July 20, 2019 
< | 1 | 2 | 3 | 4 | 5 | 6 | 7 | 8 | 9 | 10 | 11 | 12 | 13 | 14 | 15 | 16 | 17 | 18 | 19 | 20 | 21 | 22 | 23 | 24 | 25 | 26 | 27 | 28 | 29 | 30 | 31 | >
July 20, 2019 [00:08:09] *** LunarJetman <LunarJetman!LunarJetma@176.248.197.112> has joined ##vulkan[00:10:06] *** glYoda <glYoda!~kristinam@c-73-25-27-206.hsd1.or.comcast.net> has joined ##vulkan[00:15:56] *** Deluxe <Deluxe!~Deluxe@212.4.150.151> has quit IRC (Read error: Connection reset by peer)[01:08:33] *** wildlander <wildlander!~wildlande@unaffiliated/wildlander> has joined ##vulkan[02:45:28] *** mofeing <mofeing!~mofeing@226.red-83-47-83.dynamicip.rima-tde.net> has joined ##vulkan[02:50:39] *** ratchetfreak <ratchetfreak!~ratchetfr@ptr-82s3g7l9r42569r41f4.18120a2.ip6.access.telenet.be> has quit IRC (Ping timeout: 252 seconds)[02:51:26] *** brimonk <brimonk!~brian@unaffiliated/brimonk> has joined ##vulkan[02:59:42] *** LunarJetman <LunarJetman!LunarJetma@176.248.197.112> has quit IRC (Ping timeout: 272 seconds)[03:22:26] *** mofeing <mofeing!~mofeing@226.red-83-47-83.dynamicip.rima-tde.net> has quit IRC (Quit: mofeing)[05:50:15] *** felipealmeida <felipealmeida!~user@191.191.10.198> has joined ##vulkan[05:51:25] <felipealmeida> hi, I'm enabling VK_KHR_push_descriptor extension when creating the device and later trying to vkGetDeviceProcAddr, but it returns NULL[05:51:46] <felipealmeida> if I use vkGetInstanceProcAddr, it doesn't return NULL, but calls NULL later in the dispatch[05:51:59] <felipealmeida> for "vkCmdPushDescriptorSetKHR"[05:52:03] <felipealmeida> am I doing something wrong?[05:52:14] <felipealmeida> I'm using linux and intel hardware[05:52:34] <felipealmeida> the VK_KHR_push_descriptor is shown if enumerated[05:53:24] <felipealmeida> I recompiled vulkan-icd-loader and it shows it calling disp->CmdPushDescriptorSetKHR in gdb and it shows that to be NULL in disp[07:02:29] *** ryp <ryp!ryp@gateway/vpn/privateinternetaccess/ryp> has joined ##vulkan[07:45:57] *** slime <slime!~slime73@blk-215-81-93.eastlink.ca> has quit IRC (Quit: This computer has gone to sleep)[08:56:41] *** glYoda <glYoda!~kristinam@c-73-25-27-206.hsd1.or.comcast.net> has quit IRC (Quit: glYoda)[08:57:26] *** bpmedley <bpmedley!~bpm@2600:1700:eb20:6050:f5d9:d657:90cf:3967> has quit IRC (Quit: Textual IRC Client: www.textualapp.com)[08:59:03] *** glYoda <glYoda!~kristinam@c-73-25-27-206.hsd1.or.comcast.net> has joined ##vulkan[09:55:21] *** ryp <ryp!ryp@gateway/vpn/privateinternetaccess/ryp> has quit IRC (Ping timeout: 268 seconds)[10:28:08] *** ryp <ryp!ryp@gateway/vpn/privateinternetaccess/ryp> has joined ##vulkan[10:40:46] *** ryp <ryp!ryp@gateway/vpn/privateinternetaccess/ryp> has quit IRC (Ping timeout: 272 seconds)[10:42:21] *** ryp <ryp!ryp@gateway/vpn/privateinternetaccess/ryp> has joined ##vulkan[10:55:34] *** ville <ville!~ville@212-149-214-47.bb.dnainternet.fi> has quit IRC (Quit:)[10:57:28] *** ville <ville!~ville@212-149-214-47.bb.dnainternet.fi> has joined ##vulkan[10:59:42] *** ImQ009 <ImQ009!~ImQ009@unaffiliated/imq009> has joined ##vulkan[11:06:38] *** ciaala <ciaala!~crypt@2a02:120b:2c1f:5f40:6ef0:49ff:feee:4777> has joined ##vulkan[11:12:28] *** ciaala <ciaala!~crypt@2a02:120b:2c1f:5f40:6ef0:49ff:feee:4777> has quit IRC (Quit: Konversation terminated!)[11:20:12] *** Deluxe <Deluxe!~Deluxe@212.4.150.151> has joined ##vulkan[11:25:59] *** ratchetfreak <ratchetfreak!~ratchetfr@ptr-82s3g7l9r42569r41f4.18120a2.ip6.access.telenet.be> has joined ##vulkan[11:58:13] *** wildlander <wildlander!~wildlande@unaffiliated/wildlander> has quit IRC (Quit: Konversation terminated!)[12:55:51] *** ryp <ryp!ryp@gateway/vpn/privateinternetaccess/ryp> has quit IRC (Ping timeout: 258 seconds)[13:20:01] *** penguin42 <penguin42!~dg@cpc109017-salf6-2-0-cust428.10-2.cable.virginm.net> has joined ##vulkan[13:39:58] *** Karyon <Karyon!~Karyon@unaffiliated/karyon> has quit IRC (Quit: ZNC - https://znc.in)[13:41:32] *** Karyon <Karyon!~Karyon@unaffiliated/karyon> has joined ##vulkan[13:47:12] *** ryp <ryp!ryp@gateway/vpn/privateinternetaccess/ryp> has joined ##vulkan[14:21:59] *** slime <slime!~slime73@24.215.81.93> has joined ##vulkan[15:09:15] *** ciaala <ciaala!~crypt@2a02:120b:2c1f:5f40:6ef0:49ff:feee:4777> has joined ##vulkan[15:42:17] *** derhass <derhass!~derhass@dslb-188-102-016-172.188.102.pools.vodafone-ip.de> has joined ##vulkan[15:48:13] *** ryp <ryp!ryp@gateway/vpn/privateinternetaccess/ryp> has quit IRC (Ping timeout: 246 seconds)[16:12:09] *** ryp <ryp!ryp@gateway/vpn/privateinternetaccess/ryp> has joined ##vulkan[16:59:16] *** ryp <ryp!ryp@gateway/vpn/privateinternetaccess/ryp> has quit IRC (Ping timeout: 246 seconds)[17:04:08] *** mofeing <mofeing!~mofeing@226.red-83-47-83.dynamicip.rima-tde.net> has joined ##vulkan[17:32:13] *** reduz <reduz!sid223967@godotengine/reduz> has joined ##vulkan[17:33:30] <reduz> I have a problem with the validation layer.. it reports an error that does not seem to exist (says a descriptor is not bound when to me it seems clearly bound) and then crashes.. but when I run without the validation layer, everything is fine[17:34:00] <reduz> making a simple test case is difficult, so I'm wondering whether there is a way to record and reproduce the vulkan commands to submit a bug report[17:45:17] *** mofeing <mofeing!~mofeing@226.red-83-47-83.dynamicip.rima-tde.net> has left ##vulkan[17:58:52] *** ryp <ryp!ryp@gateway/vpn/privateinternetaccess/ryp> has joined ##vulkan[18:05:24] <reduz> there seems to be a vktrace, but unfortunately I can't compile it on linux, seems kind of bitrot[18:12:49] *** karroffel <karroffel!~karroffel@ip-178-203-190-17.hsi10.unitymediagroup.de> has quit IRC (Read error: Connection reset by peer)[18:13:04] *** karroffel8 <karroffel8!~karroffel@ip-178-203-190-17.hsi10.unitymediagroup.de> has joined ##vulkan[18:56:52] *** karroffel8 <karroffel8!~karroffel@ip-178-203-190-17.hsi10.unitymediagroup.de> has quit IRC (Read error: Connection reset by peer)[18:57:12] *** karroffel81 <karroffel81!~karroffel@ip-178-203-190-17.hsi10.unitymediagroup.de> has joined ##vulkan[18:58:38] <felipealmeida> how do I debug vulkan applications beside validation layer? I'm recording a buffer and then submitting later[18:58:43] <felipealmeida> however, nothing appears in the screen[18:59:02] <felipealmeida> no validation errors[18:59:57] <Yatekii> felipealmeida: did you try renderdoc?[19:00:12] <felipealmeida> nope, thanks[19:00:58] <felipealmeida> I'll try that[19:06:53] <reduz> well I just have no way to report a bug in the validation layers i guess[19:07:38] <reduz> compiling vktrace wants me to replace my system headers for custom built ones, even if i try to build the same version I have installed in the system[19:27:35] *** ColaEuphoria_ <ColaEuphoria_!~ColaEupho@unaffiliated/colaeuphoria> has quit IRC (Remote host closed the connection)[19:31:34] <Yatekii> hmm how tightly are index buffers and vertex buffers coupled?[19:31:56] <Yatekii> because when i change the vertex layout, hell breaks loose[19:33:15] <Yatekii> my indices are suddenly "random" and extremely high, even tho on the CPU all the indices are right[20:00:41] <mefesto> Yatekii: did you verify that the indices in the gpu buffer were uploaded correctly? (renderdoc)[20:01:35] <Yatekii> mefesto: well in renderdoc they are completely wrong[20:01:43] <Yatekii> on the cpu they are right[20:01:56] <Yatekii> the question is why it doesnt work anymore[20:02:10] <Yatekii> before i changed the vertex layout everything worked[20:02:12] <mefesto> do you have any code to share?[20:02:18] <Yatekii> i don't understand[20:02:51] <Yatekii> yes i do, but i don't think it helps :/ its rust and wgpu-rs ;) no c/plain vulkan[20:02:53] <mefesto> its helpful to show the code in question so that other people can see what it is you are attempting to do[20:03:20] <Yatekii> mefesto: the point is that i didn't change anything about the indexbuffers[20:03:43] <Yatekii> so it must be the vertex buffers influencing the index buffers[20:03:52] <Yatekii> for whatever dumb reason[20:03:58] <mefesto> no i dn't think it "must be" that :)[20:04:50] <Yatekii> yeah well since i didn't change anything else it con't be anything else logically[20:05:05] <Yatekii> *can't[20:05:34] <Yatekii> i mean i can show you the code and the diffs. you wont find any changes other than the vertex layout :/[20:06:05] <Yatekii> i can commit it and show it in a bit. i only have the old layout on GH atm[20:06:14] <mefesto> whatever you feel comfortable with. typically the more information provided the more likely someone in here will see what the issue is[20:06:21] <Yatekii> yeah[20:07:34] <Yatekii> also, folks, how easy should 10kk tris be on an iGPU?[20:07:46] <Yatekii> *not tris, vertices[20:08:03] <Yatekii> because atm the iGPU kinda dies. my shader is plain simple[20:08:26] <Yatekii> but a mate of mine says 10kk tris are a piece of cake for the iGPU[20:11:11] <mefesto> verts in the same buffer? how many draw calls? how often does the pipeline switch? 10k verts sound like cake... but there are plenty of details we don't know[20:14:51] <Yatekii> mefesto: no, 10KK Xd 10000000[20:15:04] <Yatekii> i will tell you more details later[20:21:09] <Yatekii> mefesto: i have 20 renderpasses all with the same pipeline. i have the different passes only to clear the stencil maybe i can make this more intelligent[20:21:30] <Yatekii> i have 10million vertices[20:22:12] <mefesto> "different passes only to clear the stencil"?[20:22:28] <Yatekii> i do 13k drawcalls (that's like 200us more than 500 drawcalls, both around 6.5ms total)[20:23:15] <Yatekii> well by that i mean i could make it one renderpass theoretically but then clearing the stencil would not be possible[20:23:40] <Yatekii> (i just figured it's possible i think)[20:24:14] <Yatekii> if you'd like i can share code/a build if youd be in the mood to have a look[20:24:46] <Yatekii> the problem is: i have looked many hours into it but in the end i don't really know what to look for[20:24:59] <mefesto> i'll skim some code but i won't be running anything :)[20:39:04] *** ryp <ryp!ryp@gateway/vpn/privateinternetaccess/ryp> has quit IRC (Ping timeout: 246 seconds)[21:04:56] <Yatekii> mefesto: ah sure, I appreciate any kind of help ^^ I just thought to see the GPU pipeline state could help, hence t6he offer to hand out a build :)[21:05:15] <Yatekii> (I could share a renderdoc trace but replaying might not work on different GPUs)[21:06:31] <Yatekii> mefesto: https://github.com/Yatekii/sailor/tree/new-vertex-format here you go :)[21:07:52] <Yatekii> pipeline setup is here: https://github.com/Yatekii/sailor/blob/new-vertex-format/src/drawing/painter.rs#L340 https://github.com/Yatekii/sailor/blob/new-vertex-format/config/shader.vert shader is here :)[21:15:19] <felipealmeida> is there anyway to reuse a command buffer already recorded to draw just some part?[21:15:25] <felipealmeida> like clipping?[21:21:52] <mefesto> Yatekii: I'm not sure how current these docs are for wgpu-native but it looks like (on line 88) that they are defaulting to u16 index types, whereas your code is using u32: https://docs.rs/wgpu-native/0.2.1/src/wgpu_native/command/render.rs.html#70-94[21:28:08] <Yatekii> mefesto: interesting. let me dig some[21:28:16] <Yatekii> because I was using 32 before without an issue[21:28:23] <Yatekii> so I wonder why it's an issue now ^ ^[21:28:28] <Yatekii> *32bit[21:28:49] <mefesto> seems like on master branch you can configure it via the pass.index_state.format value: https://github.com/gfx-rs/wgpu/blob/master/wgpu-native/src/command/render.rs#L310[21:29:40] <mefesto> good luck. gtg[21:32:03] <Yatekii> mefesto: https://github.com/gfx-rs/wgpu/blob/183058d618e8bffd36e061039663a733be7b0b85/wgpu-native/src/command/render.rs#L310 yep upstream it's fixed[21:32:09] <Yatekii> mefesto: thanks a lot :) have fun![21:32:10] <Yatekii> cya[21:58:44] *** RicardasSim <RicardasSim!~RicardasS@78-63-172-44.static.zebra.lt> has joined ##vulkan[22:09:48] *** ciaala <ciaala!~crypt@2a02:120b:2c1f:5f40:6ef0:49ff:feee:4777> has quit IRC (Quit: Konversation terminated!)[22:13:56] *** LunarJetman <LunarJetman!LunarJetma@176.248.197.112> has joined ##vulkan[22:54:46] *** ImQ009 <ImQ009!~ImQ009@unaffiliated/imq009> has quit IRC (Quit: Leaving)