diff --git a/archive/threads/NOTimothyLottes/2022-12-15-amds-windows-drivers-are-often-unfortunately/2022-12-15-matiasgoldberg-have-you-tried-setting-the-resolution-from-device.md b/archive/threads/NOTimothyLottes/2022-12-15-amds-windows-drivers-are-often-unfortunately/2022-12-15-matiasgoldberg-have-you-tried-setting-the-resolution-from-device.md new file mode 100644 index 0000000..4f93ba5 --- /dev/null +++ b/archive/threads/NOTimothyLottes/2022-12-15-amds-windows-drivers-are-often-unfortunately/2022-12-15-matiasgoldberg-have-you-tried-setting-the-resolution-from-device.md @@ -0,0 +1,89 @@ +--- +title: "@NOTimothyLottes Have you tried setting the resolution from Device Manager-> Monitor Name-> Properties ?" +type: archive +source: twitter +source_url: "https://x.com/matiasgoldberg/status/1603482063436406788" +author: "Matías N. Goldberg" +handle: matiasgoldberg +post_id: "1603482063436406788" +date: 2022-12-15 +archived: 2026-08-26 +draft: false +tags: + - archive + - twitter + - NOTimothyLottes +description: "@NOTimothyLottes Have you tried setting the resolution from Device Manager-> Monitor Name-> Properties ?" +in_reply_to: "" +parent_post_id: "1603479655574642688" +--- + +## Source + +- URL: https://x.com/matiasgoldberg/status/1603482063436406788 +- Author: Matías N. Goldberg (@matiasgoldberg) +- Posted: 2022-12-15 20:08:09 + +## Branch + +**1/** **@matiasgoldberg** ^1603482063436406788 + +**@NOTimothyLottes** + +Have you tried setting the resolution from Device Manager-> Monitor Name-> Properties ? + +Via normal ways Windows/Amd decides to always use 1080p signal and upscale. + +But on that old dialog it actually respects the resolution/hz I set + +**2/** **@NOTimothyLottes** ^1603484063230316544 + +**@matiasgoldberg** + +I get 'Generic PnP Monitor' there and no properity that seems to apply towards setting frequency limits or modes. + +**3/** **@matiasgoldberg** ^1603491500171616281 + +**@NOTimothyLottes** + +OK I was trying to recall from memory. I am at computer now. + +Go to Advanced Display -> Display Properties for + +Then Adapter -> List All Modes + +Select one you DON'T want -> OK -> Apply; +then List All Modes again, select the one you want -> OK -> Apply + +![](https://pbs.twimg.com/media/FkC-_hpWYB82qEx?format=png&name=orig) +![](https://pbs.twimg.com/media/FkC_BJhWYBQwma-?format=jpg&name=orig) + +**4/** **@NOTimothyLottes** ^1603492205229924374 + +**@matiasgoldberg** + +The mode I'm building isn't standard, so it never shows up in the list. But thanks for the tip anyway. + +**5/** **@NOTimothyLottes** ^1603493252417609740 + +**@matiasgoldberg** + +Even using the "Custom Resolution Utility" to override and place in the correct display V&H ranges into the display registry data, with a reboot, still won't fix the problem with AMD's drivers. It still just won't accept the mode. This is latest AMD driver too. + +![](https://pbs.twimg.com/media/FkC_wsCWYBUIMFF?format=png&name=orig) + +**6/** **@NOTimothyLottes** ^1603494786010341379 + +**@matiasgoldberg** + +Related, one of the amazing things possible on even current NVIDIA drivers is to run 240p 120Hz modes on standard old 31KHz VGA CRTs for perfect scanlines for vintage gaming. Also won't work on AMD due to this driver bug. Same trick as used on MiSTer here: https://www.retrorgb.com/mister-240p-120hz-on-a-vga-crt-monitor.html + +**7/** **@NOTimothyLottes** ^1603522358475726849 + +**@matiasgoldberg** + +Ok here we go. Lets see how long it takes to get a one line fix in an AMD public driver. Bug filed same as a standard user would file a bug, here on the AMD forums: https://community.amd.com/t5/drivers-software/amd-driver-quot-custom-resolutions-quot-broken-please-fix-it-s-a/m-p/567137/highlight/true#M165847 + +## Related + +- Spine: [[archive/threads/NOTimothyLottes/2022-12-15-amds-windows-drivers-are-often-unfortunately]] diff --git a/archive/threads/NOTimothyLottes/2022-12-15-amds-windows-drivers-are-often-unfortunately/index.md b/archive/threads/NOTimothyLottes/2022-12-15-amds-windows-drivers-are-often-unfortunately/index.md new file mode 100644 index 0000000..6c90f24 --- /dev/null +++ b/archive/threads/NOTimothyLottes/2022-12-15-amds-windows-drivers-are-often-unfortunately/index.md @@ -0,0 +1,54 @@ +--- +title: "AMD's windows drivers are often unfortunately broken for CRT users! Even after clicking legal blurb 'allow custom resolutions', when creating a modeline which works on CRT (works with an NVIDIA GPU too), AMD's driver fails with this." +type: archive +source: twitter +source_url: "https://x.com/NOTimothyLottes/status/1603479655574642688" +author: "NOTimothyLottes" +handle: NOTimothyLottes +post_id: "1603479655574642688" +date: 2022-12-15 +archived: 2026-08-26 +draft: false +tags: + - archive + - twitter + - NOTimothyLottes +description: "AMD's windows drivers are often unfortunately broken for CRT users! Even after clicking legal blurb 'allow custom resolutions', when creating a modeline which works on CRT (works with an NVIDIA GPU too), AMD's driver fails with this." +in_reply_to: "" +--- + +## Source + +- URL: https://x.com/NOTimothyLottes/status/1603479655574642688 +- Author: NOTimothyLottes (@NOTimothyLottes) +- Posted: 2022-12-15 19:58:35 + +## Thread + +**1/** **@NOTimothyLottes** ^1603479655574642688 + +AMD's windows drivers are often unfortunately broken for CRT users! Even after clicking legal blurb 'allow custom resolutions', when creating a modeline which works on CRT (works with an NVIDIA GPU too), AMD's driver fails with this. TRIVIAL FIX AMD, JUST DO WHAT THE USER ASKS! + +![](https://pbs.twimg.com/media/FkCyXNnWYA4BHve?format=png&name=orig) + +Branches: [[archive/threads/NOTimothyLottes/2022-12-15-amds-windows-drivers-are-often-unfortunately/2022-12-15-matiasgoldberg-have-you-tried-setting-the-resolution-from-device]] + +**2/** **@NOTimothyLottes** ^1603480412105367554 + +Checking the modeline against whatever windows returns won't work, because HDMI to VGA converters don't always return the right DCC info. + +**3/** **@NOTimothyLottes** ^1603480608818225160 + +Very frustrating to be forced to switch to NVIDIA all the time simply to use low latency displays at the desired resolution. + +**4/** **@DomWitczak** ^1603740961951174656 + +**@NOTimothyLottes** + +I'll take this to the display team but it's going to take some time before anything actually happens on the driver's end, with the holiday break round the corner. + +**5/** **@NOTimothyLottes** ^1603741952419176449 + +**@DomWitczak** + +Thanks. And yes, the greatly appreciated holiday tech shutdown is upon some of us already :) diff --git a/archive/threads/NOTimothyLottes/index.md b/archive/threads/NOTimothyLottes/index.md index 51ba7ae..3aa856d 100644 --- a/archive/threads/NOTimothyLottes/index.md +++ b/archive/threads/NOTimothyLottes/index.md @@ -14,6 +14,7 @@ tags: - [[archive/threads/NOTimothyLottes/2022-11-08-example-of-bind-everything-once-layout-aliasing]] - [[archive/threads/NOTimothyLottes/2022-11-09-if-only-pc-vendors-with-open-isas]] - [[archive/threads/NOTimothyLottes/2022-11-13-my-prefered-method-of-managing-application-data]] +- [[archive/threads/NOTimothyLottes/2022-12-15-amds-windows-drivers-are-often-unfortunately]] - [[archive/threads/NOTimothyLottes/2022-12-17-rdna3-isa-guide-comments]] - [[archive/threads/NOTimothyLottes/2023-04-28-https-gpuopen-com-learn-rgp-1-15-enhanced-isa]] - [[archive/threads/NOTimothyLottes/2023-05-03-found-some-of-my-old-blogs-thread-looking-back-an]] diff --git a/archive/threads/SebAaltonen/2022-11-22-ive-long-time-talked-about-killing-the-graphics/2022-11-22-AlenL-few-more-decades-and-will-finally-be-back-to.md b/archive/threads/SebAaltonen/2022-11-22-ive-long-time-talked-about-killing-the-graphics/2022-11-22-AlenL-few-more-decades-and-will-finally-be-back-to.md new file mode 100644 index 0000000..c8f510f --- /dev/null +++ b/archive/threads/SebAaltonen/2022-11-22-ive-long-time-talked-about-killing-the-graphics/2022-11-22-AlenL-few-more-decades-and-will-finally-be-back-to.md @@ -0,0 +1,37 @@ +--- +title: "@SebAaltonen Few more decades and will finally be back to where we never should have left from - writing HW registers." +type: archive +source: twitter +source_url: "https://x.com/AlenL/status/1595196334453161984" +author: "Alen Ladavac" +handle: AlenL +post_id: "1595196334453161984" +date: 2022-11-22 +archived: 2026-08-26 +draft: false +tags: + - archive + - twitter + - SebAaltonen +description: "@SebAaltonen Few more decades and will finally be back to where we never should have left from - writing HW registers." +in_reply_to: "" +parent_post_id: "1595091915472764928" +--- + +## Source + +- URL: https://x.com/AlenL/status/1595196334453161984 +- Author: Alen Ladavac (@AlenL) +- Posted: 2022-11-22 23:23:38 + +## Branch + +**1/** **@AlenL** ^1595196334453161984 + +**@SebAaltonen** + +Few more decades and will finally be back to where we never should have left from - writing HW registers. 👍 + +## Related + +- Spine: [[archive/threads/SebAaltonen/2022-11-22-ive-long-time-talked-about-killing-the-graphics]] diff --git a/archive/threads/SebAaltonen/2022-11-22-ive-long-time-talked-about-killing-the-graphics/2022-11-22-PLT_cheater-killing-the-graphics-api-what-does-that-mean-how.md b/archive/threads/SebAaltonen/2022-11-22-ive-long-time-talked-about-killing-the-graphics/2022-11-22-PLT_cheater-killing-the-graphics-api-what-does-that-mean-how.md new file mode 100644 index 0000000..2f415f3 --- /dev/null +++ b/archive/threads/SebAaltonen/2022-11-22-ive-long-time-talked-about-killing-the-graphics/2022-11-22-PLT_cheater-killing-the-graphics-api-what-does-that-mean-how.md @@ -0,0 +1,37 @@ +--- +title: "@SebAaltonen \"killing the graphics api\" - what does that mean? how do you \"kill\" an api?" +type: archive +source: twitter +source_url: "https://x.com/PLT_cheater/status/1595094244204904448" +author: "Multiple Systems" +handle: PLT_cheater +post_id: "1595094244204904448" +date: 2022-11-22 +archived: 2026-08-26 +draft: false +tags: + - archive + - twitter + - SebAaltonen +description: "@SebAaltonen \"killing the graphics api\" - what does that mean? how do you \"kill\" an api?" +in_reply_to: "" +parent_post_id: "1595091915472764928" +--- + +## Source + +- URL: https://x.com/PLT_cheater/status/1595094244204904448 +- Author: Multiple Systems (@PLT_cheater) +- Posted: 2022-11-22 16:37:57 + +## Branch + +**1/** **@PLT_cheater** ^1595094244204904448 + +**@SebAaltonen** + +"killing the graphics api" - what does that mean? how do you "kill" an api? + +## Related + +- Spine: [[archive/threads/SebAaltonen/2022-11-22-ive-long-time-talked-about-killing-the-graphics]] diff --git a/archive/threads/SebAaltonen/2022-11-22-ive-long-time-talked-about-killing-the-graphics/2022-11-22-SebAaltonen-hopefully-in-the-future-the-existing-vulkan.md b/archive/threads/SebAaltonen/2022-11-22-ive-long-time-talked-about-killing-the-graphics/2022-11-22-SebAaltonen-hopefully-in-the-future-the-existing-vulkan.md new file mode 100644 index 0000000..d934613 --- /dev/null +++ b/archive/threads/SebAaltonen/2022-11-22-ive-long-time-talked-about-killing-the-graphics/2022-11-22-SebAaltonen-hopefully-in-the-future-the-existing-vulkan.md @@ -0,0 +1,37 @@ +--- +title: "Hopefully in the future the existing Vulkan descriptor APIs can be removed." +type: archive +source: twitter +source_url: "https://x.com/SebAaltonen/status/1595092429878837253" +author: "Sebastian Aaltonen" +handle: SebAaltonen +post_id: "1595092429878837253" +date: 2022-11-22 +archived: 2026-08-26 +draft: false +tags: + - archive + - twitter + - SebAaltonen +description: "Hopefully in the future the existing Vulkan descriptor APIs can be removed." +in_reply_to: "" +parent_post_id: "1595091915472764928" +--- + +## Source + +- URL: https://x.com/SebAaltonen/status/1595092429878837253 +- Author: Sebastian Aaltonen (@SebAaltonen) +- Posted: 2022-11-22 16:30:45 + +## Branch + +**1/** **@SebAaltonen** ^1595092429878837253 + +Hopefully in the future the existing Vulkan descriptor APIs can be removed. This is so much simpler. Much less API surface and simpler bookkeeping. + +I am really pleased about this API. Best thing that has happened in a long time :) + +## Related + +- Spine: [[archive/threads/SebAaltonen/2022-11-22-ive-long-time-talked-about-killing-the-graphics]] diff --git a/archive/threads/SebAaltonen/2022-11-22-ive-long-time-talked-about-killing-the-graphics/2022-11-22-danoli3-hooley-dooley.md b/archive/threads/SebAaltonen/2022-11-22-ive-long-time-talked-about-killing-the-graphics/2022-11-22-danoli3-hooley-dooley.md new file mode 100644 index 0000000..36086e6 --- /dev/null +++ b/archive/threads/SebAaltonen/2022-11-22-ive-long-time-talked-about-killing-the-graphics/2022-11-22-danoli3-hooley-dooley.md @@ -0,0 +1,37 @@ +--- +title: "@SebAaltonen Hooley Dooley" +type: archive +source: twitter +source_url: "https://x.com/danoli3/status/1595178104607363075" +author: "Danoli3" +handle: danoli3 +post_id: "1595178104607363075" +date: 2022-11-22 +archived: 2026-08-26 +draft: false +tags: + - archive + - twitter + - SebAaltonen +description: "@SebAaltonen Hooley Dooley" +in_reply_to: "" +parent_post_id: "1595091915472764928" +--- + +## Source + +- URL: https://x.com/danoli3/status/1595178104607363075 +- Author: Danoli3 (@danoli3) +- Posted: 2022-11-22 22:11:11 + +## Branch + +**1/** **@danoli3** ^1595178104607363075 + +**@SebAaltonen** + +Hooley Dooley + +## Related + +- Spine: [[archive/threads/SebAaltonen/2022-11-22-ive-long-time-talked-about-killing-the-graphics]] diff --git a/archive/threads/SebAaltonen/2022-11-22-ive-long-time-talked-about-killing-the-graphics/2022-11-22-vingt_2-yes-please.md b/archive/threads/SebAaltonen/2022-11-22-ive-long-time-talked-about-killing-the-graphics/2022-11-22-vingt_2-yes-please.md new file mode 100644 index 0000000..f878f6e --- /dev/null +++ b/archive/threads/SebAaltonen/2022-11-22-ive-long-time-talked-about-killing-the-graphics/2022-11-22-vingt_2-yes-please.md @@ -0,0 +1,37 @@ +--- +title: "@SebAaltonen Yes! Please!!" +type: archive +source: twitter +source_url: "https://x.com/vingt_2/status/1595180307392577536" +author: "Vingt-2" +handle: vingt_2 +post_id: "1595180307392577536" +date: 2022-11-22 +archived: 2026-08-26 +draft: false +tags: + - archive + - twitter + - SebAaltonen +description: "@SebAaltonen Yes! Please!!" +in_reply_to: "" +parent_post_id: "1595091915472764928" +--- + +## Source + +- URL: https://x.com/vingt_2/status/1595180307392577536 +- Author: Vingt-2 (@vingt_2) +- Posted: 2022-11-22 22:19:56 + +## Branch + +**1/** **@vingt_2** ^1595180307392577536 + +**@SebAaltonen** + +Yes! Please!! + +## Related + +- Spine: [[archive/threads/SebAaltonen/2022-11-22-ive-long-time-talked-about-killing-the-graphics]] diff --git a/archive/threads/SebAaltonen/2022-11-22-ive-long-time-talked-about-killing-the-graphics/2022-11-23-JoaoBapt-one-of-the-biggest-roadblocks-when-i-was-trying.md b/archive/threads/SebAaltonen/2022-11-22-ive-long-time-talked-about-killing-the-graphics/2022-11-23-JoaoBapt-one-of-the-biggest-roadblocks-when-i-was-trying.md new file mode 100644 index 0000000..a321daa --- /dev/null +++ b/archive/threads/SebAaltonen/2022-11-22-ive-long-time-talked-about-killing-the-graphics/2022-11-23-JoaoBapt-one-of-the-biggest-roadblocks-when-i-was-trying.md @@ -0,0 +1,75 @@ +--- +title: "@SebAaltonen One of the biggest roadblocks when I was trying to learn Vulkan was the descriptor API." +type: archive +source: twitter +source_url: "https://x.com/JoaoBapt/status/1595315606349307904" +author: "João Baptista 🇧🇷" +handle: JoaoBapt +post_id: "1595315606349307904" +date: 2022-11-23 +archived: 2026-08-26 +draft: false +tags: + - archive + - twitter + - SebAaltonen +description: "@SebAaltonen One of the biggest roadblocks when I was trying to learn Vulkan was the descriptor API." +in_reply_to: "" +parent_post_id: "1595091915472764928" +--- + +## Source + +- URL: https://x.com/JoaoBapt/status/1595315606349307904 +- Author: João Baptista 🇧🇷 (@JoaoBapt) +- Posted: 2022-11-23 07:17:34 + +## Branch + +**1/** **@JoaoBapt** ^1595315606349307904 + +**@SebAaltonen** + +One of the biggest roadblocks when I was trying to learn Vulkan was the descriptor API. I never fully understood them, and tbh the way Direct3D and Metal do is way cleaner and simpler (the argument buffers not that much, but you get it). + +**2/** **@SebAaltonen** ^1595316396346540039 + +**@JoaoBapt** + +Descriptor sets and argument buffers simply abstract a linear chunk of memory (pointer + size) containing descriptors. But the abstractions make this look super complicated. The idea is that you can change the pointer quickly at runtime (to change materials bindings, etc). + +**3/** **@JoaoBapt** ^1595317094375215104 + +**@SebAaltonen** + +Yeah, that part I got, you could even set the root constants and descriptors on D3D or just *plain write to them with the CPU* on the PS5. The argument buffer API is nice as well, though having to ‘useResource’ every one of them seems to me as defeating the point entirely. + +**4/** **@SebAaltonen** ^1595318728840855553 + +**@JoaoBapt** + +You want to use heaps instead of calling useResource separately for each argument buffer. + +https://developer.apple.com/documentation/metal/mtlrendercommandencoder/3043402-useheap?language=objc + +**5/** **@JoaoBapt** ^1595355410210357248 + +**@SebAaltonen** + +Yeah, I forgot that they existed 😅 it’s been being a nightmare to learn all three APIs, especially because I skipped D3D11 (learned OpenGL instead), so it’s been quite a journey. + +**6/** **@SebAaltonen** ^1595359995012517890 + +**@JoaoBapt** + +Yeah. Metal heaps is more similar to Vulkan. Have sub-allocate things manually. Going to be reusing a lot of my memory management code from the Vulkan backend in the Metal backend. + +**7/** **@JoaoBapt** ^1595360142727548928 + +**@SebAaltonen** + +Makes sense. Though in this case you probably can try to repurpose VMA for that? + +## Related + +- Spine: [[archive/threads/SebAaltonen/2022-11-22-ive-long-time-talked-about-killing-the-graphics]] diff --git a/archive/threads/SebAaltonen/2022-11-22-ive-long-time-talked-about-killing-the-graphics/2022-11-23-RandomPedroJ-when-i-started-with-modern-graphics-apis-i-e.md b/archive/threads/SebAaltonen/2022-11-22-ive-long-time-talked-about-killing-the-graphics/2022-11-23-RandomPedroJ-when-i-started-with-modern-graphics-apis-i-e.md new file mode 100644 index 0000000..2d9e0e2 --- /dev/null +++ b/archive/threads/SebAaltonen/2022-11-22-ive-long-time-talked-about-killing-the-graphics/2022-11-23-RandomPedroJ-when-i-started-with-modern-graphics-apis-i-e.md @@ -0,0 +1,37 @@ +--- +title: "@SebAaltonen @SheriefFYI When I started with modern graphics APIs (i.e., Vulkan) I felt like I was not understanding something because of how far it was from the idea of 'close to the metal' that intro readings stated." +type: archive +source: twitter +source_url: "https://x.com/RandomPedroJ/status/1595378159540748294" +author: "Pedro J. Estébanez — Pedrocorp" +handle: RandomPedroJ +post_id: "1595378159540748294" +date: 2022-11-23 +archived: 2026-08-26 +draft: false +tags: + - archive + - twitter + - SebAaltonen +description: "@SebAaltonen @SheriefFYI When I started with modern graphics APIs (i.e., Vulkan) I felt like I was not understanding something because of how far it was from the idea of 'close to the metal' that intro readings stated." +in_reply_to: "" +parent_post_id: "1595091915472764928" +--- + +## Source + +- URL: https://x.com/RandomPedroJ/status/1595378159540748294 +- Author: Pedro J. Estébanez — Pedrocorp (@RandomPedroJ) +- Posted: 2022-11-23 11:26:08 + +## Branch + +**1/** **@RandomPedroJ** ^1595378159540748294 + +**@SebAaltonen** **@SheriefFYI** + +When I started with modern graphics APIs (i.e., Vulkan) I felt like I was not understanding something because of how far it was from the idea of 'close to the metal' that intro readings stated. Descriptor sets/pools were a major point of confusion. + +## Related + +- Spine: [[archive/threads/SebAaltonen/2022-11-22-ive-long-time-talked-about-killing-the-graphics]] diff --git a/archive/threads/SebAaltonen/2022-11-22-ive-long-time-talked-about-killing-the-graphics/2022-11-23-SvetlinTotev-another-vulkan-w.md b/archive/threads/SebAaltonen/2022-11-22-ive-long-time-talked-about-killing-the-graphics/2022-11-23-SvetlinTotev-another-vulkan-w.md new file mode 100644 index 0000000..ec18795 --- /dev/null +++ b/archive/threads/SebAaltonen/2022-11-22-ive-long-time-talked-about-killing-the-graphics/2022-11-23-SvetlinTotev-another-vulkan-w.md @@ -0,0 +1,37 @@ +--- +title: "@SebAaltonen Another Vulkan W." +type: archive +source: twitter +source_url: "https://x.com/SvetlinTotev/status/1595435479574380545" +author: "Svetlin Totev" +handle: SvetlinTotev +post_id: "1595435479574380545" +date: 2022-11-23 +archived: 2026-08-26 +draft: false +tags: + - archive + - twitter + - SebAaltonen +description: "@SebAaltonen Another Vulkan W." +in_reply_to: "" +parent_post_id: "1595091915472764928" +--- + +## Source + +- URL: https://x.com/SvetlinTotev/status/1595435479574380545 +- Author: Svetlin Totev (@SvetlinTotev) +- Posted: 2022-11-23 15:13:54 + +## Branch + +**1/** **@SvetlinTotev** ^1595435479574380545 + +**@SebAaltonen** + +Another Vulkan W. + +## Related + +- Spine: [[archive/threads/SebAaltonen/2022-11-22-ive-long-time-talked-about-killing-the-graphics]] diff --git a/archive/threads/SebAaltonen/2022-11-22-ive-long-time-talked-about-killing-the-graphics/2022-11-23-ThisIsJBernard-finally-was-waiting-on-this-one-now-just-want.md b/archive/threads/SebAaltonen/2022-11-22-ive-long-time-talked-about-killing-the-graphics/2022-11-23-ThisIsJBernard-finally-was-waiting-on-this-one-now-just-want.md new file mode 100644 index 0000000..bf0c4a8 --- /dev/null +++ b/archive/threads/SebAaltonen/2022-11-22-ive-long-time-talked-about-killing-the-graphics/2022-11-23-ThisIsJBernard-finally-was-waiting-on-this-one-now-just-want.md @@ -0,0 +1,37 @@ +--- +title: "@SebAaltonen Finally, was waiting on this one." +type: archive +source: twitter +source_url: "https://x.com/ThisIsJBernard/status/1595255142366216192" +author: "J Bernard" +handle: ThisIsJBernard +post_id: "1595255142366216192" +date: 2022-11-23 +archived: 2026-08-26 +draft: false +tags: + - archive + - twitter + - SebAaltonen +description: "@SebAaltonen Finally, was waiting on this one." +in_reply_to: "" +parent_post_id: "1595091915472764928" +--- + +## Source + +- URL: https://x.com/ThisIsJBernard/status/1595255142366216192 +- Author: J Bernard (@ThisIsJBernard) +- Posted: 2022-11-23 03:17:19 + +## Branch + +**1/** **@ThisIsJBernard** ^1595255142366216192 + +**@SebAaltonen** + +Finally, was waiting on this one. Now just want offsetof support in dxc to ease unpacking data from byte buffers. + +## Related + +- Spine: [[archive/threads/SebAaltonen/2022-11-22-ive-long-time-talked-about-killing-the-graphics]] diff --git a/archive/threads/SebAaltonen/2022-11-22-ive-long-time-talked-about-killing-the-graphics/2022-11-23-Venemo-afaik-that-is-only-fully-true-on-amd-gpus.md b/archive/threads/SebAaltonen/2022-11-22-ive-long-time-talked-about-killing-the-graphics/2022-11-23-Venemo-afaik-that-is-only-fully-true-on-amd-gpus.md new file mode 100644 index 0000000..363f7c8 --- /dev/null +++ b/archive/threads/SebAaltonen/2022-11-22-ive-long-time-talked-about-killing-the-graphics/2022-11-23-Venemo-afaik-that-is-only-fully-true-on-amd-gpus.md @@ -0,0 +1,43 @@ +--- +title: "@SebAaltonen AFAIK that is only fully true on AMD GPUs, everyone else has to go through some pain to implement this extension." +type: archive +source: twitter +source_url: "https://x.com/Venemo/status/1595310115401474048" +author: "Timur Kristóf" +handle: Venemo +post_id: "1595310115401474048" +date: 2022-11-23 +archived: 2026-08-26 +draft: false +tags: + - archive + - twitter + - SebAaltonen +description: "@SebAaltonen AFAIK that is only fully true on AMD GPUs, everyone else has to go through some pain to implement this extension." +in_reply_to: "" +parent_post_id: "1595091915472764928" +--- + +## Source + +- URL: https://x.com/Venemo/status/1595310115401474048 +- Author: Timur Kristóf (@Venemo) +- Posted: 2022-11-23 06:55:45 + +## Branch + +**1/** **@Venemo** ^1595310115401474048 + +**@SebAaltonen** + +AFAIK that is only fully true on AMD GPUs, everyone else has to go through some pain to implement this extension. + +**2/** **@SebAaltonen** ^1595312553776893952 + +**@Venemo** + +Nvidia has also similar scalar instructions and can store descriptors in scalar registers. Their CUDA docs list the new instructions that were added in Turing. Latest 3 generations of Nvidia should be fine. + +## Related + +- Spine: [[archive/threads/SebAaltonen/2022-11-22-ive-long-time-talked-about-killing-the-graphics]] diff --git a/archive/threads/SebAaltonen/2022-11-22-ive-long-time-talked-about-killing-the-graphics/2022-11-23-zhaijialong-this-extension-still-lacks-the-ability-to-create.md b/archive/threads/SebAaltonen/2022-11-22-ive-long-time-talked-about-killing-the-graphics/2022-11-23-zhaijialong-this-extension-still-lacks-the-ability-to-create.md new file mode 100644 index 0000000..e009097 --- /dev/null +++ b/archive/threads/SebAaltonen/2022-11-22-ive-long-time-talked-about-killing-the-graphics/2022-11-23-zhaijialong-this-extension-still-lacks-the-ability-to-create.md @@ -0,0 +1,37 @@ +--- +title: "@SebAaltonen This extension still lacks the ability to create descriptors in shaders from buffer memory." +type: archive +source: twitter +source_url: "https://x.com/zhaijialong/status/1595246365730570240" +author: "zhaijialong" +handle: zhaijialong +post_id: "1595246365730570240" +date: 2022-11-23 +archived: 2026-08-26 +draft: false +tags: + - archive + - twitter + - SebAaltonen +description: "@SebAaltonen This extension still lacks the ability to create descriptors in shaders from buffer memory." +in_reply_to: "" +parent_post_id: "1595091915472764928" +--- + +## Source + +- URL: https://x.com/zhaijialong/status/1595246365730570240 +- Author: zhaijialong (@zhaijialong) +- Posted: 2022-11-23 02:42:26 + +## Branch + +**1/** **@zhaijialong** ^1595246365730570240 + +**@SebAaltonen** + +This extension still lacks the ability to create descriptors in shaders from buffer memory. We need that :) + +## Related + +- Spine: [[archive/threads/SebAaltonen/2022-11-22-ive-long-time-talked-about-killing-the-graphics]] diff --git a/archive/threads/SebAaltonen/2022-11-22-ive-long-time-talked-about-killing-the-graphics/index.md b/archive/threads/SebAaltonen/2022-11-22-ive-long-time-talked-about-killing-the-graphics/index.md new file mode 100644 index 0000000..e2e0155 --- /dev/null +++ b/archive/threads/SebAaltonen/2022-11-22-ive-long-time-talked-about-killing-the-graphics/index.md @@ -0,0 +1,53 @@ +--- +title: "I've long time talked about killing the graphics API." +type: archive +source: twitter +source_url: "https://x.com/SebAaltonen/status/1595091915472764928" +author: "Sebastian Aaltonen" +handle: SebAaltonen +post_id: "1595091915472764928" +date: 2022-11-22 +archived: 2026-08-26 +draft: false +tags: + - archive + - twitter + - SebAaltonen +description: "I've long time talked about killing the graphics API." +in_reply_to: "" +--- + +## Source + +- URL: https://x.com/SebAaltonen/status/1595091915472764928 +- Author: Sebastian Aaltonen (@SebAaltonen) +- Posted: 2022-11-22 16:28:42 + +## Thread + +**1/** **@SebAaltonen** ^1595091915472764928 + +I've long time talked about killing the graphics API. First step: Kill descriptor APIs. Descriptors are just memory, you can put them in buffers, load them in shader and copy/update them in GPU timeline. + +This new Khronos extension does exactly that: +https://www.khronos.org/blog/vk-ext-descriptor-buffer + +Branches: [[archive/threads/SebAaltonen/2022-11-22-ive-long-time-talked-about-killing-the-graphics/2022-11-22-SebAaltonen-hopefully-in-the-future-the-existing-vulkan]], [[archive/threads/SebAaltonen/2022-11-22-ive-long-time-talked-about-killing-the-graphics/2022-11-22-PLT_cheater-killing-the-graphics-api-what-does-that-mean-how]], [[archive/threads/SebAaltonen/2022-11-22-ive-long-time-talked-about-killing-the-graphics/2022-11-22-danoli3-hooley-dooley]], [[archive/threads/SebAaltonen/2022-11-22-ive-long-time-talked-about-killing-the-graphics/2022-11-22-vingt_2-yes-please]], [[archive/threads/SebAaltonen/2022-11-22-ive-long-time-talked-about-killing-the-graphics/2022-11-22-AlenL-few-more-decades-and-will-finally-be-back-to]], [[archive/threads/SebAaltonen/2022-11-22-ive-long-time-talked-about-killing-the-graphics/2022-11-23-zhaijialong-this-extension-still-lacks-the-ability-to-create]], [[archive/threads/SebAaltonen/2022-11-22-ive-long-time-talked-about-killing-the-graphics/2022-11-23-ThisIsJBernard-finally-was-waiting-on-this-one-now-just-want]], [[archive/threads/SebAaltonen/2022-11-22-ive-long-time-talked-about-killing-the-graphics/2022-11-23-Venemo-afaik-that-is-only-fully-true-on-amd-gpus]], [[archive/threads/SebAaltonen/2022-11-22-ive-long-time-talked-about-killing-the-graphics/2022-11-23-JoaoBapt-one-of-the-biggest-roadblocks-when-i-was-trying]], [[archive/threads/SebAaltonen/2022-11-22-ive-long-time-talked-about-killing-the-graphics/2022-11-23-RandomPedroJ-when-i-started-with-modern-graphics-apis-i-e]], [[archive/threads/SebAaltonen/2022-11-22-ive-long-time-talked-about-killing-the-graphics/2022-11-23-SvetlinTotev-another-vulkan-w]] + +**2/** **@NOTimothyLottes** ^1595158732706811905 + +**@SebAaltonen** + +Yes, closer to what many of us wanted from VK from the beginning. Article missed this: DYNAMIC on AMD at least was good because it can be SGPR preload (removes an indirection at runtime). + +**3/** **@NOTimothyLottes** ^1595159948669104128 + +**@SebAaltonen** + +Unfortunatly though I'd expect some are going to abuse the API: use max descriptor size for everything all in one giant array. Will kill cache efficiency. Not sure if this new API supports the alternatives to fix that. + +**4/** **@rianflo** ^1595161710905929728 + +**@NOTimothyLottes** **@SebAaltonen** + +I am using VK_KHR_push_descriptor for now. diff --git a/archive/threads/SebAaltonen/index.md b/archive/threads/SebAaltonen/index.md index adc74cd..0cec890 100644 --- a/archive/threads/SebAaltonen/index.md +++ b/archive/threads/SebAaltonen/index.md @@ -9,6 +9,7 @@ tags: - SebAaltonen --- +- [[archive/threads/SebAaltonen/2022-11-22-ive-long-time-talked-about-killing-the-graphics]] - [[archive/threads/SebAaltonen/2024-04-04-this-is-how-slow-code-feels-like]] - [[archive/threads/SebAaltonen/2024-05-03-why-should-one-be-able-to-code-without]] - [[archive/threads/SebAaltonen/2024-12-18-fixed-all-our-vulkan-bugs]]