July 14, 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 14, 2019 [00:20:04] *** ryp <ryp!ryp@gateway/vpn/privateinternetaccess/ryp> has joined ##vulkan[00:33:22] *** ryp <ryp!ryp@gateway/vpn/privateinternetaccess/ryp> has quit IRC (Ping timeout: 245 seconds)[01:58:25] *** Guy1524 <Guy1524!6c2cd738@pool-108-44-215-56.clppva.fios.verizon.net> has joined ##vulkan[01:58:43] <Guy1524> Hello all, I am facing an issue where my code works w/ the validation layers on, but crashes inside the driver when it's off[01:58:52] <Guy1524> and the validation layers don't give me any messages[01:59:03] <Guy1524> do any of you have an idea for why this could happen[02:09:03] *** ratchetfreak <ratchetfreak!~ratchetfr@ptr-82s3g7nyo2zjr8pdfah.18120a2.ip6.access.telenet.be> has quit IRC (Ping timeout: 250 seconds)[02:21:50] *** Deluxe <Deluxe!~Deluxe@212.4.150.151> has quit IRC (Remote host closed the connection)[04:47:36] *** Snoras <Snoras!~Snoras@gateway/tor-sasl/snoras> has quit IRC (Ping timeout: 260 seconds)[04:49:36] *** penguin42 <penguin42!~dg@cpc109017-salf6-2-0-cust428.10-2.cable.virginm.net> has quit IRC (Remote host closed the connection)[04:49:50] *** Snoras <Snoras!~Snoras@gateway/tor-sasl/snoras> has joined ##vulkan[04:57:49] *** DrGonzo <DrGonzo!~DrGonzo@modemcable080.164-57-74.mc.videotron.ca> has quit IRC (Read error: Connection reset by peer)[05:07:21] *** tristianc <tristianc!~tristianc@pool-72-83-128-193.washdc.fios.verizon.net> has quit IRC (Quit: Konversation terminated!)[05:37:58] *** Guy1524 <Guy1524!6c2cd738@pool-108-44-215-56.clppva.fios.verizon.net> has quit IRC (Remote host closed the connection)[06:58:11] *** xieta <xieta!xieta@unaffiliated/xieta> has quit IRC (Quit: ZNC - https://znc.in)[08:03:35] *** ryp <ryp!ryp@gateway/vpn/privateinternetaccess/ryp> has joined ##vulkan[08:40:48] *** ville <ville!~ville@212-149-214-47.bb.dnainternet.fi> has joined ##vulkan[08:55:25] *** R2robot <R2robot!~R2robot@unaffiliated/r2ro> has quit IRC (Ping timeout: 268 seconds)[09:20:31] *** ipe <ipe!~ipe@91-156-247-55.elisa-laajakaista.fi> has joined ##vulkan[09:38:37] *** RicardasSim <RicardasSim!~RicardasS@78-63-172-44.static.zebra.lt> has quit IRC (Quit: Leaving)[09:51:18] *** Andrej1 <Andrej1!~Andrej@BSN-143-212-5.dynamic.siol.net> has quit IRC (Read error: Connection reset by peer)[11:12:40] *** R2robot <R2robot!~R2robot@unaffiliated/r2ro> has joined ##vulkan[11:25:01] *** Deluxe <Deluxe!~Deluxe@212.4.150.151> has joined ##vulkan[12:21:58] *** pa <pa!~pa@unaffiliated/pa> has quit IRC (Ping timeout: 246 seconds)[12:24:09] *** pa <pa!~pa@unaffiliated/pa> has joined ##vulkan[12:38:56] *** ratchetfreak <ratchetfreak!~ratchetfr@ptr-82s3g7nyo2zjr8pdfah.18120a2.ip6.access.telenet.be> has joined ##vulkan[12:40:31] *** trit <trit!~trit@pool-100-36-33-252.washdc.fios.verizon.net> has joined ##vulkan[13:33:59] *** penguin42 <penguin42!~dg@cpc109017-salf6-2-0-cust428.10-2.cable.virginm.net> has joined ##vulkan[14:16:46] *** Deluxe_ <Deluxe_!~Deluxe@212.4.150.151> has joined ##vulkan[14:21:36] *** Deluxe_ <Deluxe_!~Deluxe@212.4.150.151> has quit IRC (Remote host closed the connection)[14:23:41] *** Deluxe_ <Deluxe_!~Deluxe@212.4.150.151> has joined ##vulkan[14:24:04] *** Deluxe_ <Deluxe_!~Deluxe@212.4.150.151> has quit IRC (Remote host closed the connection)[15:12:37] *** RicardasSim <RicardasSim!~RicardasS@78-63-172-44.static.zebra.lt> has joined ##vulkan[15:13:55] *** felipealmeida <felipealmeida!~user@191.191.10.198> has joined ##vulkan[15:14:08] <felipealmeida> hi, is there any downsides to ask for as much queues as possible?[15:14:11] <felipealmeida> even though some might go unused?[15:16:28] *** haasn <haasn!~haasn@HSI-KBW-085-216-096-210.hsi.kabelbw.de> has quit IRC (Quit: WeeChat 2.6-dev)[15:21:15] *** haasn <haasn!~haasn@HSI-KBW-085-216-096-210.hsi.kabelbw.de> has joined ##vulkan[15:21:25] <felipealmeida> and, would it make sense to have more than one family with VK_QUEUE_GRAPHICS_BIT in the same physical device?[16:09:49] <[exa]> felipealmeida: how much is 'as much as possible'?[16:10:36] <felipealmeida> well, all available queues from the queue family[16:14:57] <[exa]> so not by vkGetDeviceQueue, but earlier when you create the device, right?[16:15:16] <felipealmeida> in VkDeviceQueueCreateInfo[16:15:35] <felipealmeida> yes, when creating the logical device[16:21:54] <[exa]> ok I realized I have no idea why there'd be a second layer of queue types between logical and hardware[16:22:10] <bnieuwenhuizen> what do you mean second layer of queue types?[16:23:40] <bnieuwenhuizen> felipealmeida: I'm not sure about as much queues as possible from a single type, but for the Mesa AMD drivers we might eventually use the queue types you asked for to help determine whether we can enable compression on a specific image[16:24:19] <felipealmeida> bnieuwenhuizen: by using a different VK_QUEUE_*_BIT ?[16:25:05] <bnieuwenhuizen> right[16:25:14] <felipealmeida> I thought queues were readily available, but intel seems to have just one in my GPD Pocket 2[16:25:38] <bnieuwenhuizen> it all depends on how many queues the HW really has[16:25:42] <felipealmeida> and I must be careful not to use the same family for presentation[16:26:07] <felipealmeida> bnieuwenhuizen: does that mean that not using all queues I'm letting resources idle in HW?[16:26:39] <bnieuwenhuizen> felipealmeida: using you main queue for presentation is generally OK[16:26:41] <felipealmeida> does queues work like mailbox for actual hardware? or is it something completely different?[16:27:01] <bnieuwenhuizen> the answer to idling resources is HW specific[16:27:05] <felipealmeida> I'd like to minimize contention as much as possible[16:27:19] <felipealmeida> so multiple queues allows me to use more threads with less contention[16:27:38] <felipealmeida> at least in my understanding[16:27:45] <bnieuwenhuizen> but for AMD all the resources are pretty much shared between all the graphics and compute queues, while the transfer queue is dedicated HW (though still shared bandwidth)[16:27:46] <felipealmeida> I may be completely wrong though[16:28:23] <felipealmeida> bnieuwenhuizen: so I can use the full potential in AMD by using just the main queue, right?[16:28:32] <bnieuwenhuizen> roughly, yes[16:28:52] <felipealmeida> but I'd still pay for unnecessarily locking if more queues are available[16:29:01] <bnieuwenhuizen> not really[16:29:17] <bnieuwenhuizen> locks are typically only taken during the VkQueueSubmit itself[16:29:43] <bnieuwenhuizen> and on really high level. You should be doing on the order of 10 QueueSubmits per frame anyway[16:29:45] <felipealmeida> yes, but submit is not really that cheap[16:30:05] <bnieuwenhuizen> fair[16:30:19] <felipealmeida> it does take a considerable amount of time from a frame time[16:31:38] <felipealmeida> anyway, it would be possible that for some hardware, not using all queues could leave part of the hardware idle[16:31:48] <felipealmeida> which makes for a stronger case for using as much queues as possible[16:32:35] <bnieuwenhuizen> felipealmeida: not that the vulkan spec gives no guidance at all on what options should be fast[16:32:49] <bnieuwenhuizen> so multiple queues might be fast, might be slow, it is just not specified[16:33:01] <felipealmeida> yeah, a bit a reasoning about the API could make it easier[16:33:43] <felipealmeida> indeed, anyway, testing is the only way it seems[16:35:15] <felipealmeida> it could be that some queues only upload to slower hw, like slower memory bus like there was in 980ti. So not using it could make latency better.[16:51:18] *** dreda <dreda!~elmindred@h87-96-185-93.cust.a3fiber.se> has quit IRC (Remote host closed the connection)[17:00:42] *** trit <trit!~trit@pool-100-36-33-252.washdc.fios.verizon.net> has quit IRC (Quit: Leaving)[17:05:25] *** dreda <dreda!~elmindred@h87-96-185-93.cust.a3fiber.se> has joined ##vulkan[17:08:14] *** ville <ville!~ville@212-149-214-47.bb.dnainternet.fi> has quit IRC (Quit:)[17:10:45] *** ville <ville!~ville@212-149-214-47.bb.dnainternet.fi> has joined ##vulkan[18:02:50] *** ImQ009 <ImQ009!~ImQ009@unaffiliated/imq009> has joined ##vulkan[18:07:02] *** slime <slime!~slime73@hlfxns016cw-156-34-18-74.dhcp-dynamic.fibreop.ns.bellaliant.net> has quit IRC (Quit: This computer has gone to sleep)[18:27:53] *** tristianc <tristianc!~tristianc@pool-72-83-128-193.washdc.fios.verizon.net> has joined ##vulkan[18:59:28] *** Andrej1 <Andrej1!~Andrej@BSN-143-212-5.dynamic.siol.net> has joined ##vulkan[19:00:08] *** tristianc <tristianc!~tristianc@pool-72-83-128-193.washdc.fios.verizon.net> has quit IRC (Quit: Konversation terminated!)[19:03:56] *** LunarJetman <LunarJetman!LunarJetma@94.14.216.120> has joined ##vulkan[19:11:28] *** ryp <ryp!ryp@gateway/vpn/privateinternetaccess/ryp> has quit IRC (Ping timeout: 268 seconds)[19:17:03] *** felipealmeida <felipealmeida!~user@191.191.10.198> has quit IRC (Remote host closed the connection)[19:30:07] *** glYoda <glYoda!~kristinam@c-73-25-27-206.hsd1.or.comcast.net> has quit IRC (Quit: glYoda)[19:32:28] *** glYoda <glYoda!~kristinam@c-73-25-27-206.hsd1.or.comcast.net> has joined ##vulkan[19:51:32] *** glYoda <glYoda!~kristinam@c-73-25-27-206.hsd1.or.comcast.net> has quit IRC (Quit: glYoda)[19:52:04] *** glYoda <glYoda!~kristinam@c-73-25-27-206.hsd1.or.comcast.net> has joined ##vulkan[20:27:54] *** ryp <ryp!ryp@gateway/vpn/privateinternetaccess/ryp> has joined ##vulkan[20:47:04] *** emersont1 <emersont1!~emersont1@2a00:23c5:2e3b:dc00:591a:20c6:b6f6:a4a9> has joined ##vulkan[20:57:09] *** glYoda <glYoda!~kristinam@c-73-25-27-206.hsd1.or.comcast.net> has quit IRC (Quit: glYoda)[21:00:45] *** penguin42 <penguin42!~dg@cpc109017-salf6-2-0-cust428.10-2.cable.virginm.net> has quit IRC (Remote host closed the connection)[21:06:29] *** glYoda <glYoda!~kristinam@c-73-25-27-206.hsd1.or.comcast.net> has joined ##vulkan[21:11:16] *** emersont1 <emersont1!~emersont1@2a00:23c5:2e3b:dc00:591a:20c6:b6f6:a4a9> has left ##vulkan ("Leaving")[21:44:23] *** slime <slime!~slime73@24.215.81.93> has joined ##vulkan[22:01:37] *** DrGonzo <DrGonzo!~DrGonzo@modemcable080.164-57-74.mc.videotron.ca> has joined ##vulkan[22:06:28] *** emersont1 <emersont1!~emersont1@2a00:23c5:2e3b:dc00:591a:20c6:b6f6:a4a9> has joined ##vulkan[22:12:52] *** emersont1 <emersont1!~emersont1@2a00:23c5:2e3b:dc00:591a:20c6:b6f6:a4a9> has quit IRC (Remote host closed the connection)[22:14:55] *** emersont1 <emersont1!~emersont1@2a00:23c5:2e3b:dc00:591a:20c6:b6f6:a4a9> has joined ##vulkan[22:15:17] *** slime <slime!~slime73@24.215.81.93> has quit IRC (Quit: This computer has gone to sleep)[22:16:12] *** slime <slime!~slime73@24.215.81.93> has joined ##vulkan[22:18:29] *** ImQ009 <ImQ009!~ImQ009@unaffiliated/imq009> has quit IRC (Quit: Leaving)[22:44:19] *** emersont1 <emersont1!~emersont1@2a00:23c5:2e3b:dc00:591a:20c6:b6f6:a4a9> has quit IRC (Remote host closed the connection)[22:59:41] *** Deluxe <Deluxe!~Deluxe@212.4.150.151> has quit IRC (Read error: Connection reset by peer)[23:26:45] *** slime <slime!~slime73@24.215.81.93> has quit IRC (Quit: This computer has gone to sleep)[23:51:00] *** ciaala <ciaala!~crypt@2a02:120b:2c1f:5f40:6ef0:49ff:feee:4777> has joined ##vulkan