01:52:07 | * | derpydoo joined #nim |
02:34:30 | FromDiscord | <Robyn [She/Her]> I am making a parser for the WIT file format and calling it wit's end, and no-one can stop me |
02:35:33 | FromDiscord | <Elegantbeef> @Robyn [She/Her] how is it not called witless and self deprecating |
02:37:09 | FromDiscord | <Robyn [She/Her]> In reply to @Elegantbeef "<@524288464422830095> how is it": Beef, you are an absolute genius |
02:37:17 | FromDiscord | <Robyn [She/Her]> I am now naming it that instead |
02:40:21 | FromDiscord | <Robyn [She/Her]> nested comments, wtf... |
03:02:02 | * | derpydoo quit (Quit: derpydoo) |
03:36:47 | * | nils` quit (Ping timeout: 272 seconds) |
04:11:19 | * | SchweinDeBurg quit (Quit: WeeChat 4.7.0-dev) |
04:24:02 | * | cornfeedhobo quit (Quit: ZNC - https://znc.in) |
04:32:47 | * | cornfeedhobo joined #nim |
04:52:56 | * | SchweinDeBurg joined #nim |
06:45:37 | * | ntat joined #nim |
06:57:33 | * | coldfeet joined #nim |
07:59:17 | * | andy-turner joined #nim |
08:27:17 | * | derpydoo joined #nim |
08:49:09 | * | nils` joined #nim |
09:25:44 | * | SchweinDeBurg quit (Ping timeout: 272 seconds) |
09:27:09 | * | SchweinDeBurg joined #nim |
09:45:24 | * | SchweinDeBurg quit (Ping timeout: 245 seconds) |
09:47:22 | * | SchweinDeBurg joined #nim |
10:24:09 | FromDiscord | <tauruuuuuus> Beef solving the most difficult problem in programming |
10:24:27 | FromDiscord | <tauruuuuuus> Ever thought of doing naming as a service? |
11:18:07 | * | beholders_eye joined #nim |
11:20:28 | * | SchweinDeBurg quit (Ping timeout: 252 seconds) |
11:22:13 | * | SchweinDeBurg joined #nim |
11:40:54 | FromDiscord | <heysokam> NAAS could be easily confused with "nagging as a service"↵but maybe he could do both at once, if he is resourceful.. hmmm |
11:59:09 | * | beholders_eye quit (Ping timeout: 245 seconds) |
12:04:44 | * | beholders_eye joined #nim |
12:05:37 | * | derpydoo quit (Ping timeout: 252 seconds) |
12:08:30 | * | nils` quit (Ping timeout: 246 seconds) |
12:10:14 | * | void09_ quit () |
12:11:01 | * | void09 joined #nim |
13:11:16 | * | beholders_eye quit (Ping timeout: 252 seconds) |
13:18:08 | * | andy-turner quit (Quit: Leaving) |
13:18:25 | * | andy-turner joined #nim |
13:30:40 | * | beholders_eye joined #nim |
14:13:10 | * | alexdaguy joined #nim |
14:54:45 | FromDiscord | <pmunch> In reply to @battery.acid.bubblegum "I am making a": WIT file format? |
14:55:30 | FromDiscord | <pmunch> I'm working on my epaper x11 driver thingy, currently implementing dithering as a shader |
14:55:55 | FromDiscord | <Robyn [She/Her]> In reply to @pmunch "WIT file format?": Yeah for WebAssembly's component model stuff |
14:56:33 | FromDiscord | <pmunch> Aah |
14:59:05 | * | alexdaguy quit (Quit: a) |
15:27:38 | FromDiscord | <pmunch> Hmm, so there was an update to the shady library 5 months ago, but no new version was drawn. It simultaneously has stuff and need and breaks things |
15:44:43 | * | rockcavera joined #nim |
16:04:18 | FromDiscord | <linuxbro> What's the recommended library for cryptographic operations in nim? |
16:04:48 | FromDiscord | <linuxbro> Trying to do AES-GCM or ChaCha20-Poly1305 |
16:08:24 | * | nils` joined #nim |
16:10:27 | FromDiscord | <_nenc> try nimcrypto |
16:17:33 | FromDiscord | <linuxbro> That's what I was looking at. Just wasn't sure if it was the "standard" |
16:18:13 | FromDiscord | <linuxbro> No chachapoly20, but it's got rihndael, so gtg |
16:18:20 | FromDiscord | <linuxbro> (edit) "rihndael," => "rijndael," |
16:49:22 | FromDiscord | <blackmius> In reply to @linuxbro "What's the recommended library": you can try extend openssl bindings i think |
16:58:41 | FromDiscord | <linuxbro> Will that still result in static binaries? |
16:59:53 | FromDiscord | <linuxbro> Ah, looks like you can force a static link |
17:02:38 | * | nils` quit (Ping timeout: 252 seconds) |
17:09:24 | * | coldfeet quit (Quit: Lost terminal) |
17:10:50 | FromDiscord | <blackmius> sent a code paste, see https://play.nim-lang.org/#pasty=PrIRLMVK |
17:11:26 | FromDiscord | <blackmius> (edit) "https://play.nim-lang.org/#pasty=EDLwnATK" => "https://play.nim-lang.org/#pasty=EJWdKZcL" |
17:12:38 | FromDiscord | <linuxbro> Perfect, this should help me in my quest to chacha real smooth |
17:13:02 | FromDiscord | <linuxbro> I definitely didn't pick Chacha20 over other options because it has a funny name |
17:58:04 | * | nils` joined #nim |
18:01:03 | * | beholders_eye quit (Ping timeout: 244 seconds) |
18:51:20 | * | beholders_eye joined #nim |
19:10:08 | FromDiscord | <linuxbro> I don't use VSCode all that much. How do I configure it to build and run a file when I click the "run" triangle in the top right? It only builds it currently |
19:10:23 | FromDiscord | <linuxbro> I'm using the official nim-lang.org extension |
19:13:24 | * | beholders_eye quit (Ping timeout: 246 seconds) |
20:12:20 | * | beholders_eye joined #nim |
20:26:26 | FromDiscord | <pmunch> Hmm, been a while since I've seen @treeform around. Is it just time zone difference and me getting healthier day/night cycles? |
20:27:05 | FromDiscord | <treeform> I just been busy with traveling and moving states and chaning jobs... |
20:27:09 | FromDiscord | <treeform> some times life gets in the way |
20:27:16 | FromDiscord | <pmunch> Oh damn, lots of big changes! |
20:27:35 | FromDiscord | <pmunch> What're you up to now? |
20:27:44 | FromDiscord | <treeform> AI what else? |
20:27:56 | FromDiscord | <pmunch> Haha, fair enough |
20:28:14 | FromDiscord | <pmunch> Using it for what though? We can't just all go around saying we work with AI |
20:28:38 | FromDiscord | <treeform> I was doing applied AI, now switched to theoretical research. |
20:29:00 | FromDiscord | <pmunch> Oh cool, like actually working on AI, not just using AI for something |
20:29:25 | FromDiscord | <pmunch> New employer, or just different position? |
20:29:31 | FromDiscord | <treeform> manly multi agent systems and alignment |
20:29:37 | FromDiscord | <treeform> new employer |
20:30:40 | FromDiscord | <treeform> To bad the AI stuff is so heavy into python and cuda. |
20:30:49 | FromDiscord | <treeform> I wish I could use Nim more |
20:32:00 | FromDiscord | <pmunch> Yeah that'd be really cool |
20:32:15 | FromDiscord | <pmunch> Our resident AI guy is the only one not coding in Nim |
20:32:20 | FromDiscord | <pmunch> Granted there's only three of us so far |
20:32:31 | FromDiscord | <treeform> What is your AI project? |
20:34:07 | FromDiscord | <pmunch> Well, we do power line monitoring, and part of that is figuring out how much power can go through a line while staying within thermal limit. Our units have a lot of sensors and we have a couple different AI projects. Some working with image data trying to detect various things. And then we are also trying to predict conditions into the future so that power deliveries can be planned better. |
20:36:13 | FromDiscord | <treeform> That's pretty cool. The the Ai is the only part not in Nim? Impressive. |
20:37:07 | FromDiscord | <pmunch> Well all the stuff that we write is in Nim, from firmware to most of our data processing. But we use some off the shelf stuff as well |
20:45:49 | FromDiscord | <pmunch> Quick question about Shady by the way, have you tried any more complex examples of compute shaders? |
20:47:06 | FromDiscord | <pmunch> Trying to get some dithering to work for my epaper laptop project. It works on the CPU, but not on the GPU. So I'm guessing that I don't set up the buffers correctly |
20:52:26 | FromDiscord | <fabric.input_output> sent a code paste, see https://play.nim-lang.org/#pasty=VABDbYmu |
20:53:16 | FromDiscord | <fabric.input_output> because it sure is weird as hell |
20:53:48 | FromDiscord | <fabric.input_output> when I assign `Foo[int](10)` to `a` it outputs `ptr 10` |
20:53:50 | FromDiscord | <fabric.input_output> ?? |
20:53:52 | FromDiscord | <pmunch> Hmm, that is indeed weird |
20:54:36 | FromDiscord | <pmunch> 2.2.2? |
20:55:03 | FromDiscord | <fabric.input_output> tried with latest on the playground |
20:55:15 | FromDiscord | <fabric.input_output> oh wait `ptr X` X is the value |
20:56:34 | FromDiscord | <pmunch> Seems to be a change between 1.6.20 and 2.0.0 |
20:57:08 | FromDiscord | <pmunch> 1.6.20 gives a pointer to the value 0 |
20:57:34 | FromDiscord | <pmunch> Oh wait, it's just the repr of pointers which changed |
21:02:00 | * | ntat quit (Quit: leaving) |
21:02:33 | * | ntat joined #nim |
21:02:47 | * | ntat quit (Client Quit) |
21:10:08 | FromDiscord | <fabric.input_output> man why isn't opengl working with nim opengl |
21:10:15 | FromDiscord | <fabric.input_output> I tried it in C++ it worked |
21:10:25 | FromDiscord | <fabric.input_output> (edit) "opengl" => "glfw" |
21:12:01 | FromDiscord | <fabric.input_output> the functions aren't being loaded or something |
21:28:54 | FromDiscord | <pmunch> Linking issue? |
21:33:11 | FromDiscord | <treeform> In reply to @pmunch "Quick question about Shady": No but my friend manage to get Nim and Cuda/HIP working together: https://github.com/monofuel/hippo |
21:41:51 | FromDiscord | <pmunch> Well I'm not trying to do anything fancy, just to dither something |
21:45:28 | FromDiscord | <pmunch> I've got a pattern and an image, and then I basically need to sample every pixel in the image, compare it to values from the pattern and decide if the output pixel is black or white |
21:48:28 | FromDiscord | <pmunch> That part works on the CPU, but not as a shader, so I'm guessing I just need to set up the pattern and screen buffer correctly |
21:50:07 | FromDiscord | <pmunch> Hmm, either that or texelFetch using a Sampler2D doesn't work correctly.. |
21:54:59 | FromDiscord | <treeform> I am not sure. Shaders can be quite tricky. |
21:55:20 | FromDiscord | <treeform> I never used shady in production so I am not as confident in it. |
22:15:16 | FromDiscord | <pmunch> Hmm, yeah shaders can be a bit of a pain.. |
22:16:17 | FromDiscord | <pmunch> I had hoped Shady would make it a bit easier.. |
22:29:44 | * | andy-turner quit (Quit: Leaving) |
22:58:42 | FromDiscord | <Elegantbeef> @pmunch still have an issue with a shader? |
23:03:09 | FromDiscord | <pmunch> Yup |
23:03:49 | FromDiscord | <pmunch> I just can't seem to get this working |
23:05:47 | FromDiscord | <pmunch> @ElegantBeef |
23:06:32 | FromDiscord | <Elegantbeef> What's currently happening? |
23:08:56 | FromDiscord | <Elegantbeef> I did read what you said 😄 |
23:13:30 | FromDiscord | <pmunch> Currently I just get a black result |
23:15:33 | FromDiscord | <pmunch> What I'm trying to do is a simple pattern dither, compare a pixel in the input with a pixel in the pattern (position mod pattern size). If the pixel in the input is darker than the pixel in the pattern then the output pixel should be black, otherwise it should be white. |
23:16:06 | FromDiscord | <pmunch> I've modified the compute1.nim example from shady. And it works when I run it on the CPU, but the GPU just gives a black output |
23:16:37 | FromDiscord | <pmunch> The mess I've got right now: https://pasty.ee/RzALiazD |
23:17:00 | FromDiscord | <Elegantbeef> Atleast it's self contained |
23:17:15 | FromDiscord | <pmunch> Yeah it's a fairly simple thing |
23:17:22 | FromDiscord | <pmunch> Which just makes it extra frustrating |
23:17:30 | FromDiscord | <pmunch> 90% of that code is just the compute example |
23:18:17 | FromDiscord | <pmunch> I'm tempted to revert to SamplerBuffer and just load all the pixels as one long buffer and do the indexing myself |
23:19:31 | FromDiscord | <Elegantbeef> `glUniform1i(patternBufferLoc, 1)`..? |
23:20:35 | FromDiscord | <pmunch> Again, mostly copied from the example 😅 If I've understood this correctly that binds `patternDimension` in the shader to the 1st buffer |
23:20:43 | FromDiscord | <Elegantbeef> No |
23:20:51 | FromDiscord | <pmunch> Ah.. |
23:21:06 | * | om3ga quit (Ping timeout: 272 seconds) |
23:21:13 | FromDiscord | <Elegantbeef> that sets the value of `patternBufferLoc` to be `1` |
23:21:48 | FromDiscord | <Elegantbeef> that should be your `patternTextureId` |
23:21:59 | FromDiscord | <pmunch> Huh |
23:22:37 | FromDiscord | <Elegantbeef> Actually wait no it should be `glBindTextureUnit` followed by `glUinform` |
23:23:18 | FromDiscord | <pmunch> glBindTextureUnit? |
23:23:53 | FromDiscord | <pmunch> Here is where I got the idea from by the way: https://github.com/treeform/shady/blob/master/examples/compute1.nim#L99 |
23:24:42 | FromDiscord | <Elegantbeef> The `1` seems off to me |
23:25:38 | * | om3ga joined #nim |
23:25:53 | FromDiscord | <pmunch> glBindTextureUnit seems to only be available in 4.5, this code is using 4.3. Not that it matters terribly much |
23:26:01 | FromDiscord | <Elegantbeef> sent a code paste, see https://play.nim-lang.org/#pasty=VlNqCQBR |
23:26:26 | FromDiscord | <pmunch> I think it's the same 1 as the one in GL_TEXTURE1 |
23:26:26 | FromDiscord | <Elegantbeef> Ah perhaps in older opengl that's fine, I tend to stick to 4.5+ features 😄 |
23:27:06 | FromDiscord | <pmunch> Probably a good idea |
23:27:18 | FromDiscord | <Elegantbeef> Ignore my picking of that code then, quite close to what I recall but yea perhaps not the issue |
23:27:27 | FromDiscord | <Elegantbeef> I cannot get this to build locally so cannot do much investigating |
23:29:17 | FromDiscord | <pmunch> Oh shit, yeah I had to much about with shady and patch a thing in Windy to get this to compile.. |
23:29:27 | FromDiscord | <pmunch> Probably not the best start come to think of it |
23:29:48 | FromDiscord | <pmunch> But that was just to build the compute1.nim |
23:30:13 | FromDiscord | <Elegantbeef> One day I'll expand my framework to make compute shaders usable 😄 |
23:31:17 | FromDiscord | <Elegantbeef> I would suggest hooking up the opengl debug proc to see if anything stands out |
23:31:46 | FromDiscord | <Elegantbeef> https://github.com/beef331/truss3d/blob/master/src/truss3D.nim#L65-L91 if you want to steal parts of mine |
23:33:03 | FromDiscord | <pmunch> Ooh, some debugging output would indeed be nice |
23:33:30 | FromDiscord | <Elegantbeef> line 109 and 110 are also important |
23:36:09 | * | m5zs7k quit (Ping timeout: 244 seconds) |
23:40:18 | * | m5zs7k joined #nim |
23:47:26 | FromDiscord | <pmunch> https://pasty.ee/FJcWvzwQ |
23:47:30 | FromDiscord | <pmunch> Hmm, nothing too interesting |
23:48:03 | FromDiscord | <Elegantbeef> Well that atleast means you aren't holding the API wrong and the shader is just somehow wrong |
23:48:35 | FromDiscord | <pmunch> Indeed |
23:51:01 | FromDiscord | <Elegantbeef> Don't know if renderdoc works with compute, but you can always look at that |