February 16, 2017 
< | 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 | >
February 16, 2017 [00:05:53] *** joshsyn <joshsyn!~swoorup@14-203-132-32.static.tpgi.com.au> has quit IRC (Ping timeout: 240 seconds)[00:11:09] *** cactuz <cactuz!~cactus@0140800230.0.fullrate.ninja> has quit IRC (Quit: cactuz)[00:16:38] *** ShadowIce <ShadowIce!~pyoro@unaffiliated/shadowice-x841044> has quit IRC (Quit: Leaving)[00:20:00] *** CaptainLex <CaptainLex!~CaptainLe@12.104.244.5> has joined ##vulkan[00:21:59] *** Joefish <Joefish!~Joefish@p200300C6F3C57500CB54CD8F3EA22981.dip0.t-ipconnect.de> has quit IRC (Ping timeout: 252 seconds)[00:22:29] *** HuntsMan <HuntsMan!~hunts@cpc104822-sgyl39-2-0-cust136.18-2.cable.virginm.net> has joined ##vulkan[00:25:17] *** bjz <bjz!~bjz@14-201-122-110.static.tpgi.com.au> has quit IRC (Quit: My MacBook Pro has gone to sleep. ZZZzzz…)[00:28:18] *** afl_ext <afl_ext!~afl_ext3@unaffiliated/afl-ext/x-2796036> has quit IRC (Quit: Leaving)[00:42:46] *** bjz <bjz!~bjz@14-201-122-110.static.tpgi.com.au> has joined ##vulkan[00:57:01] *** bjz <bjz!~bjz@14-201-122-110.static.tpgi.com.au> has quit IRC (Quit: My MacBook Pro has gone to sleep. ZZZzzz…)[00:58:19] *** bjz <bjz!~bjz@14-201-122-110.static.tpgi.com.au> has joined ##vulkan[01:01:35] *** CaptainLex <CaptainLex!~CaptainLe@12.104.244.5> has quit IRC (Ping timeout: 252 seconds)[01:04:08] *** ZexaronS <ZexaronS!~zexpc@cpe-194-152-22-45.static.triera.net> has quit IRC (Quit: Leaving)[01:07:19] *** bjz <bjz!~bjz@14-201-122-110.static.tpgi.com.au> has quit IRC (Quit: My MacBook Pro has gone to sleep. ZZZzzz…)[01:14:01] *** derhass <derhass!~derhass@ipservice-092-216-126-046.092.216.pools.vodafone-ip.de> has quit IRC (Quit: leaving)[01:25:35] *** ratchetfreak <ratchetfreak!~ratchetfr@ptr-82s3g7omx242auk5xjq.18120a2.ip6.access.telenet.be> has quit IRC (Read error: Connection reset by peer)[01:26:00] *** ratchetfreak <ratchetfreak!~ratchetfr@ptr-82s3g7omx242auk5xjq.18120a2.ip6.access.telenet.be> has joined ##vulkan[01:41:08] *** Kingsquee <Kingsquee!~kingsley@d172-218-58-109.bchsia.telus.net> has joined ##vulkan[01:42:20] *** parasite_ <parasite_!~parasite@mar75-4-82-230-46-11.fbx.proxad.net> has joined ##vulkan[01:44:08] *** spossiba <spossiba!~spossiba@unaffiliated/stryx/x-3871776> has quit IRC (Quit: leaving)[01:45:10] *** spossiba <spossiba!~spossiba@unaffiliated/stryx/x-3871776> has joined ##vulkan[01:56:57] *** BearishMushroom <BearishMushroom!~BearishMu@90-231-174-194-no159.tbcn.telia.com> has quit IRC (Read error: Connection reset by peer)[01:57:16] *** ector <ector!~asdf@37-247-11-157.customers.ownit.se> has quit IRC ()[02:19:48] *** john1 <john1!~john@host86-147-123-91.range86-147.btcentralplus.com> has joined ##vulkan[02:24:10] *** john1 <john1!~john@host86-147-123-91.range86-147.btcentralplus.com> has quit IRC (Ping timeout: 240 seconds)[02:24:49] *** ratchetfreak <ratchetfreak!~ratchetfr@ptr-82s3g7omx242auk5xjq.18120a2.ip6.access.telenet.be> has quit IRC (Ping timeout: 255 seconds)[03:02:35] *** spossiba <spossiba!~spossiba@unaffiliated/stryx/x-3871776> has quit IRC (Ping timeout: 252 seconds)[03:06:14] *** spossiba <spossiba!~spossiba@unaffiliated/stryx/x-3871776> has joined ##vulkan[03:12:02] *** glYoda <glYoda!~MTLYoda@c-71-236-230-251.hsd1.or.comcast.net> has joined ##vulkan[03:25:55] *** kdub <kdub!~kdub@d28-23-138-68.dim.wideopenwest.com> has quit IRC (Quit: goodnight)[03:28:52] *** TechnoCrunch <TechnoCrunch!~TechnoCru@101.100.137.146> has joined ##vulkan[03:31:04] *** w-t-h_ <w-t-h_!~chatzilla@72.67.52.110> has joined ##vulkan[04:13:56] *** GMpow3srV <GMpow3srV!~goedelmac@55d46705.access.ecotel.net> has joined ##vulkan[04:17:48] *** GMpow2srV <GMpow2srV!~goedelmac@55d463d2.access.ecotel.net> has quit IRC (Ping timeout: 256 seconds)[04:30:01] *** Kingsquee <Kingsquee!~kingsley@d172-218-58-109.bchsia.telus.net> has quit IRC (Read error: Connection reset by peer)[04:30:18] *** Kingsquee <Kingsquee!~kingsley@d172-218-58-109.bchsia.telus.net> has joined ##vulkan[04:31:57] *** spossiba <spossiba!~spossiba@unaffiliated/stryx/x-3871776> has quit IRC (Ping timeout: 240 seconds)[04:33:06] *** ville <ville!~ville@87-95-81-153.bb.dnainternet.fi> has quit IRC (Ping timeout: 256 seconds)[04:35:02] *** spossiba <spossiba!~spossiba@unaffiliated/stryx/x-3871776> has joined ##vulkan[04:37:32] *** Biolunar_ <Biolunar_!Biolunar@dslb-084-061-188-206.084.061.pools.vodafone-ip.de> has joined ##vulkan[04:40:46] *** Biolunar <Biolunar!Biolunar@dslb-094-220-175-028.094.220.pools.vodafone-ip.de> has quit IRC (Ping timeout: 245 seconds)[04:53:06] *** lpghatguy <lpghatguy!~LPGhatguy@63.153.99.104> has joined ##vulkan[04:55:08] *** ville <ville!~ville@37-33-125-214.bb.dnainternet.fi> has joined ##vulkan[05:00:04] *** w-t-h_ <w-t-h_!~chatzilla@72.67.52.110> has quit IRC (Ping timeout: 255 seconds)[05:01:57] *** doomlord <doomlord!~textual@host86-148-102-241.range86-148.btcentralplus.com> has quit IRC (Quit: My MacBook Pro has gone to sleep. ZZZzzz…)[05:02:05] *** john1 <john1!~john@host86-147-123-91.range86-147.btcentralplus.com> has joined ##vulkan[05:13:34] *** nidefawl <nidefawl!~nidefawl@p57B2FA06.dip0.t-ipconnect.de> has quit IRC ()[05:14:15] *** ZexaronS <ZexaronS!~zexpc@cpe-194-152-22-45.static.triera.net> has joined ##vulkan[05:16:35] *** w-t-h_ <w-t-h_!~chatzilla@72.67.52.110> has joined ##vulkan[05:41:15] *** w-t-h_ <w-t-h_!~chatzilla@72.67.52.110> has quit IRC (Read error: Connection reset by peer)[05:42:22] *** GMpow2 <GMpow2!~5@2a02:810a:83c0:cb14:404e:fa8:2ab3:2bca> has quit IRC (Ping timeout: 255 seconds)[06:01:24] *** nidefawl <nidefawl!~nidefawl@p57B2FA06.dip0.t-ipconnect.de> has joined ##vulkan[06:02:33] *** spossiba <spossiba!~spossiba@unaffiliated/stryx/x-3871776> has quit IRC (Ping timeout: 258 seconds)[06:05:47] *** spossiba <spossiba!~spossiba@unaffiliated/stryx/x-3871776> has joined ##vulkan[06:07:04] *** tambre <tambre!~tambre@a5be-f20e-d303-198d-0e80-8a31-07d0-2001.dyn.estpak.ee> has joined ##vulkan[06:57:57] *** robot-beethoven <robot-beethoven!~robot-bee@c-73-65-29-166.hsd1.mn.comcast.net> has joined ##vulkan[07:04:02] *** nsf <nsf!~nsf@jiss.convex.ru> has joined ##vulkan[07:08:18] *** joshsyn <joshsyn!~swoorup@14-203-132-32.static.tpgi.com.au> has joined ##vulkan[07:58:18] *** sla_ro|master <sla_ro|master!~sla.ro@95.76.45.217> has joined ##vulkan[08:14:01] *** spossiba <spossiba!~spossiba@unaffiliated/stryx/x-3871776> has quit IRC (Ping timeout: 255 seconds)[08:14:37] *** spossiba <spossiba!~spossiba@unaffiliated/stryx/x-3871776> has joined ##vulkan[08:21:32] *** ratchetfreak <ratchetfreak!~ratchetfr@ptr-82s3g7omx242auk5xjq.18120a2.ip6.access.telenet.be> has joined ##vulkan[08:26:10] *** ratchetfreak <ratchetfreak!~ratchetfr@ptr-82s3g7omx242auk5xjq.18120a2.ip6.access.telenet.be> has quit IRC (Ping timeout: 255 seconds)[08:26:11] *** Joefish <Joefish!~Joefish@p200300C6F3CBF1009A05AA2309BCFF95.dip0.t-ipconnect.de> has joined ##vulkan[08:57:49] *** slidercrank <slidercrank!~slidercra@unaffiliated/slidercrank> has quit IRC (Quit: siht gnidaer er'uoy fi emit eerf hcum oot evah uoy :tiuQ)[08:59:01] *** realitix <realitix!~realitix@92.103.166.6> has joined ##vulkan[09:02:50] *** spossiba <spossiba!~spossiba@unaffiliated/stryx/x-3871776> has quit IRC (Ping timeout: 256 seconds)[09:06:16] *** spossiba <spossiba!~spossiba@unaffiliated/stryx/x-3871776> has joined ##vulkan[09:16:34] *** doomlord <doomlord!~textual@host86-148-102-241.range86-148.btcentralplus.com> has joined ##vulkan[09:21:21] *** ImQ009 <ImQ009!~ImQ009@unaffiliated/imq009> has joined ##vulkan[09:27:25] *** robot-beethoven <robot-beethoven!~robot-bee@c-73-65-29-166.hsd1.mn.comcast.net> has quit IRC (Quit: leaving)[09:37:53] *** ravior <ravior!~crapitea@89.121.200.106> has joined ##vulkan[09:41:46] *** slidercrank <slidercrank!~slidercra@unaffiliated/slidercrank> has joined ##vulkan[09:48:59] *** TechnoCrunch <TechnoCrunch!~TechnoCru@101.100.137.146> has quit IRC (Read error: Connection reset by peer)[09:49:45] <neure> sharpneli, ping[09:51:08] <sharpneli> pong[09:51:34] <neure> which preprocessor you use?[09:51:46] <neure> i tried simple cl /p but it complains about #version etc.[09:53:56] <sharpneli> bsd licensed cpp. But we insert the version tag afterwards. So the one going to the preprocessor has no extraneous tags[09:55:47] *** ratchetfreak <ratchetfreak!c18021f8@gateway/web/freenode/ip.193.128.33.248> has joined ##vulkan[09:56:11] *** bjz <bjz!~bjz@14-201-122-110.static.tpgi.com.au> has joined ##vulkan[10:04:24] *** xkpe <xkpe!~xkpe@212.250.200.210> has quit IRC (Read error: Connection reset by peer)[10:14:11] *** ZexaronS <ZexaronS!~zexpc@cpe-194-152-22-45.static.triera.net> has quit IRC (Quit: Leaving)[10:33:35] *** spossiba <spossiba!~spossiba@unaffiliated/stryx/x-3871776> has quit IRC (Ping timeout: 240 seconds)[10:34:25] *** lpghatguy <lpghatguy!~LPGhatguy@63.153.99.104> has quit IRC (Quit: I'll be back.)[10:34:40] *** spossiba <spossiba!~spossiba@unaffiliated/stryx/x-3871776> has joined ##vulkan[10:36:42] *** elect <elect!~elect@business-092-079-135-041.static.arcor-ip.net> has joined ##vulkan[10:42:33] <neure> so looks like mcpp is decent preprocessor, but it also dislikes #version and I cant get rid of couple of stdc defines[10:47:53] <ratchetfreak> do you really need a C preprocessor?[10:48:14] <ratchetfreak> could any other text manip tool help?[10:51:11] <neure> the C preprocessor is pretty standard[10:54:50] *** sla_ro|master <sla_ro|master!~sla.ro@95.76.45.217> has quit IRC ()[10:57:59] <ratchetfreak> but far from the best designed language[11:03:09] <neure> sharpneli, what about #extension ?[11:03:32] <neure> it would be great if there was an actual, standard preprocessor for glsl[11:04:14] <sharpneli> neure: We don't use extensions.[11:04:15] <sharpneli> Yap[11:04:31] <sharpneli> If we do need extensions then we'll have to insert them. but for now only version tag is needed[11:06:17] <sharpneli> ratchetfreak: For convenience it's just good to have support for basic #include tags etc.[11:09:54] *** realitix <realitix!~realitix@92.103.166.6> has quit IRC (Read error: Connection reset by peer)[11:10:05] *** realitix <realitix!~realitix@92.103.166.6> has joined ##vulkan[11:12:14] <exDM69> Cheery: regarding your reddit post: it would be great if you didn't spread unsubstantiated rumors or present subjective opinions as facts[11:13:40] <exDM69> I understand your frustration with the API but come on...[11:20:12] <sharpneli> What post is this about?[11:21:40] *** joshsyn <joshsyn!~swoorup@14-203-132-32.static.tpgi.com.au> has quit IRC (Ping timeout: 255 seconds)[11:30:39] *** Rangar <Rangar!~Dave@124.197.29.80> has joined ##vulkan[11:31:17] <fazias_> be glad that you can fill commandlists of unspecified lengths.[11:31:29] *** fazias_ is now known as fazias[11:33:28] *** spossiba <spossiba!~spossiba@unaffiliated/stryx/x-3871776> has quit IRC (Ping timeout: 260 seconds)[11:34:59] *** spossiba <spossiba!~spossiba@unaffiliated/stryx/x-3871776> has joined ##vulkan[11:35:10] *** Lucretia <Lucretia!~laguest@pdpc/supporter/active/lucretia> has quit IRC (Ping timeout: 255 seconds)[11:35:49] *** bjz <bjz!~bjz@14-201-122-110.static.tpgi.com.au> has quit IRC (Quit: My MacBook Pro has gone to sleep. ZZZzzz…)[11:38:34] <ratchetfreak> glslang has a only-preprocess flag: https://github.com/KhronosGroup/glslang/blob/master/StandAlone/StandAlone.cpp#L939[11:40:57] *** john1 <john1!~john@host86-147-123-91.range86-147.btcentralplus.com> has quit IRC (Quit: WeeChat 1.7)[11:42:20] <fazias> I think it's funny to think that "pipelined" would somehow disappear... because thats the basic idea when you want more throughput.[11:43:00] *** bjz <bjz!~bjz@14-201-122-110.static.tpgi.com.au> has joined ##vulkan[11:43:11] <sharpneli> ratchetfreak: Doesn't support #include[11:43:21] <sharpneli> ratchetfreak: Which was the biggest reason for us to use our own[11:43:30] <fazias> shaderc supports include[11:43:34] *** timperi <timperi!~timperi@mobile-user-2e84ba-78.dhcp.inet.fi> has quit IRC (Ping timeout: 264 seconds)[11:43:53] <sharpneli> Yes. But it's not glslang[11:44:02] <fazias> but it backends to glslang...? no?[11:44:03] <sharpneli> And 'sides. It was trivial to throw it trough a normal CPP[11:44:09] *** timperi <timperi!~timperi@mobile-user-2e84ba-78.dhcp.inet.fi> has joined ##vulkan[11:45:06] <sharpneli> Btw. What do you mean pipelined would disappear?[11:45:14] <fazias> well, just that one reddit post.[11:45:21] <sharpneli> What reddit post?[11:45:28] <fazias> https://www.reddit.com/r/vulkan/comments/5u36km/terraforming_the_vulkan/[11:45:36] <fazias> not like vulkan reddit has that many posts.. :p[11:45:46] *** john1 <john1!~john@host86-147-123-91.range86-147.btcentralplus.com> has joined ##vulkan[11:46:05] <fazias> It only sounds like someone needs higher level api and shouldn't be using vulkan.[11:47:37] <fazias> there is nothing hostile in vulkan, it mainly shows you what actually needs to be done on fairly close level to what hardware is.[11:49:22] *** bjz <bjz!~bjz@14-201-122-110.static.tpgi.com.au> has quit IRC (Ping timeout: 240 seconds)[11:49:35] *** bjz_ <bjz_!~bjz@104.222.140.91> has joined ##vulkan[11:55:59] <sharpneli> Yah. Sounds exactly like that[11:56:59] <sharpneli> From his WarpGPU "There is a hint that barrier insertion can be abstracted away. LunarG validation layers appears to be able to point out the missing barriers.". Yes. Barriers can be abstracted away. It's just heavy on CPU side and automatically you cant always do optimal barriers. Hence the whole reason why Vulkan exposes them explicitly.[12:05:29] <sharpneli> The rumours in his post are pure FUD tho. We're writing Vulkan quite well without doing any dialup to anyone. And wtf is it about things stuffed to vulkan stop mattering? Things like what?[12:28:29] *** Hunts <Hunts!~hunts@dhcp9.eps.hw.ac.uk> has joined ##vulkan[12:31:31] <tomaka> I guess the reason why they think vulkan will stop mattering is that OpenGL 1 stopped mattering in the past[12:31:50] <tomaka> but such a transition is not going to happen again any time soon[12:32:09] <sharpneli> Yeah. Considering that OpenGL is almost 30 years old now[12:32:46] <sharpneli> And you can still write completely well working OpenGL 1 application and it works nice and dandy.[12:36:25] <tomaka> to be fair that's possible because OpenGL is higher-level and because CPUs are much faster than in the past[12:37:13] <tomaka> if CPUs were the same speed as in the past, the translation would probably kill performances[12:37:26] <sharpneli> Sure. But OpenGL nowadays is not any slower than before. It's just that before simple rasterization was so expensive that the function call overhead didn't matter at all[12:38:05] <fazias> gpu's are so fast that they need to be fed ahead of time with work to keep them busy.[12:38:36] <tomaka> yeah I confused myself with that cpu speed thing[12:38:42] <tomaka> whatever[12:39:17] *** jkilpatr_ <jkilpatr_!~jkilpatr@2602:30a:c7da:b600:1202:b5ff:fe3b:d41d> has quit IRC (Ping timeout: 276 seconds)[12:39:32] <sharpneli> Yup. The bottlenecks have changed.[12:52:37] *** HuntsMan <HuntsMan!~hunts@cpc104822-sgyl39-2-0-cust136.18-2.cable.virginm.net> has quit IRC (Read error: Connection reset by peer)[12:53:01] *** HuntsMan <HuntsMan!~hunts@cpc104822-sgyl39-2-0-cust136.18-2.cable.virginm.net> has joined ##vulkan[12:57:12] <ratchetfreak> and they may change again 10 years into the future but that is what vulkan 2.0 is for[13:00:22] *** jkilpatr <jkilpatr!jkilpatr@nat/redhat/x-kjsmkutrlzyqxqlv> has joined ##vulkan[13:01:40] <Prf_Jakob> There are a couple of things that makes OpenGL slow for the CPU, you need lots of calls to change certain state, and in the drawcall you need to do a bunch of validation.[13:02:22] <Prf_Jakob> Those are the two big things.[13:02:28] <ratchetfreak> and there are sync pitfalls that can create massive pipeline salls[13:02:41] <ratchetfreak> *stalls[13:02:46] <Prf_Jakob> Thats a GPU thing, but yes there is that as well. :)[13:04:32] <ratchetfreak> one of the biggest ones is readPixels that only "recently" got fixed with PBOs, even though the fix should have been (and was) straightforward[13:04:41] <Prf_Jakob> One of the things you could do with OpenGL is put all of the state onto the shader program object, that would allow you to quickly switch between states and you could do validation up front.[13:06:12] <Prf_Jakob> ratchetfreak: Recently? ARB_pbo was published 2004.[13:07:08] <ratchetfreak> but only got into core in 4.5[13:13:42] <ratchetfreak> and worst of all there is no DSA version of ReadPixels and most comments on PBO seem to make things more complicated than they really are[13:13:48] <Prf_Jakob> But its been ubiquities since around that time.[13:13:52] *** Hunts <Hunts!~hunts@dhcp9.eps.hw.ac.uk> has quit IRC (Ping timeout: 240 seconds)[13:14:10] <Prf_Jakob> That is bad tho :-/[13:14:11] *** kdub <kdub!~kdub@d28-23-138-68.dim.wideopenwest.com> has joined ##vulkan[13:16:38] *** tarceri <tarceri!~tarceri@CPE-120-144-40-53.lnse5.win.bigpond.net.au> has quit IRC (Remote host closed the connection)[13:17:46] *** nidefawl <nidefawl!~nidefawl@p57B2FA06.dip0.t-ipconnect.de> has quit IRC (Ping timeout: 255 seconds)[13:20:30] *** nidefawl <nidefawl!~nidefawl@p57B2FA06.dip0.t-ipconnect.de> has joined ##vulkan[13:27:36] *** nidefawl <nidefawl!~nidefawl@p57B2FA06.dip0.t-ipconnect.de> has quit IRC (Ping timeout: 240 seconds)[13:31:19] *** tacchinotacchi <tacchinotacchi!~tacchinot@net-31-27-221-94.cust.vodafonedsl.it> has quit IRC (Ping timeout: 268 seconds)[13:32:32] *** Kingsquee <Kingsquee!~kingsley@d172-218-58-109.bchsia.telus.net> has quit IRC (Quit: https://i.imgur.com/qicT3GK.gif)[13:32:59] *** Ralith_ <Ralith_!~ralith@c-24-56-236-197.customer.broadstripe.net> has joined ##vulkan[13:33:26] *** tarceri <tarceri!~tarceri@CPE-120-144-40-53.lnse5.win.bigpond.net.au> has joined ##vulkan[13:33:48] <neure> what you mean shaderc is not glslang?[13:33:57] <neure> it internal uses glslang[13:34:09] <neure> > glslangvalidator -V -E Phong.frag -o Phong.frag.pre[13:34:09] <neure> glslangvalidator: Error can't use -E when linking is selected (use -h for usage)[13:34:17] <neure> > glslangvalidator -E Phong.frag -o Phong.frag.pre[13:34:17] <neure> glslangvalidator: Error no binary generation requested (e.g., -V) (use -h for usage)[13:34:25] <neure> wtf[13:34:30] *** Ralith__ <Ralith__!~ralith@c-24-143-67-131.customer.broadstripe.net> has quit IRC (Ping timeout: 258 seconds)[13:34:32] <neure> im not sure how to use -E[13:35:03] <sharpneli> neure: As in it's not the same project. It would add an extra dependency and if glslang gets updated but shaderc lags we'd be stuck[13:35:25] *** sla_ro|master <sla_ro|master!~sla.ro@95.76.45.217> has joined ##vulkan[13:35:33] <sharpneli> Gotta see in future if a change is needed. Right now this approach works nicely.[13:35:35] <neure> does -E actually work somehow?[13:36:01] *** bjz_ <bjz_!~bjz@104.222.140.91> has quit IRC (Quit: My MacBook Pro has gone to sleep. ZZZzzz…)[13:45:32] <Cheery> exDM69: I'm a bit uncertain whether that post is resulting in anything that is in my interests. I could have been a bit more modest.[13:46:00] <exDM69> yes, you could have...[13:46:19] <exDM69> the "rumors" you mention have no basis in reality and I don't know where you've heard that kind of stuff[13:46:42] <baldurk> a bit more modest and a hell of a lot less crazy[13:47:57] *** Andrej1 <Andrej1!~Andrej@86.61.99.245> has quit IRC (Ping timeout: 240 seconds)[13:48:24] *** Andrej1 <Andrej1!~Andrej@86.61.104.172> has joined ##vulkan[13:53:25] <sharpneli> Cheery: Where did the rumours came from btw?[13:53:31] *** sla_ro|master <sla_ro|master!~sla.ro@95.76.45.217> has quit IRC ()[13:54:09] <Cheery> sharpneli: I got a friend that keeps claiming those things. It has bothered me whether he's right or wrong.[13:54:33] <sharpneli> What have been his arguments to support those claims?[13:56:33] <Cheery> I think he's been mostly observing the spec, reading up things. I don't think he has made many of arguments supporting the claims.[13:58:06] <exDM69> repeating unsubstantiated rumors from someone with no credibility whatsoever is harmful[13:58:10] <sharpneli> Without any arguments it's kinda hard to refute them. And it's also a good reason just to consider them being randomly thrown out into air[13:58:22] <Cheery> he's doing webgl for living.[13:58:34] *** bjz <bjz!~bjz@14-201-122-110.static.tpgi.com.au> has joined ##vulkan[13:58:43] <exDM69> ... which proves nothing[13:59:04] <ratchetfreak> *insert comment about webdev not knowing about perf critical programming*[13:59:22] <exDM69> even if he'd have connections to khronos members, it still wouldn't be very credible[13:59:55] <sharpneli> I wonder what he even meant by "most of things stuffed into Vulkan will stop to matter in just few years" as no examples were given[14:00:23] <neure> what are you talking about?[14:00:37] <sharpneli> neure: https://xkcd.com/386/[14:00:43] <sharpneli> https://www.reddit.com/r/vulkan/comments/5u36km/terraforming_the_vulkan/[14:00:59] <neure> lol[14:01:43] <sharpneli> But yeah. If your friend has only practical experience with webgl then yes it's likely that he has a view that it requires hotline to AMD and Nvidia[14:02:19] <sharpneli> With practical experience of OpenGL, DirectX, OpenCL and CUDA it's really quite straightforward to use in a neutral way that's still performant.[14:02:52] <neure> people can build fancy abstractions on top of vulkan[14:03:02] <neure> so everything that is desired in the post can be achieved[14:03:37] <neure> if vulkan directly was something nice and abstracted, it would be restricting[14:03:43] <neure> vulkan as is has no such restrictions[14:04:00] <neure> thats about all worth saying about that post[14:04:18] <sharpneli> Well it still kinda is abstracted. As low as one can go considering the variety of hardware we have.[14:04:19] <neure> you want something nice? make it or buy it, vulkan can make it happen[14:04:40] <neure> yeah[14:04:57] <neure> it is not unexpected that people still today wish vulkan was "nice"[14:05:03] <neure> those people will never disappear[14:05:40] <neure> vulkan is enabling technology[14:06:06] <neure> "nice A" and "nice B" and "nice unity" etc. are all possible[14:13:19] <Cheery> neure: are those beliefs or facts?[14:13:58] <Cheery> well for certain this is a fact: people can build fancy abstractions on top of vulkan.[14:13:59] *** BearishMushroom <BearishMushroom!~BearishMu@90-231-174-194-no159.tbcn.telia.com> has joined ##vulkan[14:15:06] <sharpneli> That's the whole point of Vulkan. So driver writers wouldn't have to write the fancy abstraction and the driver can be really thin[14:15:11] <Cheery> but how practical is it, how much does it cost in different ways, can it be done in usual time constraints?[14:15:48] <sharpneli> Use an opensource or paid middleware to provide the abstraction for you. Far better than dealing with the bugs that Qualcomm provided nice abstraction has a bug so it has to be used differently compared to someone elses[14:16:11] <sharpneli> It's all about saving the driver writers all having to support this massive middleware blob.[14:17:21] <sharpneli> On Windows the situation is bit better in that sense. The GPU vendors write against WDDM and most of the massive middleware is provided by Microsoft.[14:20:39] <Cheery> sharpneli: do you know some details about abstracting away pipeline barriers?[14:20:57] <Cheery> sharpneli: or any documents discussing that subject?[14:22:08] <Cheery> on the early look I thought that automatic barrier insertion is difficult.[14:23:40] *** joshsyn <joshsyn!~swoorup@14-203-132-32.static.tpgi.com.au> has joined ##vulkan[14:23:45] *** spossiba <spossiba!~spossiba@unaffiliated/stryx/x-3871776> has quit IRC (Ping timeout: 260 seconds)[14:24:08] <Cheery> but then I thought it's one of the things that maybe isn't important to abstract away.[14:25:21] *** spossiba <spossiba!~spossiba@unaffiliated/stryx/x-3871776> has joined ##vulkan[14:27:17] <neure> you could have a layer do all synchronizations for you[14:27:20] <neure> like opengl does[14:27:54] <neure> but then it would prevent you from doing things optimally, with granularity that is based on knowledge on your specific case[14:28:47] <neure> vulkan allows both manual-optimal and middleware-general-but-it-works[14:30:00] <neure> inserting barriers is not difficult. you just need to know what depends on what[14:30:25] <Cheery> sort of.. that's the problem yeah.[14:30:46] <neure> if you dont know what depends on what, you dont basically know what you are doing[14:30:58] <Cheery> and you need to go in abstraction layers a lot upwards if you want to know what depends on what.[14:31:16] <neure> no, not really[14:31:57] <neure> nothing says you need to have many abstraction layers[14:32:10] *** bjz <bjz!~bjz@14-201-122-110.static.tpgi.com.au> has quit IRC (Quit: My MacBook Pro has gone to sleep. ZZZzzz…)[14:32:16] <neure> you can have all the dependency information in your one and only layer[14:33:19] <sharpneli> Cheery: It all depends. You can either do the traditional style state tracking. It's brittle and slow but works. We are doing task graph based approach.[14:33:35] <Cheery> um.. if I got this roughly right.. to automate the barrier insertion you have to know: which resources are going to be written to, which resources are going to be read.[14:33:43] <sharpneli> Cheery: As in shadow pass is described as outputting to a rendertarget and the opaque pass then reads from that. Then we just walk trough that graph and produce a scheduling that way.[14:33:56] <neure> Cheery, that information is present in SPIR-V debug info[14:34:04] <sharpneli> Yup[14:34:34] <fazias> in most of my projects I have concept of SRV and UAV (shader resource and Unordered access)[14:34:43] <fazias> basically latter means "read/write" and first one means "read"[14:35:27] <fazias> also automatic barriers is good if you don't have that many draws calls.[14:35:46] <sharpneli> There are bunch of ways of doing it. And it is kinda interesting as now people are exploring the possible abstractions. We are few years out until we have bunch of best practices for it[14:36:08] <sharpneli> fazias: We have automated barriers and ~10k drawcalls.[14:36:28] <fazias> :)[14:36:45] <fazias> I should write the "begin/end" splitbarrier variation soon[14:37:11] <fazias> I haven't optimized it much but you could generate commandlists multithreaded anyway.[14:37:27] <neure> sharpneli, was it you who recommended parsing spirv debug info?[14:37:36] <neure> anyway that was an excellent suggestion[14:37:43] <neure> couldnt do it any other way[14:37:47] <sharpneli> Yah. We also use the same task graph for generating command lists in multithreaded fashion and also deduce in how many queues we can submit it into[14:37:58] <fazias> basically takes about 1ms, and there is atleast 50% or more overhead[14:38:01] <sharpneli> neure: Can't remember. Likely not. so thank someone else :D[14:43:08] <Cheery> the little while I've worked on Vulkan. I feel it's been a bit like "I've been there before".[14:43:29] <Cheery> but it was opengl back then[14:43:49] *** z3r0_ <z3r0_!~z3r0@197.234.219.80> has joined ##vulkan[14:44:50] <fazias> though my barrier systems doesn't need to be able to handle 10k drawcalls[14:46:24] <sharpneli> fazias: Consider that most drawcalls happen in groups. As an example a single renderpass can have abritrary amount of drawcalls and for any barrier analysis it's just a single unit[14:46:49] <Cheery> for example.. Vulkan can be instantiated without a windowing system. But if you want to draw on the screen you need to get the WSI up before you instantiate, so you know whether the device can draw to the window you want to draw to.[14:46:49] <sharpneli> It's more about the amount of read/write relationships in general[14:46:52] <fazias> this is on dx12, so everything is barrier.[14:47:07] <fazias> vulkan is easier, yes. less things to track.[14:47:10] <sharpneli> Ah yeah. To be honest the DX12 resource system sucks[14:47:13] <fazias> nah[14:47:15] <sharpneli> Especially as they change things[14:47:15] <fazias> its trivial[14:47:26] <fazias> I have subresource accuracy tracker working.[14:47:32] <Cheery> so in the effect Vulkan provided some things more, but if you want to do same as in opengl, you seem to face similar restrictions.[14:47:38] <sharpneli> It has this weird concept of buffer also having layouts, and then they bypassed that by giving implicit layout promotions[14:47:45] <fazias> I already beat the dx11 driver atleast on amd.[14:47:47] <sharpneli> Which makes it clunky to use[14:47:54] <sharpneli> Yeah that's not hard for fresh code[14:47:56] <fazias> but not by much, though my code is a bit unoptimized.[14:48:00] <fazias> nope[14:48:05] <sharpneli> But compared to Vulkan the resource system is horrible[14:48:12] <fazias> this is on codebase which was written dx12 like[14:48:28] <sharpneli> In that case nice[14:48:36] <sharpneli> You're porting existing codebase?[14:48:37] <fazias> but there are some unsolved problems[14:48:44] <fazias> like uploading way too much data.[14:48:50] <fazias> but I have solutions for those alreayd.[14:49:01] <fazias> kind of yeah. Cannot say much about it.[14:49:02] <sharpneli> you're uploading stuff you don't need to upload?[14:50:33] <fazias> the case is that you have limited amount of upload heap[14:50:51] <fazias> and in our codebase we have all sorts of fancy stuff[14:51:10] <fazias> so someone might suddenly want to put hundreds of megs of updating texture stuff.[14:52:58] <fazias> so basically I can split that up in the backend and submit to gpu in small pieces.[14:53:22] <sharpneli> Yup. No other option relaly[14:53:40] *** rpb <rpb!~rpb@199.68.195.102> has quit IRC (Ping timeout: 240 seconds)[14:54:16] <fazias> I'm excited to see what frostbite has to say about their framegraph at gdc[14:55:25] <sharpneli> Oh damn. That actually sounds pretty much like what we are doing. Based on http://schedule.gdconf.com/session/framegraph-extensible-rendering-architecture-in-frostbite[14:55:59] <sharpneli> + we're going to right direction. - we weren't the ones giving the first talk :D[14:56:58] <fazias> :)[14:59:01] <sharpneli> But honestly it's such an obvious direction that would be weird not to see someone else doing it too[14:59:47] <fazias> There are few ways to go at it, but I'm interested to see how their api looks. I know only how one can do the "commandlist" type well.[15:01:40] <sharpneli> I bet even they have some things in traditional commandlists. As an example transparent objects. API order is really important there. So drawing itself is still in commandlist fashion.[15:08:35] *** joshsyn <joshsyn!~swoorup@14-203-132-32.static.tpgi.com.au> has quit IRC (Ping timeout: 252 seconds)[15:14:10] <ratchetfreak> or they have a sort transparent operation before submitting the commands[15:15:57] <sharpneli> Sure. I just meant that the abstraction likely still has traces of commandlists.[15:16:01] <sharpneli> But yeah. Might not be. hard to say.[15:18:36] *** ravior <ravior!~crapitea@89.121.200.106> has quit IRC (Quit: Konversation terminated!)[15:20:40] *** Serus <Serus!~Serus@unaffiliated/serus> has quit IRC (Ping timeout: 240 seconds)[15:23:24] *** Serus <Serus!~Serus@unaffiliated/serus> has joined ##vulkan[15:31:20] *** z3r0_ <z3r0_!~z3r0@197.234.219.80> has quit IRC (Quit: Leaving)[15:32:47] *** nsf <nsf!~nsf@jiss.convex.ru> has quit IRC (Quit: WeeChat 1.7)[15:45:17] *** tacchinotacchi <tacchinotacchi!~tacchinot@151.35.109.181> has joined ##vulkan[15:48:35] *** nidefawl_ <nidefawl_!~nidefawl@104.244.76.248> has joined ##vulkan[15:49:15] *** tacchinotacchi_ <tacchinotacchi_!~tacchinot@net-188-152-21-183.cust.dsl.teletu.it> has joined ##vulkan[15:49:32] *** Lucretia <Lucretia!~laguest@pdpc/supporter/active/lucretia> has joined ##vulkan[15:50:08] *** tacchinotacchi <tacchinotacchi!~tacchinot@151.35.109.181> has quit IRC (Ping timeout: 260 seconds)[15:52:01] *** nidefawl_ is now known as nidefawl[16:01:46] *** afl_ext <afl_ext!~afl_ext3@unaffiliated/afl-ext/x-2796036> has joined ##vulkan[16:09:03] *** jkilpatr <jkilpatr!jkilpatr@nat/redhat/x-kjsmkutrlzyqxqlv> has quit IRC (Quit: Leaving)[16:09:21] *** jkilpatr <jkilpatr!jkilpatr@nat/redhat/x-irzqpurtdgsejmoh> has joined ##vulkan[16:20:23] *** mflow <mflow!~tty@unaffiliated/shiningthrough> has joined ##vulkan[16:29:11] *** sla_ro|master <sla_ro|master!~sla.ro@95.76.45.217> has joined ##vulkan[16:37:16] *** ector <ector!~asdf@37-247-11-157.customers.ownit.se> has joined ##vulkan[16:38:48] *** ville <ville!~ville@37-33-125-214.bb.dnainternet.fi> has quit IRC (Ping timeout: 240 seconds)[16:39:23] *** mflow <mflow!~tty@unaffiliated/shiningthrough> has quit IRC ()[16:40:09] *** tacchinotacchi_ <tacchinotacchi_!~tacchinot@net-188-152-21-183.cust.dsl.teletu.it> has quit IRC (Remote host closed the connection)[16:42:43] *** afl_ext <afl_ext!~afl_ext3@unaffiliated/afl-ext/x-2796036> has quit IRC (Quit: Leaving)[16:42:55] *** afl_ext <afl_ext!~afl_ext3@unaffiliated/afl-ext/x-2796036> has joined ##vulkan[16:46:56] *** tacchinotacchi <tacchinotacchi!~tacchinot@net-188-152-21-183.cust.dsl.teletu.it> has joined ##vulkan[16:54:16] *** Feedz <Feedz!~Feedz@unaffiliated/feedz> has joined ##vulkan[17:06:57] *** xissburg <xissburg!~xissburg@unaffiliated/xissburg> has quit IRC (Ping timeout: 240 seconds)[17:11:05] *** xissburg <xissburg!~xissburg@unaffiliated/xissburg> has joined ##vulkan[17:19:53] *** Hunts <Hunts!~hunts@dhcp9.eps.hw.ac.uk> has joined ##vulkan[17:44:10] *** CaptainLex <CaptainLex!~CaptainLe@12.104.244.5> has joined ##vulkan[17:57:26] *** CaptainLex <CaptainLex!~CaptainLe@12.104.244.5> has quit IRC (Ping timeout: 252 seconds)[18:00:35] *** kuldeep <kuldeep!~kuldeepdh@unaffiliated/kuldeepdhaka> has quit IRC (Remote host closed the connection)[18:01:09] *** kuldeep <kuldeep!~kuldeepdh@unaffiliated/kuldeepdhaka> has joined ##vulkan[18:01:32] *** CaptainLex <CaptainLex!~CaptainLe@12.104.244.5> has joined ##vulkan[18:04:54] *** ravior <ravior!~crapitea@5-13-193-187.residential.rdsnet.ro> has joined ##vulkan[18:21:10] *** Joefish <Joefish!~Joefish@p200300C6F3CBF1009A05AA2309BCFF95.dip0.t-ipconnect.de> has left ##vulkan ("WeeChat 1.7")[18:31:35] *** realitix <realitix!~realitix@92.103.166.6> has quit IRC (Ping timeout: 240 seconds)[18:34:17] *** nsf <nsf!~nsf@jiss.convex.ru> has joined ##vulkan[18:42:49] *** ratchetfreak <ratchetfreak!c18021f8@gateway/web/freenode/ip.193.128.33.248> has quit IRC (Ping timeout: 260 seconds)[18:51:20] *** Hunts <Hunts!~hunts@dhcp9.eps.hw.ac.uk> has quit IRC (Ping timeout: 252 seconds)[18:52:22] *** glYoda <glYoda!~MTLYoda@c-71-236-230-251.hsd1.or.comcast.net> has quit IRC (Ping timeout: 240 seconds)[18:56:01] *** glYoda <glYoda!~MTLYoda@c-71-236-230-251.hsd1.or.comcast.net> has joined ##vulkan[19:12:17] *** slidercrank <slidercrank!~slidercra@unaffiliated/slidercrank> has quit IRC (Quit: Leaving)[19:15:13] *** realitix <realitix!~realitix@80.12.38.163> has joined ##vulkan[19:19:08] *** realitix <realitix!~realitix@80.12.38.163> has quit IRC (Ping timeout: 240 seconds)[19:21:40] *** ratchetfreak <ratchetfreak!~ratchetfr@ptr-82s3g7omx242auk5xjq.18120a2.ip6.access.telenet.be> has joined ##vulkan[19:22:02] *** ratchetfreak <ratchetfreak!~ratchetfr@ptr-82s3g7omx242auk5xjq.18120a2.ip6.access.telenet.be> has quit IRC (Read error: Connection reset by peer)[19:22:19] *** jkilpatr <jkilpatr!jkilpatr@nat/redhat/x-irzqpurtdgsejmoh> has quit IRC (Quit: Leaving)[19:22:25] *** ratchetfreak <ratchetfreak!~ratchetfr@ptr-82s3g7omx242auk5xjq.18120a2.ip6.access.telenet.be> has joined ##vulkan[19:22:52] *** jkilpatr <jkilpatr!jkilpatr@nat/redhat/x-pfwcmgaknckxurhf> has joined ##vulkan[19:25:10] *** kevr <kevr!~kevr@unaffiliated/kevr> has quit IRC (Ping timeout: 240 seconds)[19:36:59] *** elect <elect!~elect@business-092-079-135-041.static.arcor-ip.net> has quit IRC (Ping timeout: 260 seconds)[19:42:29] *** kevr <kevr!~kevr@unaffiliated/kevr> has joined ##vulkan[19:57:31] *** tacchinotacchi <tacchinotacchi!~tacchinot@net-188-152-21-183.cust.dsl.teletu.it> has quit IRC (Remote host closed the connection)[20:06:11] *** derhass <derhass!~derhass@ipservice-092-216-126-046.092.216.pools.vodafone-ip.de> has joined ##vulkan[20:09:26] *** tacchinotacchi <tacchinotacchi!~tacchinot@net-188-152-21-183.cust.vodafonedsl.it> has joined ##vulkan[20:18:04] *** David3k <David3k!~David3k@130.105.229.199> has joined ##vulkan[20:32:11] *** slidercrank <slidercrank!~slidercra@unaffiliated/slidercrank> has joined ##vulkan[20:42:17] *** jkilpatr <jkilpatr!jkilpatr@nat/redhat/x-pfwcmgaknckxurhf> has quit IRC (Quit: Leaving)[20:42:35] *** jkilpatr <jkilpatr!jkilpatr@nat/redhat/x-fwmkirndxngppccb> has joined ##vulkan[20:46:09] *** jkilpatr <jkilpatr!jkilpatr@nat/redhat/x-fwmkirndxngppccb> has quit IRC (Client Quit)[20:46:50] *** jkilpatr <jkilpatr!jkilpatr@nat/redhat/x-jyfsnwgyvadldwof> has joined ##vulkan[21:03:24] *** bjz <bjz!~bjz@14-201-122-110.static.tpgi.com.au> has joined ##vulkan[21:10:06] *** GMpow2 <GMpow2!~5@2a02:810a:83c0:cb14:f9b6:2120:c731:85bd> has joined ##vulkan[21:28:18] *** tambre <tambre!~tambre@a5be-f20e-d303-198d-0e80-8a31-07d0-2001.dyn.estpak.ee> has quit IRC (Ping timeout: 258 seconds)[21:40:12] *** ravior <ravior!~crapitea@5-13-193-187.residential.rdsnet.ro> has quit IRC (Remote host closed the connection)[21:40:53] *** ravior <ravior!~crapitea@5-13-193-187.residential.rdsnet.ro> has joined ##vulkan[22:17:32] *** bjz <bjz!~bjz@14-201-122-110.static.tpgi.com.au> has quit IRC (Quit: My MacBook Pro has gone to sleep. ZZZzzz…)[22:18:41] *** sla_ro|master <sla_ro|master!~sla.ro@95.76.45.217> has quit IRC ()[22:21:25] *** afl_ext <afl_ext!~afl_ext3@unaffiliated/afl-ext/x-2796036> has quit IRC (Quit: Leaving)[22:22:20] *** john1 <john1!~john@host86-147-123-91.range86-147.btcentralplus.com> has quit IRC (Ping timeout: 276 seconds)[22:27:16] *** toor_ <toor_!~toor@76.ip-51-254-204.eu> has quit IRC (Ping timeout: 240 seconds)[22:32:10] *** drewbarbs <drewbarbs!~drewbarbs@c-69-143-7-106.hsd1.va.comcast.net> has quit IRC (Ping timeout: 264 seconds)[22:34:25] *** jkilpatr <jkilpatr!jkilpatr@nat/redhat/x-jyfsnwgyvadldwof> has quit IRC (Ping timeout: 255 seconds)[22:34:29] *** drewbarbs <drewbarbs!~drewbarbs@2601:141:202:d351:5893:d352:985:1174> has joined ##vulkan[22:39:54] *** toor_ <toor_!~toor@eth-east-parth2-46-193-65-52.wb.wifirst.net> has joined ##vulkan[22:43:42] *** bjz <bjz!~bjz@14-201-122-110.static.tpgi.com.au> has joined ##vulkan[22:46:19] *** toor_ <toor_!~toor@eth-east-parth2-46-193-65-52.wb.wifirst.net> has quit IRC (Ping timeout: 268 seconds)[22:46:34] *** charlie5 <charlie5!~rod@202-159-169-118.dyn.iinet.net.au> has quit IRC (Quit: Leaving.)[22:53:12] *** bjz <bjz!~bjz@14-201-122-110.static.tpgi.com.au> has quit IRC (Quit: My MacBook Pro has gone to sleep. ZZZzzz…)[22:58:52] *** toor_ <toor_!~toor@76.ip-51-254-204.eu> has joined ##vulkan[23:05:09] *** nsf <nsf!~nsf@jiss.convex.ru> has quit IRC (Quit: WeeChat 1.7)[23:10:53] *** jkilpatr <jkilpatr!~jkilpatr@2602:30a:c7da:b600:1202:b5ff:fe3b:d41d> has joined ##vulkan[23:24:49] *** ImQ009 <ImQ009!~ImQ009@unaffiliated/imq009> has quit IRC (Quit: Leaving)[23:37:45] *** rpb <rpb!~rpb@199.68.195.102> has joined ##vulkan[23:41:58] *** Feedz <Feedz!~Feedz@unaffiliated/feedz> has quit IRC (Quit: Leaving)[23:44:07] *** rpb <rpb!~rpb@199.68.195.102> has quit IRC (Remote host closed the connection)[23:46:45] *** rpb <rpb!~rpb@199.68.195.102> has joined ##vulkan[23:49:10] *** glYoda <glYoda!~MTLYoda@c-71-236-230-251.hsd1.or.comcast.net> has quit IRC (Quit: glYoda)