| 00:01.22 | ``Erik | I would recommend procuring a non-linux machine... be it private procurement, dual boot on existing hw, or convincing the company to supply you... |
| 00:01.56 | ``Erik | given your interest in shark, I'd ask the company about shipping you a g[45] based mac... |
| 00:02.01 | Maloeran | FreeBSD then? That'll install on the desktop |
| 00:02.06 | Maloeran | Or OSX *shivers* |
| 00:02.43 | ``Erik | in your dicussions with the company... I use a dual g5 osX tiger tower as my daily machine... |
| 00:03.02 | ``Erik | and I'm the one on the hook to tell arl "aye" or "nay" on your progress |
| 00:03.21 | ``Erik | so with that argument, you might be able to get a free dual g5 tower out of the company... |
| 00:03.22 | ``Erik | :) |
| 00:03.43 | Maloeran | Understood :) |
| 00:04.09 | ``Erik | heh, don't be scared of me :) |
| 00:04.26 | Maloeran | I don't really see how G4/G5 hardware could be a significant target though, I understand about the OSX part |
| 00:04.27 | ``Erik | I mean, I'm installing X on my dual opteron... :) |
| 00:04.48 | ``Erik | erm |
| 00:04.50 | ``Erik | quad, rather |
| 00:05.02 | ``Erik | I'm just noting what I have on my desk |
| 00:05.13 | ``Erik | and, uh, it happens to have better support for profiling... O:-) |
| 00:06.32 | ``Erik | (also bear in mind... company $'s are on a far different scale than you&me dollars...) |
| 00:06.34 | Maloeran | True, I needed a full memory manager anyway. I'm presently hesitating about the way to priotirize lower blocks in order to allow blocks to be freed when memory use drops |
| 00:07.31 | ``Erik | personally, I believe that: the greater variety of hw and os you have available; the better software you are able to produce. |
| 00:08.09 | ``Erik | if you want, either lee or myself will talk to the company to ensure that you are adequately equiped |
| 00:08.50 | Maloeran | Thanks but don't bother them. The pay is quite enough, once I receive it anyway |
| 00:09.17 | ``Erik | ... |
| 00:09.23 | ``Erik | dude, they could send you half a dozen machines without blinking... |
| 00:12.15 | Maloeran | I don't need that though. I would be very satisfied with just the pay to acquire what I need, really |
| 00:14.08 | Maloeran | Any thoughts on priotirizing memory blocks to allow some to be freed? I just aimed for a "best fit" allocation, but memory use then obviously remains high after use |
| 00:14.34 | ``Erik | it's a topic of "OS" research... |
| 00:14.53 | ``Erik | a guess at lru or mru would probably be "adequate" |
| 00:21.51 | Maloeran | realloc() does not garantee to preserve data in place if you shrink, which I often found annoying ; fortunately that one does |
| 01:01.13 | ``Erik | realloc doesn't guarantee to preserve data in place period. |
| 01:32.47 | *** join/#brlcad digitalfredy (n=digitalf@200.71.62.161) | |
| 01:39.46 | *** join/#brlcad danfalck (n=dan@pool-71-111-76-8.ptldor.dsl-w.verizon.net) | |
| 03:47.02 | Twingy | hi dan |
| 03:51.54 | danfalck | Twingy: hi |
| 03:52.04 | Twingy | oh, hi dan |
| 03:52.14 | Twingy | I just finished bolt holes |
| 03:52.17 | danfalck | cool |
| 03:52.30 | Twingy | I am going to add Duplicate on Sunday |
| 03:52.32 | danfalck | I did a bolt hole calculator in python for my stuff here |
| 03:52.41 | Twingy | and Rotation angle into Sketch |
| 03:52.43 | danfalck | tkinter gui |
| 03:53.05 | Twingy | Bolt Hole goes in a Sketch |
| 03:53.19 | Twingy | so you can have your extrusion applied to each hole |
| 03:53.26 | Twingy | i.e. give each on a taper |
| 03:53.55 | Twingy | or a recess to do flush mounting |
| 03:54.05 | danfalck | can you output drilling canned cycles ie G81 |
| 03:54.16 | Twingy | is that pecking? |
| 03:54.25 | danfalck | just straight drilling |
| 03:54.36 | danfalck | pecking is G82 |
| 03:54.54 | Twingy | erm |
| 03:54.59 | Twingy | not sure what 81 does for you |
| 03:55.15 | danfalck | if you have big machine, it will do it |
| 03:55.16 | Twingy | specify hole locations for drilling |
| 03:55.23 | Twingy | do what? |
| 03:55.38 | Twingy | grab a drill bit? |
| 03:55.45 | danfalck | drill without pecking, depends on material, coolant etc |
| 03:55.55 | Twingy | I just mill the holes |
| 03:56.56 | Twingy | welp, got to get to bed for solar installation tomorrow |
| 03:57.03 | Twingy | night |
| 03:57.05 | danfalck | talk to you later |
| 03:57.07 | danfalck | have fun |
| 15:43.56 | Maloeran | Hey Erik, please share the results if you try the new prep on OSX, I would be interested to know if the memory manager fixed the slowness |
| 16:01.06 | Maloeran | It's 3.5 seconds on the M1 here, I should be able to get that to about 2 seconds |
| 17:21.36 | *** join/#brlcad DTRemenak (n=DTRemena@adsl-68-126-0-210.dsl.irvnca.pacbell.net) | |
| 19:29.49 | ``Erik | hum, on my 1.25 g4 lappie, the prep took 36.6 seconds with the latest code, mal |
| 19:56.33 | Maloeran | Gah. |
| 19:56.57 | Maloeran | And that's on the M1 model? *grumbles* |
| 19:58.28 | Maloeran | Was it an improvement in any way at least?... Do you think you can put up any new figures on where the time is being spent? This is puzzling |
| 20:02.41 | Maloeran | I guess the assumption of an overzealous free() that releases pages constantly, just to allocate them again which require zero'ing them, was wrong |
| 20:03.58 | Maloeran | My 1.8ghz Sempron takes 5 seconds |
| 20:18.43 | *** join/#brlcad digitalfredy (n=digitalf@200.71.62.161) | |