**** BEGIN LOGGING AT Fri Oct 30 02:59:57 2020 Oct 30 07:12:44 hm... 6 pin uart not working on default images Oct 30 07:12:51 i thought it was enabled by default Oct 30 07:13:02 sorry, iot image Oct 30 07:34:44 nvm, looks like my ftdi breakout is backwards Oct 30 07:35:31 i guess there's no standard for that. given i have what looks like the average board. Oct 30 07:35:38 surprising. Oct 30 07:36:02 they can be in either orientation. even though the boards look identical. Oct 30 07:48:53 bz2S: normally pin 1 (gnd) is a black wire Oct 30 07:49:55 and marked with a white dot on the BBB Oct 30 07:50:32 i have an ftdi with 6 pin 0.1" female connector Oct 30 07:51:23 they have a standard pin order, though in principle you can remove the individual leads from the connector frame and rearrange them Oct 30 07:52:39 well, it's not pin order that is backwards... but Oct 30 07:52:56 if you look at the breakout, you'd assume they'd all have the ic on top Oct 30 07:53:20 and the pins the same. e.g. gnd is on left of ic with connector facing out... Oct 30 07:53:32 but ic can be either on top or bottom Oct 30 07:53:34 relative to connector Oct 30 07:53:52 dunno what you mean by "all" .. the standard "6-pin ftdi" is a cable, i.e. the IC is integrated into the (oversized) usb connector Oct 30 07:55:00 if you have a random breakout then I wouldn't assume anything about it at all but explicitly check Oct 30 07:55:06 https://dave.cheney.net/wp-content/uploads/2013/09/photo.jpg Oct 30 07:55:13 his breakout has ic on inside. but mine is on outside. Oct 30 07:55:17 both boards look similar. Oct 30 07:55:23 that's all. Oct 30 07:55:54 I mean, the pins are clearly labeled anyway Oct 30 08:02:44 it is important to check such things... you're lucky the serial console header of the BBB doesn't connect directly to the cpu (like the expansion headers) but goes through a fairly robust buffer IC, otherwise you would have destroyed the BBB Oct 30 13:02:35 5 minutes after my realization last evening, i realized that in my case the 1/x is always 1 and the division is completely unecessary Oct 30 13:35:45 Sweet! Oct 30 14:21:43 Konsgnx: I was already puzzled by why you'd need division (by a non-constant) Oct 30 14:23:37 in spline computation you need a fractional value for where you are between 2 points. I didn't notice that my table was 1D rather than 2d so with a constant step of 1, there is no relevance to the division, I could just mask out the integer part of the value and use the remainder. Oct 30 14:24:58 the first realization was that for a given table you could also pre-compute the inverse of each step size to get a value to multiply by rather than divide. Oct 30 14:28:30 the division is by the time resolution of your lookup table, which is therefore only done when computing the table itself Oct 30 14:29:52 yup, assuming x is time(what i was calling step). in my case, the step is just an index. Oct 30 14:31:04 yeah, with "time" I meant the argument to the function you're evaluating... phase would be a better word Oct 30 19:32:34 I got a jitter, scary joke... Oct 30 19:32:43 Forget it. Oct 30 19:34:10 For those of you who read my awful post yesterday. Sorry. I am having 'puter issues on my end. It is not MachineKit. Oct 30 20:22:44 * GenTooMan wondered what you were complaining about. Oct 30 20:53:44 dang it... Oct 31 00:12:17 set_ https://munk.org/typecast/2015/09/14/rounding-out-baby-wedge-week-all-the-brother-ep-series/ Oct 31 01:00:14 I just had my first steak night! Sign me up! Oh and about that link...off to look. Oct 31 01:00:32 What in the hell was that? Oct 31 01:02:34 That link is like a propanganda site w/ a small photo of a typecast. Oct 31 01:06:21 Tomorrow is Halloween. Boo! Oct 31 01:15:15 Just for you dwellers, that is a double entendre. Oct 31 01:15:25 oh boy i just bought a beagle-AI Oct 31 01:15:25 The "Boo!" part. Oct 31 01:15:41 Thank you for saving the day. Oct 31 01:17:57 GenTooMan just bugged me by sending me an odd link. Oct 31 01:32:22 i'm so excited for these DSPs Oct 31 01:35:23 DSPs, yes! Finally, someone will make a, well hopefully, post about DSP and the AI. Oct 31 01:37:40 ayjay_t: I personally quite like the architecture of these DSPs... haven't actually played with the C66x yet, but I have experience with its predecessor the C674x Oct 31 01:38:36 i wish i had your background zmatt Oct 31 01:38:49 because i am really going to be fumbling around in the dark Oct 31 01:38:52 we will call it an adventure Oct 31 01:40:40 definitely never before seen a cpu with that many branch delay slots ;) and its hardware-assisted software pipelined loop functionality is really cool Oct 31 02:27:47 One shot to make the MachineKit image work. That is all you get. Holmes. Oct 31 02:33:01 eMMC! Oct 31 02:35:36 It is almost 10:00 and it is going to be a late night. Scarey-Dairy! Oct 31 02:46:37 Anyway, I am most likely wrong. Forget it. Oct 31 02:46:38 Sheesh. Oct 31 02:54:41 FIVE MINUTES! **** ENDING LOGGING AT Sat Oct 31 02:59:57 2020