{"id":13738,"date":"2026-05-28T10:01:29","date_gmt":"2026-05-28T17:01:29","guid":{"rendered":"https:\/\/devblogs.microsoft.com\/directx\/?p=13738"},"modified":"2026-05-28T12:31:06","modified_gmt":"2026-05-28T19:31:06","slug":"announcing-agilitysdk-721-preview-and-more-shader-model-6-10-features","status":"publish","type":"post","link":"https:\/\/devblogs.microsoft.com\/directx\/announcing-agilitysdk-721-preview-and-more-shader-model-6-10-features\/","title":{"rendered":"Announcing AgilitySDK 721 Preview and an addition to LinAlg in Shader Model 6.10"},"content":{"rendered":"<h2 style=\"text-align: left;\">Overview<\/h2>\n<hr \/>\n<p style=\"text-align: left;\">Today, we are pleased to announce Agility SDK 1.721-preview with new D3D features along with DXC 1.10.2605.4 adding VectorAccumulate to LinAlg in the existing Shader Model 6.10 preview.\u00a0 The previous 1.720-preview and SM 6.10 preview <a href=\"https:\/\/devblogs.microsoft.com\/directx\/shader-model-6-10-agilitysdk-720-preview\/\">blog<\/a> is still relevant, as all of its features are available here too.<\/p>\n<p style=\"text-align: left;\">Here are the new features.\u00a0 More detail is further below as well as download and driver links.<\/p>\n<ul style=\"text-align: left;\">\n<li>LinAlg VectorAccumulate\n<ul style=\"text-align: left;\">\n<li>Addition to existing LinAlg support in SM 6.10 preview<\/li>\n<\/ul>\n<\/li>\n<li>Partial Programs<\/li>\n<li>GUID Texture Layout<\/li>\n<li>UAVs of Depth<\/li>\n<\/ul>\n<h3>Downloads<\/h3>\n<ul style=\"text-align: left;\">\n<li>AgilitySDK 1.721-preview: <a href=\"https:\/\/devblogs.microsoft.com\/directx\/directx12agility\/\">https:\/\/devblogs.microsoft.com\/directx\/directx12agility\/<\/a><\/li>\n<li style=\"text-align: left;\">DXC 1.10.2605.4: <a href=\"https:\/\/github.com\/microsoft\/DirectXShaderCompiler\/releases\">https:\/\/github.com\/microsoft\/DirectXShaderCompiler\/releases<\/a><\/li>\n<\/ul>\n<h3>Drivers<\/h3>\n<p>See <em>Appendix &gt; Feature Support<\/em> for the full table of each feature&#8217;s supported hardware.<\/p>\n<table style=\"width: 99.1509%; height: 96px;\">\n<tbody>\n<tr style=\"height: 24px;\">\n<td style=\"width: 12.7753%; height: 24px;\" width=\"145\"><strong>IHV<\/strong><\/td>\n<td style=\"width: 138.228%; height: 24px;\" width=\"479\"><strong>Driver Link(s)<\/strong><\/td>\n<\/tr>\n<tr style=\"height: 24px;\">\n<td style=\"width: 12.7753%; height: 24px;\" width=\"145\">AMD<\/td>\n<td style=\"width: 138.228%; height: 24px;\" width=\"479\"><a href=\"https:\/\/www.amd.com\/en\/resources\/support-articles\/release-notes\/RN-RAD-MS-AGILITY-SDK-26-10-07-02.html\">AMD Developer Preview Edition 26.10.07.02<\/a><\/td>\n<\/tr>\n<tr style=\"height: 24px;\">\n<td style=\"width: 12.7753%; height: 24px;\" width=\"145\">Intel<\/td>\n<td style=\"width: 138.228%; height: 24px;\" width=\"479\"><span data-olk-copy-source=\"MessageBody\">These features will be supported in an upcoming driver.<\/span><\/td>\n<\/tr>\n<tr style=\"height: 24px;\">\n<td style=\"width: 12.7753%; height: 24px;\" width=\"145\">NVIDIA<\/td>\n<td style=\"width: 138.228%; height: 24px;\" width=\"479\"><span data-olk-copy-source=\"MessageBody\">Contact your developer relations representative for in-development driver access.\u00a0 Partial Programs support will be available in a future driver.<\/span><\/td>\n<\/tr>\n<\/tbody>\n<\/table>\n<p>The WARP software device supports all these features as well:\u00a0<a href=\"https:\/\/www.nuget.org\/packages\/Microsoft.Direct3D.WARP\/1.65535.20-preview\">https:\/\/www.nuget.org\/packages\/Microsoft.Direct3D.WARP\/1.65535.20-preview<\/a><\/p>\n<h2 style=\"text-align: left;\"><span class=\"TextRun SCXW4956687 BCX8\" lang=\"EN-US\" xml:lang=\"EN-US\" data-contrast=\"none\"><span class=\"NormalTextRun SCXW4956687 BCX8\" data-ccp-parastyle=\"heading 1\">Features<\/span><\/span><\/h2>\n<hr \/>\n<h4>HLSL LinAlg VectorAccumulate<\/h4>\n<p><span data-teams=\"true\">The latest preview release of the DXC compiler adds support for the remaining LinAlg operation, VectorAccumulate. It is documented in the <a id=\"menur78a\" class=\"fui-Link ___1q1shib f2hkw1w f3rmtva f1ewtqcl fyind8e f1k6fduh f1w7gpdv fk6fouc fjoy568 figsok6 f1s184ao f1mk8lai fnbmjn9 f1o700av f13mvf36 f1cmlufx f9n3di6 f1ids18y f1tx3yz7 f1deo86v f1eh06m1 f1iescvh fhgqx19 f1olyrje f1p93eir f1nev41a f1h8hb77 f1lqvz6u f10aw75t fsle3fq f17ae5zn\" title=\"https:\/\/github.com\/microsoft\/hlsl-specs\/blob\/main\/proposals\/0035-linalg-matrix.md\" href=\"https:\/\/github.com\/microsoft\/hlsl-specs\/blob\/main\/proposals\/0035-linalg-matrix.md\" target=\"_blank\" rel=\"noreferrer noopener\" aria-label=\"Link LinAlg spec\">LinAlg spec<\/a>. It also resolves several LinAlg preview bugs.<\/span><\/p>\n<p><span data-teams=\"true\">See the full release notes here: <a href=\"https:\/\/github.com\/microsoft\/DirectXShaderCompiler\/releases\/tag\/v1.10.2605.24\">https:\/\/github.com\/microsoft\/DirectXShaderCompiler\/releases\/tag\/v1.10.2605.24<\/a><\/span><\/p>\n<h4>Partial Programs<\/h4>\n<p>Some game titles have a huge number of unique PSO combinations that they need to be compiled before the game loads to avoid stutter. Precompiling these PSO heavy titles in advance for many different hardware configurations for distribution through advanced shader delivery would take a significant amount of time and create duplicate effort. To address this, we are creating partial graphics programs. Partial graphics programs split the pipeline creation into two steps: create partial pre-rasterization and pixel shader programs containing common state used by different graphics pipelines, then link them together in a generic program with other state.<\/p>\n<p>See this dedicated blog post for more about this feature: <a href=\"https:\/\/devblogs.microsoft.com\/directx\/partial-graphics-programs\/\">https:\/\/devblogs.microsoft.com\/directx\/partial-graphics-programs\/<\/a><\/p>\n<h4>GUID Texture Layout<\/h4>\n<p><span data-teams=\"true\">D3D12 GUID Texture Layout is an extensible mechanism that lets applications specify exact texture memory layouts via standardized public or vendor-specific private GUIDs (extending D3D12_TEXTURE_LAYOUT), enabling predictable, contiguous, CPU-accessible layouts (e.g., row-major NV12) that improve performance and power efficiency in scenarios like camera\/video processing pipelines.<\/span><\/p>\n<p>Spec here: <a href=\"https:\/\/github.com\/microsoft\/DirectX-Specs\/tree\/master\/d3d\/D3D12GUIDTextureLayout.md\">https:\/\/github.com\/microsoft\/DirectX-Specs\/tree\/master\/d3d\/D3D12GUIDTextureLayout.md<\/a><\/p>\n<h4>UAVs of Depth<\/h4>\n<p><span data-teams=\"true\">With this Agility SDK update and supporting driver you can now create Unordered Access Views (UAV) of Depth Stencil resources, enabling direct read\/write access to both the depth and stencil planes.<\/span><\/p>\n<p>Spec here:\u00a0<a href=\"https:\/\/github.com\/microsoft\/DirectX-Specs\/tree\/master\/d3d\/D3D12UAVofDepthBuffer.md\">https:\/\/github.com\/microsoft\/DirectX-Specs\/tree\/master\/d3d\/D3D12UAVofDepthBuffer.md<\/a><\/p>\n<h4>DXDump Files<\/h4>\n<p><span data-teams=\"true\">This release includes some additional functionality for DirectX Dump Files, but there is no way to use it just yet. Stay tuned for updates about this in the coming weeks.<\/span><\/p>\n<h4><\/h4>\n<h2 style=\"text-align: left;\">PIX<\/h2>\n<hr \/>\n<p style=\"text-align: left;\">PIX supports all features released here. See the PIX release blog:\u00a0<a href=\"https:\/\/devblogs.microsoft.com\/pix\/pix-2605-28-preview\/\">https:\/\/devblogs.microsoft.com\/pix\/pix-2605-28-preview\/<\/a><\/p>\n<h2 style=\"text-align: left;\">Appendix<\/h2>\n<hr \/>\n<h4><strong>Feature Support<\/strong><\/h4>\n<p>Given drivers described in <em>Overview<\/em> &gt; <em>Drivers<\/em>:<\/p>\n<table style=\"width: 99.9182%; height: 216px;\">\n<tbody>\n<tr style=\"height: 24px;\">\n<td style=\"width: 19.755%; height: 24px;\" width=\"125\"><\/td>\n<td style=\"width: 25.5743%; height: 24px;\" width=\"163\"><strong>AMD<\/strong><\/td>\n<td style=\"width: 25.7274%; height: 24px;\" width=\"164\"><strong>Intel<\/strong><\/td>\n<td style=\"width: 78.073%; height: 24px;\" width=\"172\"><strong>NVIDIA<\/strong><\/td>\n<\/tr>\n<tr style=\"height: 24px;\">\n<td style=\"width: 19.755%; height: 24px;\" width=\"125\">LinAlg VectorAccumulate<\/td>\n<td style=\"width: 25.5743%; height: 24px;\" width=\"163\"><span data-olk-copy-source=\"MessageBody\">AMD Radeon\u2122 RX 9000 series hardware<\/span><\/td>\n<td style=\"width: 25.7274%; height: 24px;\" width=\"164\"><span data-olk-copy-source=\"MessageBody\">Xe2 or newer hardware (in a future driver)<\/span><\/td>\n<td style=\"width: 78.073%; height: 24px;\" width=\"172\">All RTX hardware<\/td>\n<\/tr>\n<tr style=\"height: 48px;\">\n<td style=\"width: 19.755%; height: 48px;\" width=\"125\">Partial Programs<\/td>\n<td style=\"width: 25.5743%; height: 48px;\" width=\"163\"><span data-olk-copy-source=\"MessageBody\">AMD Radeon\u2122 RX 7000 and 9000 series hardware<\/span><\/td>\n<td style=\"width: 25.7274%; height: 48px;\" width=\"164\"><span data-olk-copy-source=\"MessageBody\">Xe2 or newer hardware (in a future driver)<\/span><\/td>\n<td style=\"width: 78.073%; height: 48px;\" width=\"172\">All RTX hardware (in a future driver)<\/td>\n<\/tr>\n<tr style=\"height: 48px;\">\n<td style=\"width: 19.755%; height: 48px;\" width=\"125\">GUID Texture Layout<\/td>\n<td style=\"width: 25.5743%; height: 48px;\" width=\"163\"><span data-olk-copy-source=\"MessageBody\">AMD Radeon\u2122 RX 7000 and 9000 series hardware<\/span><\/td>\n<td style=\"width: 25.7274%; height: 48px;\" width=\"164\"><span data-olk-copy-source=\"MessageBody\">Xe2 or newer hardware (in a future driver)<\/span><\/td>\n<td style=\"width: 78.073%; height: 48px;\" width=\"172\">All RTX hardware<\/td>\n<\/tr>\n<tr style=\"height: 48px;\">\n<td style=\"width: 19.755%; height: 48px;\" width=\"125\">UAVs of Depth<\/td>\n<td style=\"width: 25.5743%; height: 48px;\" width=\"163\"><span data-olk-copy-source=\"MessageBody\">AMD Radeon\u2122 RX 7000 and 9000 series hardware<\/span><\/td>\n<td style=\"width: 25.7274%; height: 48px;\" width=\"164\"><span data-olk-copy-source=\"MessageBody\">Xe2 or newer hardware (in a future driver)<\/span><\/td>\n<td style=\"width: 78.073%; height: 48px;\" width=\"172\">All RTX hardware<\/td>\n<\/tr>\n<\/tbody>\n<\/table>\n<p>The WARP software device supports all these features as well.<\/p>\n<h2 style=\"text-align: left;\"><\/h2>\n","protected":false},"excerpt":{"rendered":"<p>Overview Today, we are pleased to announce Agility SDK 1.721-preview with new D3D features along with DXC 1.10.2605.4 adding VectorAccumulate to LinAlg in the existing Shader Model 6.10 preview.\u00a0 The previous 1.720-preview and SM 6.10 preview blog is still relevant, as all of its features are available here too. Here are the new features.\u00a0 More [&hellip;]<\/p>\n","protected":false},"author":8584,"featured_media":12651,"comment_status":"open","ping_status":"closed","sticky":false,"template":"","format":"standard","meta":{"_acf_changed":false,"footnotes":""},"categories":[1],"tags":[],"class_list":["post-13738","post","type-post","status-publish","format-standard","has-post-thumbnail","hentry","category-directx"],"acf":[],"blog_post_summary":"<p>Overview Today, we are pleased to announce Agility SDK 1.721-preview with new D3D features along with DXC 1.10.2605.4 adding VectorAccumulate to LinAlg in the existing Shader Model 6.10 preview.\u00a0 The previous 1.720-preview and SM 6.10 preview blog is still relevant, as all of its features are available here too. Here are the new features.\u00a0 More [&hellip;]<\/p>\n","_links":{"self":[{"href":"https:\/\/devblogs.microsoft.com\/directx\/wp-json\/wp\/v2\/posts\/13738","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/devblogs.microsoft.com\/directx\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/devblogs.microsoft.com\/directx\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/devblogs.microsoft.com\/directx\/wp-json\/wp\/v2\/users\/8584"}],"replies":[{"embeddable":true,"href":"https:\/\/devblogs.microsoft.com\/directx\/wp-json\/wp\/v2\/comments?post=13738"}],"version-history":[{"count":2,"href":"https:\/\/devblogs.microsoft.com\/directx\/wp-json\/wp\/v2\/posts\/13738\/revisions"}],"predecessor-version":[{"id":13805,"href":"https:\/\/devblogs.microsoft.com\/directx\/wp-json\/wp\/v2\/posts\/13738\/revisions\/13805"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/devblogs.microsoft.com\/directx\/wp-json\/wp\/v2\/media\/12651"}],"wp:attachment":[{"href":"https:\/\/devblogs.microsoft.com\/directx\/wp-json\/wp\/v2\/media?parent=13738"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/devblogs.microsoft.com\/directx\/wp-json\/wp\/v2\/categories?post=13738"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/devblogs.microsoft.com\/directx\/wp-json\/wp\/v2\/tags?post=13738"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}