25 15:14:50 <ayjay_t> forming a partnership with another customer who needs this might be an option
Jun 25 15:15:15 <Letothe2nd> you have to think it the other way round. if you're in the league that can actually handle that technology, the 1.5k for the dev kit are peanuts.
Jun 25 15:15:50 <Letothe2nd> its not even 3days worth of dev time.
Jun 25 15:16:12 <Letothe2nd> on a project that might take man-months up to man-years. so who cares about 1.5k?
Jun 25 15:17:16 <ayjay_t> no no i'm talking about building the dev-kit into a product run
Jun 25 15:17:42 <Letothe2nd> .. which is almost certainly discouraged by the specs and ratings.
Jun 25 15:17:44 <ayjay_t> we don't need 5 9's uptime on the first 500 units
Jun 25 15:18:07 <ayjay_t> we really don't.
Jun 25 15:18:15 <ayjay_t> this is for microscopes
Jun 25 15:18:20 <Letothe2nd> so we're back to just what i said 15 mintues ago:
Jun 25 15:18:23 <Letothe2nd> 15:04 <@Letothe2nd> its enough for a PoC on the dev kit at most.
Jun 25 15:18:31 <Letothe2nd> just what you described
Jun 25 15:18:39 <ayjay_t> right but i still want to drive the cost of the dev kit down
Jun 25 15:18:54 <Letothe2nd> ayjay_t: well, call TI :)
Jun 25 15:18:59 <ayjay_t> haha no kidding, i mean i will
Jun 25 15:19:04 <ayjay_t> i wanted to talk to people i know i can trust first
Jun 25 15:19:07 <Letothe2nd> i'm not kidding either.
Jun 25 15:19:08 <ayjay_t> and i grew up on the beagle
Jun 25 15:20:17 <ayjay_t> i'm looking at this because i need to 64k 64k-point FFTs 512 times per image
Jun 25 15:20:26 <ayjay_t> just in case you were wondering
Jun 25 15:23:07 <Letothe2nd> i'm not wondering at all, people have funny and crazy ideas each second.
Jun 25 15:23:23 <ayjay_t> TI will also know if there is someone on the market that can do something inbetween the dev-kit and custom board solution
Jun 25 15:23:51 <Letothe2nd> all i'm saying is that if you think in 10k$, and if you have to muse about the dev kit being 1250$ or 1500$, then your project is doomed anyways.
Jun 25 15:24:09 <ayjay_t> no the project isn't doom lol
Jun 25 15:24:11 <ayjay_t> doomed*
Jun 25 15:24:24 <ayjay_t> the actual worse case scenario is we ship a 500MP microscope
Jun 25 15:24:25 <Letothe2nd> ayjay_t: i love to be proven wrong on such prediciton!
Jun 25 15:24:30 <ayjay_t> instead of a 5PG microscope
Jun 25 15:24:39 <ayjay_t> and have to 5 years for some of this technology to get cheaper and the market to mature
Jun 25 15:25:13 <ayjay_t> so i mean obviously i want to see explore just how far we can get with what he have today
Jun 25 15:25:33 <ayjay_t> we can do the 5GP images it just takes a couple hours per image
Jun 25 15:25:54 <ayjay_t> i just ordered nVidia's Xavier unit which is like this SBC with 32GB RAM and a massive GPU
Jun 25 15:25:59 <ayjay_t> it's $700 bucks
Jun 25 15:26:05 <Letothe2nd> like i said: my $.02, make of it what you deem best.
Jun 25 15:26:39 <ayjay_t> it sucks that there's no SoM
Jun 25 15:26:50 <ayjay_t> like i don't understand, the guys who make these chips don't spin them to test them?
Jun 25 15:27:04 <Letothe2nd> and as nvida probably ships 10k to 100k times the unit count of the dev kit you mentioned, thats no surprise @700$
Jun 25 15:28:12 <ayjay_t> yeah i mean, for sure, but their initial offering is a SOM
Jun 25 15:28:22 <ayjay_t> and that's a business decision that i think is really good
Jun 25 15:28:48 <ayjay_t> i don't think it's bad offering the chip too but these intermediate solutions are nice
Jun 25 15:29:13 <Letothe2nd> all i can say is, talk to your trustworthy TI sales rep.
Jun 25 15:29:55 <ayjay_t> as long as they dont try to sell me an overpriced calculator
Jun 25 15:36:39 <ayjay_t> it seems like the good market for these things is telcom
Jun 25 16:13:22 <av500> ayjay_t: I can sell you my TMS320C6678 for $750 :)
Jun 25 16:13:30 <av500> TMDSEVM6678
Jun 25 16:21:54 <ayjay_t> av500:  i mean, is it totally working?
Jun 25 16:22:21 <ayjay_t> i'm not totally memorized on these numbers yet, is that the same one i was looking at
Jun 25 16:22:27 <ayjay_t> seems like a good deal to me
Jun 25 16:25:30 <ayjay_t> lol that's not the same one i was looking at eh
Jun 25 16:26:46 <ayjay_t> 8 cores but no FFT coprocessor
Jun 25 16:40:31 <av500> ayjay_t: ok $500 :)
Jun 25 16:44:39 <zmatt> c6678 is a "shannon" device (keystone I), a pretty early c66x
Jun 25 16:45:25 <zmatt> the TCI6638K2K ("kelper") you were looking at is much more recent
Jun 25 16:45:29 <zmatt> *kepler
Jun 25 16:46:58 <zmatt> keystone II
Jun 25 16:48:09 <zmatt> also, that board is $500 new :P
Jun 25 16:48:15 <mru> hmm, that explains the "shannon" sticker on a power brick that's knocking around here
Jun 25 16:48:23 <mru> I'd forgotten what it came with
Jun 25 16:48:49 <zmatt> shannon is tci6608, c6678, c6674, c6672, c6671
Jun 25 16:58:20 <ayjay_t> so actually zmatt i'm realizing that some of my test inputs did produce conjugate symmetric transfer functions
Jun 25 16:58:31 <ayjay_t> and i've got in in my head that it might be a function of input or output size
Jun 25 17:02:25 <ayjay_t> ugh i still will have trouble removing this abs() function
Jun 25 17:08:25 <av500> $250, last price!
Jun 25 17:11:10 <ayjay_t> alright, $250, last price
Jun 25 17:11:13 <ayjay_t> i'll remember that
Jun 25 17:12:10 <zmatt> does that have silicon revision 1.0 or 2.0 ? :D
Jun 25 18:00:42 <ayjay_t> see this is what happens to rookies like me
Jun 25 18:00:49 <ayjay_t> its a whole tribe of sharks in this channel
Jun 25 18:00:56 <ayjay_t> just trying to take advatange of me
Jun 25 18:01:07 * ayjay_t looks at av500
Jun 25 18:02:28 <av500> ayjay_t: :)
Jun 25 18:02:43 <av500> I doubt its of any use to you even at $25
Jun 25 18:03:14 <av500> zmatt: it ran Linux on C66 once then I got bored
Jun 25 18:03:19 <av500> also it has a fan
Jun 25 18:59:54 <kremlin> e
Jun 25 19:24:13 <ayjay_t> lol it's been like a few years
Jun 25 19:24:15 <ayjay_t> and still,
Jun 25 19:24:15 <ayjay_t> e
Jun 25 21:03:23 * ayjay_t prays for positive real valued functions
Jun 26 00:55:46 <MattB0ne> sooooooo.........
Jun 26 00:56:45 <MattB0ne> zmatt: I am need of your wisdom
Jun 26 00:58:48 <set_> MattB0ne: Hello...what is the question?
Jun 26 00:58:59 <MattB0ne> PRU UART
Jun 26 00:59:03 <MattB0ne> you ever use that
Jun 26 00:59:21 <MattB0ne> I am using his py-uio tool
Jun 26 00:59:32 <MattB0ne> he was finishing it up
Jun 26 00:59:36 <MattB0ne> last night
Jun 26 00:59:39 <MattB0ne> need a read function
Jun 26 00:59:58 <set_> No. I have not used py-uio yet. Oh. You guys are attempting something from his library. Okay.
Jun 26 01:02:27 <set_> See, me, I am learning about temporarily changing PATH.
Jun 26 01:02:42 <set_> it is tricky.
Jun 26 01:04:32 <set_> it seems that using /etc/ld.so.conf is a better method for global changes.
Jun 26 02:47:40 <set_> is "cat /proc/ioports" supposed to show something on my end on the BBBW?
**** ENDING LOGGING AT Fri Jun 26 03:01:39 2020