Skip to main content

Browse the archive

https://github.com/shader-slang/slang
31 July 2024, 18:01:06 UTC
sort by:
Revision Author Date Message Commit Date
10470f8 Yong He 25 January 2022, 23:34:34 UTC gfx: Allow null entries in shader table. (#2096) Co-authored-by: Yong He <yhe@nvidia.com> 25 January 2022, 23:34:34 UTC
618acb2 Theresa Foley 25 January 2022, 20:59:52 UTC Add support for HLSL unorm/snorm (#2095) Read/write resource types (what D3D/HLSL often refer to as UAVs) can be broadly categorized based on whether they require an underlying format (e.g., a `DXGI_FORMAT`) for reads, or not. D3D refers to the ones that require a format as "typed" UAVs (even though a `RWStructuredBuffer<MyData>` is clearly "typed" at the HLSL level). Vulkan refers to these cases as "storage images" and "storage texel buffers." Under the D3D model, an application does not have to specify the exact format for a formatted/"typed" UAV in order for loads to work, but it *does* need to specify if an HLSL resource with a declared `float` or vector-of-`float` element type will be backed by data with a `*_UNORM` or `*_SNORM` format. This is where the `unorm` and `snorm` type modifiers come in. Superficially, it might seem that adding this feature to the Slang compiler is "just" a matter of adding the two modifiers, which is easily done with a pair of one-line `syntax` declarations in `core.meta.slang` plus the corresponding AST node types. Unfortunately the superficial view misses the detail that, to date, Slang has not had any support for *type modifiers* at all, and has only supported *declaration modifiers*. The distinction has so far not mattered, even with modifiers like `const` because, e.g., the difference between a "`const` array of `float`" and an "array of `const float`" doesn't really matter. So, adding these two modifiers required introducing a lot of infrastructure along the way. Let's walk through what needed to happen: * As described above, the actual `syntax` was added easily in the Slang stdlib * I added a new subclass of `Modifier` for `TypeModifier`s in the AST, and added the AST nodes for `unorm` and `snorm` as subclasses of that. * In order to syntactically support modifiers applied to types (e.g., `unorm float`), I needed to add a `ModifiedTypeExpr` subclass of `Expr` that represents a base type expression with one or more modifiers applied * The parser needed some subtle new logic. There are two main cases where type modifiers will come up: 1. In contexts where we might be parsing a declaration (e.g., `const unorm float a`), we need to support a list of modifiers that might freely mix type modifiers and "declaration modifiers" which are not intended to apply to types. In this case we need to split the lis tof modifiers into the type-related ones and the declaration-related ones, and attach each subset to the appropriate place. This is very important for features like C-style pointers, where in `static const float* a;`, the `static` modifier applies to the entire declaration of `a`, but the `const` modifier *only* applies to the `float` type specifier, and *not* to the outer pointer type (the actual type of `a`). 2. In contexts where we are not parsing a declaration (e.g., a generic type argument), we need to support a list of modifiers and appy them *all* to the type specifier being parsed, even if some of them might not be appropriate. * While working in the parser I implemented a certain amount of unrelated cleanup for code that was using raw `Modifier*`s to represent lists of modifiers, instead of the purpose-built `Modifiers` type. * The `_parseGenericArg` case needed specific work, because it is an important case in the grammar where we need to parse *either* a type expression or a value exprssion, but cannot easily predict which we will see. The fix implemented for now is to always try to parse modifiers and, if we see any, to assume we are in the type case. Because of the rules for how modifiers in a C-like language inhere to the type specifier (and not necessarily the entire type), we need to refactor some of the type expression parsing routines to support parsing a "suffix" of a type expression. * Note: I decided to be conservative and only make these changes in `_parseGenericArg` because that is place that is *needed* in order for user code with `unorm`/`snorm` to work, but in practice a user could still confuse our parser by using type modifiers as part of a cast (e.g., `x = (unorm float)y;`). While there is currently no reason why a user should want to do this, it *does* suggest that we need to be prepared to see type modifiers in other ambiguous "expression or type?" contexts. We have so far preferred to avoid looking up built-in syntax declarations like modifiers in expression contexts, because we want to allow users to create variable names that might conflict with some of the more surprising modifier keywords in HLSL (e.g., both `triangle` and `sample` are modifier keyword). A nuanced strategy may be required when we get around to closing this gap (which will be needed around when we want full pointer support, since a cast like `(const SomeType*)somePtr` is pretty common). * In semantic checking, we now need a `visitModifiedTypeExpr`, which visits the base expression to produce a `Type` and then checks each of the `Modifier`s attached to it. During this process we need to translate the AST-level `Modifier`s into something that can exist properly in the universe of `Type`s. We introduce a `ModifiedType` subclass of `Type`, distinct from the `ModifiedTypeExpr` subclass of `Expr`. Furthermore, we introduce a `ModifierVal` subclass of `Val`, distinct from `Modifier`/`TypeModifier`. * One unfortunate thing here is that it means we have both, e.g., `UNormModifier` to represent the parsed syntax, and `UNormModifierVal` to represent the `Type`/`Val`-level representation of the same concept. It is quite likely that we are near the point where we can/should consider having two distinct AST representations: one for freshly-parsed ASTs and one for semantically-checked ASTs. The `Type`/`Val` hierarchy clearly belongs to the latter. * No actual semantic checking is currently being applied to the `unorm` and `snorm` modifiers, although we should in principle check that they are only being applied to `float` and vector-of-`float` types. * In an attempt to simplify some of the creation logic and build a tiny bit of reusable infrastructure, I went ahead and added the skeleton of a dedupe-caching system in `ASTBuilder` so that we can easily ensure only a single `UNormModifierVal` and a single `SNormModifierVal` ever get created inside the scope of a single builder. * TODO: Thinking about this, I'm now worried the deduplication does not mean I can make the simplifications I currently do in semantic checking by assuming that any two `UNormModifierVal`s will be pointer-identical. This is because we do not currently (IIRC) have the required "bottleneck" in the compiler where all ASTs get serialized after initial checking, and then deserialized when `import`ed into a downstream module, so that every AST node during a checking step comes from a single `ASTBuilder`. Hmm... * If we can rely on deduplication to do its thing, then the `Val` and `Type` implementations of modifiers can be relatively simple. * TODO: One issue here is that the equality comparison for `ModifiedType` currently checks for the same base type and the same modifiers in the same order. This works for now when we only have a small number of type modifiers and any given type will hae at most one, but in the longer run it relies on us to implement some kind of canonicalization scheme, which would both ensure that between `Modified(T, {A, B})` and `Modified(T, {B, A})` only one is allowed (that is, a canonical ordering on modifiers), and that we do not allow `Modified(Modified(T, {A}), {B})`. * TODO: One other issues is that the `ModifiedType` case does not currently interact correctly with the `as()`-based casting for types (whereas that operation *does* interact in a semantically-correct fashion with `typedef`s). Fixing this issue in a robust way really depends on us re-architecting the `Type` system so that *any* `Type` can have modifiers attached, with modifiers affecting type identity/deduplication. * The key place where `ModifiedType` creates a complication in semantic checking is type conversion/coercion. A user is likely to declare a `RWTexture2D<unorm float>`, fetch from it (producing a value of type `unorm float`) and then assign the result to a `float` variable, prompting for a conversion from `unorm float` to `float` (because they are distinct `Type`s). * We handle this case in the core `_coerce()` operation by checking if either `toType` or `fromType` is a `ModifiedType`. If *either* one is a modified type, we apply logic to check for modifiers that are present on one and not the other. Basically we check which modifiers need to be "dropped" and which need to be "added" during conversion, and validate that these modifiers *can* be dropped/added without creating a semantic error. The only type modifiers we support right now *can* be dropped/added like this, so we are fine. * TODO: When we add more complete pointer support, we could need logic here to validate when casts between, e.g., `const int*` and `int*` should/shouldn't be allowed. * Note: Even opening the door to type modifiers at all creates the same kind of challenges for user-defined generic types (and functions!) since `MyType<int>` and `MyType<const int>` are distinct instantiations in a future where we support `const` as a type modifier. We *may* need to plan to restrict where modified types can be used, so that certain built-in generic types support modified types as arguments, but user-defined types don't (or at least might need to opt-in to get support). * The result of a `_coerce()` that drops/adds modifiers is a `ModifierCastExpr`, which is a kind of no-op AST node that merely expresses that the conversion is allowed and valid. * In IR lowering we currently do the simple thing and translate a `ModifiedType` to a distinct IR node called `AttributedType`. * The change in terminology from "modifier" to "attribute" is to follow the way that these kinds of modifiers best map to the `IRAttr` case in the IR (rather than the `IRDecoration` case). We probably ought to do a careful terminology scrub here, because having this terminology mismatch between IR and AST could be a source of confusion. * TODO: In principle, using `IRAttributedType` creates the same basic problems as using `ModifiedType`: code that is usin `as()` or similar operations to check for a specific subclass of `IRType` may not see the case they were looking for due to use of `IRAttributedType`. * Initially I had hoped to avoid the problem by having the `IRAttr`s be attached directly as operands to an otherwise-ordinary `IRType`. E.g., a lowered `unorm float4` would be an `IRVectorType` with an "extra" operand that is an `IRUNormAttr`, something like: `Vector<Float, 4, UNorm>`. This sounds great (and looks great!), but runs into the problem that it is incompatible with the way we currently represent things like generic type parameters. A generic type parameter `T` is represented as an `IRParam`, and it does *not* make sense to have an additional `IRParam` to represent `const T` or `unorm T`, etc. * The Right Way to solve this stuff at both the AST and IR levels is to avoid passing around bare `Type*` or `IRType*` in general, and instead use a value type that implements the needed policy more directly: something like a `TypeHolder` or `IRTypeHolder` (placeholder name). The `*Holder` type would abstract over the various "wrapper" nodes required to store all the additional data like attributes but, importantly, would *not* allow that extra information to be dropped or lost during operations like casting (e.g., note how the current `Type` implementation of `as()` loses information on `typedef` names, making our error messages slightly worse). This is actually quite similar to how we currently use the `DeclRef<T>` system to allow working with what is *usually* a `T*` under the hood, but in a way that ensures we don't lose track of any generic substitution information. * During C-like code emit we have a process that turns an `IRType` into a chain of declarators as needed to emit a C-like declaration with pointers, arrays, etc. The `IRAttributedType` case needs to get folded into this logic. Basically, when we see an `IRAttributedType` we immediately emit any modifiers that are required to be in a prefix position, then recursively emit the underlying type with an extra layer of declarator that tracks the modifiers, so that we can emit any modifiers that should be placed in a postfix position *after* the type. As a specific example, our C/C++ back-end would want to use the postifx option to handle `const`, because then it can properly emit stuff like `int const * const *` and not the incorrect `const const int**`. * The HLSL emit logic overrides the prefix case for handling type attributes, and uses it to emit `unorm` and `snorm` where they occur. * One unfortunate detail is that (apparently) some downstream HLSL compilers do not allow the `unorm`/`snorm` modifiers to apply to `vector<float, *>` types, even though that should be semantically valid. Instead, they only support `float`, `float2`, `float3`, and `float4` explicitly. To work around this issue, we go ahead and change our HLSL emit logic so that when we encountered 1-to-4 component vectors of `float`, `int`, or `uint` we emit the type name using the typical HLSL shorthand. This is actually a signficicant change in our HLSL output, but it both seemed like a good fix to have anyway, and was also the only obvious way to address the downstream parser shortcomings without a massive kludge. * As a result of this change the `half-texture.slang` test broke, since it was using raw HLSL as the expected output. I changed the test to do a DXIL comparison instead, which is our preferred way of testing cross-compilation behavior (since it is more robust in the face of small changes to our source output). 25 January 2022, 20:59:52 UTC
6528b1c lucy96chen 25 January 2022, 18:26:29 UTC Add implementations for resolveResource() to D3D12 and Vulkan backends (#2094) * Added implementations for resolveResource to both D3D and Vulkan backends * Simple test for resolving a multisampled resource written and confirmed to run successfully for D3D12 * Fixed a bug preventing MSAA from working in Vulkan * Changed test format to RGBA32 Float (because swiftshader) 25 January 2022, 18:26:29 UTC
7cff340 Yong He 21 January 2022, 20:13:23 UTC Add entry-point name override feature. (#2089) Co-authored-by: Yong He <yhe@nvidia.com> 21 January 2022, 20:13:23 UTC
f85bc7a Yong He 21 January 2022, 18:17:39 UTC GFX: seperated ShaderTable. (#2090) 21 January 2022, 18:17:39 UTC
11d2482 lucy96chen 20 January 2022, 00:34:58 UTC Vulkan implementations for copyTexture, copyTextureToBuffer, and textureSubresourceBarrier (#2080) * Added preliminary implementations for Vulkan's copyTexture, copyTextureToBuffer, and textureSubresourceBarrier * Simple copyTexture test working * Expanded test to use textureSubresourceBarrier() to change resource states before copying out to a buffer; Changed copyTextureToBuffer() to assert that only a single mip level is being copied; Test passes on Vulkan only * Fixed an incorrect loop condition in D3D12's textureSubresourceBarrier and changed the size of the results buffer to pre-account for padding; Test runs in D3D12 but does not pass * D3D12 test working, compareComputeResult for buffers now takes an offset into the results * Refactored texture copying tests * Second test written but does not copy correctly * Fixed texture creation in D3D12 to take into account the subresource index when copying texture data so it actually copies all slices instead of just the ones in the first array layer; Second test working on both D3D12 and Vulkan * Added a note for future tests to be added for texture copying; Fixed build errors in CUDA Co-authored-by: Yong He <yhe@nvidia.com> Co-authored-by: Theresa Foley <tfoleyNV@users.noreply.github.com> 20 January 2022, 00:34:58 UTC
b40cd14 Yong He 19 January 2022, 19:14:38 UTC More fixes to GFX d3d12. (#2084) * Fixes to GFX. * Fix binding null resource views in d3d12. * Fix array render target view creation. * Add support for more primitive topologies. * More gfx fixes. * D3D12 feature report on conservative raster, programmable sample position, barycentrics and ROV. * Add QueryPool::reset. * Fix resource setDebugName. * Dynamically expanding GPU descriptor heap. * Render passes without render targets (null frame buffer). * Fix. Co-authored-by: Yong He <yhe@nvidia.com> 19 January 2022, 19:14:38 UTC
3b41c81 Yong He 19 January 2022, 18:15:30 UTC Fixes to GFX. (#2083) * Fix binding null resource views in d3d12. * Fix array render target view creation. * Add support for more primitive topologies. Co-authored-by: Yong He <yhe@nvidia.com> 19 January 2022, 18:15:30 UTC
7dc5857 jsmall-nvidia 18 January 2022, 17:09:31 UTC Fix for issue #2069 (#2082) * #include an absolute path didn't work - because paths were taken to always be relative. * Fix unused initialized variable warning. Fixed typo found in issue 2069 causing g++11 error of access through this. 18 January 2022, 17:09:31 UTC
b2f4cb1 Yong He 14 January 2022, 21:08:46 UTC Various fixes to GFX, nested parameter block test for d3d12. (#2081) * Various fixes. * Add nested parameter block test. * Remove slang-llvm licence info * Ingore slang-llvm/ directory. * Fixup. Co-authored-by: Yong He <yhe@nvidia.com> 14 January 2022, 21:08:46 UTC
d3f904c Yong He 10 January 2022, 21:45:17 UTC Update build status in README 10 January 2022, 21:45:17 UTC
ad9abad Yong He 10 January 2022, 21:30:41 UTC Various fixes to gfx. (#2074) * Various gfx fixes. * Fixup. Co-authored-by: Yong He <yhe@nvidia.com> 10 January 2022, 21:30:41 UTC
0ac1974 Yong He 10 January 2022, 21:16:30 UTC Enable running tests in parallel. (#2078) * Enable running tests in parallel. * Fix linux build. * Add pthread dependency for slang-test. * Fix teamcity output. * Fix race condition. * Make testReporter thread safe. * Clean up. * Fix. * trigger build * Fix. Co-authored-by: Yong He <yhe@nvidia.com> Co-authored-by: Theresa Foley <tfoleyNV@users.noreply.github.com> 10 January 2022, 21:16:30 UTC
36e9276 JKot 10 January 2022, 20:57:16 UTC Fix hexadecimal escape is string literals (#2072) Co-authored-by: Theresa Foley <tfoleyNV@users.noreply.github.com> 10 January 2022, 20:57:16 UTC
4e7694f jsmall-nvidia 10 January 2022, 20:32:34 UTC Test for dynamicDispatch/resource internal error (#2071) * #include an absolute path didn't work - because paths were taken to always be relative. * Test for internal error with resource/dynamic dispatch. * Fix typo. Co-authored-by: Theresa Foley <tfoleyNV@users.noreply.github.com> 10 January 2022, 20:32:34 UTC
aee4aa3 lucy96chen 10 January 2022, 19:00:50 UTC Draw call tests for Vulkan (#2073) * Added instancing support to Vulkan, drawInstanced() test image is upside-down * Fixed inverted drawInstanced() test output by changing Vulkan viewport convention * Replaced vkCmdDraw with vkCmdDrawIndexed in all non-indirect indexed draws, drawIndexedIndirect test now working for Vulkan * Moved index and vertex buffer binding into setIndexBuffer and setVertexBuffers; Defaulted countBuffer to nullptr and countOffset to 0 and added a check for non-null countBuffer to drawIndirect and drawIndexedIndirect in Vulkan; All Vulkan draw tests working (but D3D12 tests broken) * Added support for drawInstanced and drawIndexedInstanced to D3D11 and added tests for both, however D3D11 tests are currently disabled due to readTextureResource assuming a fixed pixel size (among other possible problems); Fixed issues causing D3D12 tests to fail after major back-end fixes to get Vulkan up and running * Removed testing function for dumping images and some other commented out code * Removed some commented out code * Fix initializer list causing builds to fail (attempt 1) * Removed initializer list for VertexStreamDesc in createInputLayout() and fill in struct fields normally (build fix attempt 2) * Removed default values from VertexStreamDesc and changed all initializer lists to reflect this change * Moved applyBinding before setVertexBuffer in RenderTestApp::renderFrame() to ensure the pipeline has already been bound before vertex buffers are * Changed D3D11's readTextureResource to calculate pixel size using format-specific size information; Removed wrapper around D3D11 instanced and indexed instanced draw tests 10 January 2022, 19:00:50 UTC
8919c19 Yong He 09 January 2022, 01:07:33 UTC Fix random d3d12 crash related to readTextureResource(). (#2075) * Fix D3D12 random crash. * Trigger * Trigger * Trigger1 * Trigger2 * Trigger3 * Trigger4 * Trigger5 * Trigger6 * Trigger7 * Trigger8 * Trigger9 * Trigger10 Co-authored-by: Yong He <yhe@nvidia.com> 09 January 2022, 01:07:33 UTC
f591afd Yong He 07 January 2022, 04:44:35 UTC Fix cuda texture creation bug (#2076) Co-authored-by: Yong He <yhe@nvidia.com> 07 January 2022, 04:44:35 UTC
43b6f5c Yong He 04 January 2022, 19:24:11 UTC gfx: Fix root shader object implementation in debug layer. (#2059) * gfx: Fix root shader object implementation in debug layer. * Fix. Co-authored-by: Yong He <yhe@nvidia.com> Co-authored-by: Theresa Foley <tfoleyNV@users.noreply.github.com> 04 January 2022, 19:24:11 UTC
9d6c776 ZanderMajercik 04 January 2022, 19:05:04 UTC Buffer allocation backend. (#2045) * removed initialization of upload resource for CPU visible buffers (D3D12, Vulkan) * reverted slang-gfx.h change * declared DescBase::hasCpuAccessFlag() const to make backend changes compile under gcc * commit before checking master branch * commit before merge * commit before checking master * commit before merging upstream * revert vulkan changes * commit before testing on master * commit before merge with master * reverted bad merge changes * reverted more bad merge changes in render-d3d12.cpp * reverted buffer transition in _uploadBufferData * implemented bufferBarrier() in render-d3d12.cpp * reverted uneccesary transition in createBuffer * create staging buffer even when AccessFlag::None passed to D3D11 createBufferResource * renamed AccessFlags to MemoryType Co-authored-by: Yong He <yonghe@outlook.com> 04 January 2022, 19:05:04 UTC
1a1b2a0 jsmall-nvidia 22 December 2021, 21:50:58 UTC Hotfix/doc typo5 (#2070) * #include an absolute path didn't work - because paths were taken to always be relative. * Fix typos in introduction and conventional features. * Struct inheritance is allowed. Fix some typos. 22 December 2021, 21:50:58 UTC
447b7e0 jsmall-nvidia 21 December 2021, 21:35:34 UTC Language experiments (#2068) * #include an absolute path didn't work - because paths were taken to always be relative. * Moved to experiments. Added some more tests. * More tests around associated types. * Return interface tests. * More tests. 21 December 2021, 21:35:34 UTC
d7ed829 jsmall-nvidia 21 December 2021, 17:11:08 UTC Assorted disabled tests around generics (#2062) * #include an absolute path didn't work - because paths were taken to always be relative. * Some generic experiments. * Add some more generic tests. * More generic experiments/issues. * Some more generic tests. * Remove erroneous test. * Small improvements. * Disable test that was accidentally enabled. * Add equality-2.slang. * Some more generic tests. * Issues around type inference. * Some more generic tests. * Tuple experiment. * Generic interfaces don't seem to be supported. * Add inheritance test. * Alternative array type issue. 21 December 2021, 17:11:08 UTC
48530f4 jsmall-nvidia 21 December 2021, 17:00:22 UTC Hotfix/doc typos4 (#2067) * #include an absolute path didn't work - because paths were taken to always be relative. * Fix some typos. * Fix some more typos. 21 December 2021, 17:00:22 UTC
f813d5f jsmall-nvidia 20 December 2021, 22:52:13 UTC Shader playground improvements (#2066) * #include an absolute path didn't work - because paths were taken to always be relative. * Improve shader playground docs. 20 December 2021, 22:52:13 UTC
407f180 jsmall-nvidia 20 December 2021, 22:25:14 UTC Hotfix/doc typos3 (#2065) * #include an absolute path didn't work - because paths were taken to always be relative. * Fix typos in docs. * Fix some typos. 20 December 2021, 22:25:14 UTC
1efa01f jsmall-nvidia 20 December 2021, 21:56:01 UTC Hotfix/doc typos2 (#2064) * #include an absolute path didn't work - because paths were taken to always be relative. * Fix typos in docs. 20 December 2021, 21:56:01 UTC
3b6fff2 jsmall-nvidia 20 December 2021, 21:25:10 UTC Hotfix/doc typos (#2063) * #include an absolute path didn't work - because paths were taken to always be relative. * Typos in 07-declarations.md 20 December 2021, 21:25:10 UTC
cc709e6 Theresa Foley 17 December 2021, 17:35:23 UTC Cleanup refactoring work around the IR builder (#2061) * Cleanup refactoring work around the IR builder We have some long-term goals for the IR that require a more centralized and disciplined set of rules for how IR instructions get created/emitted. I had been working on trying to set things up so that all IR instruction creation goes through a single bottleneck point, but the non-trivial work in that branch was getting drowned out by the sheer volume of cleanup and refactoring changes. This change tries to pull together several of the more important cleanups. The big pieces are: * `IRBuilder` and `SharedIRBuilder` now protect their data members and rely on users to initialize them more directly via constructor of an `init()` method. This change affects a *bunch* of sites where `IRBuilder`s were created. I changed use sites to use the constructors whenever possible, and to use `init()` in cases where we had longer-lived builders that needed to be initialized multiple times. * The insertion location for the `IRBuilder` now uses an encapsulated type called `IRInsertLoc`. This new type can replace what used to be just two `IRInst*` fields in the builder, and also covers some new functionality (if we ever want to take advantage of it). Very little client code cares about this change, but it is still a nice cleanup in terms of making things more explicit. * The creation of an `IRModule` has been moded *out* of `IRBuilder`, because in practice we `IRBuilder` always wants to be associated with a pre-existing `IRModule` at creation time (via its `SharedIRBuilder`). There is now an `IRModule::create()` operation instead. This required changing the sequencing at many `IRModule` creation sites, since most had been contriving to make an `IRBuilder` first. There were also several cleanups because code had been carelessly using non-reference-counted pointers for `IRModule`s in ways that broke now that `IRModule::create()` always returns a `RefPtr`. * The core operations to actually allocate memory for IR instructions were moved into `IRModule` (since they interact with the memory pool that the module owns). These *were* called `createEmptyInst()` but have been renamed into `_allocateInst()`. In principle these seem like they should only be needed to be called by the `IRBuilder`, but in practice they are also needed by the IR deserialization logic. * A few core operations for emitting IR instructions that were associted with `IRBuilder` were moved to actually be methods on `IRBuilder`. First is `_findOrEmitConstant` which is the primary bottleneck for creating simple scalar constant values. Another is `_createInst` (formerly part of the templated `createInstImpl` along with `createInstWithSizeImpl`) which is the main bottleneck for allocation and initialization of any instruction other than a constant (well, the `IRModuleInst` is the other exception...). Finally, there is also `_maybeSetSourceLoc()`, which is obvious to scope inside the `IRBuilder` once it is protecting the source-location info. Notes: * The `minSizeInBytes` parameter to `_createInst()` might not actually be needed at all. At this point any `IRInst` subtypes that need data allocated for things other than their operands already get created manually via `_allocateInst` or `_findOrEmitConstant`, so I *think* we could remove that part. I will handle that in a subsequent cleanup if it turns out to be the case. * There is one IR pass (`slang-ir-string-hash.cpp`) that is using manual `_allocateInst()` instead of going through an `IRBuilder`. It could be easily cleaned up to not do so (and I will probably make that change down the line), but for now I wanted to avoid doing anything that wasn't close to pure refactoring if I could. * At this point in our design an `IRBuilder` is a very lightweight thing - it basically just owns the insertion location plus a source location to write into instructions. A lot of our code currently treats `IRBuilder`s like they are expensive and/or need to be re-used (which leads to them being used in more mutable/stateful ways). It is quite likely that as we clean up other aspects of the implementation of IR creation/emission we can make `IRBuilder` use feel more lightweight in ways that can streamline and simplify code. * The next step for this work is to identify the different paths that eventually lead to `_createInst()` being called, and unify them at a single bottleneck operation that can own the decisions around when to create an instruction vs. when to re-use an existing one (rather than those decisions being baked into the various `IRBuilder` subroutines that create instructions of the various subtypes). * fixup: gcc/clang C++ spec details 17 December 2021, 17:35:23 UTC
ac88374 lucy96chen 16 December 2021, 22:48:04 UTC gfx: Add tests for instanced, indexed instanced, and indirect draw calls (#2060) * Implemented instancing for the drawInstanced test and confirmed that test values taken at specific pixel location match expected values * Removed writeImage() helper function as this is for debugging only * Factored out shared test code and wrapped tests inside a base class with derived structs for each individual draw call variant; Added a test for indexed, instanced draws but test is currently not working correctly * Indexed instancing test finish and working; Further refactor to move code that fills the test result array and creates the vertex and color buffers to the base test class * Commented out image dump helper function at the top as it may still be needed for the remaining draw tests * Added a new struct derived from the base test class for testing indirect but non-indexed draws; Moved required command signatures for indirect draws into D3D12Device to ensure they continue to exist outside of their respective draw calls; Moved command signature creation into D3D12Device::initialize(); Small consistency cleanup changes * Added working indexed indirect draw call test * Moved expectedResult and compareComputeResult() call into getTestResults() and renamed to checkTestResults() * Rerun tests 16 December 2021, 22:48:04 UTC
f024d72 Yong He 13 December 2021, 20:40:52 UTC Fixes to `uploadTextureData`. (#2058) * Fixes to `uploadTextureData`. * Fix, Co-authored-by: Yong He <yhe@nvidia.com> 13 December 2021, 20:40:52 UTC
7c1ca35 Yong He 13 December 2021, 17:58:29 UTC gfx: Implement `setSamplePositions`. (#2056) * gfx: Implement remaining resource commands on D3D12. Includes: `textureBarrier`, `copyTexture`, `uploadTextureData`, `copyTextureToBuffer`, and `textureSubresourceBarrier`. * gfx: Implement `CurrentSize` query. * gfx: Implement `setSamplePositions`. Co-authored-by: Yong He <yhe@nvidia.com> 13 December 2021, 17:58:29 UTC
3359313 Yong He 13 December 2021, 17:57:56 UTC gfx: use uint32_t in draw calls and implement current size query. (#2055) * gfx: Implement remaining resource commands on D3D12. Includes: `textureBarrier`, `copyTexture`, `uploadTextureData`, `copyTextureToBuffer`, and `textureSubresourceBarrier`. * gfx: Implement `CurrentSize` query. Co-authored-by: Yong He <yhe@nvidia.com> 13 December 2021, 17:57:56 UTC
c1064a2 Yong He 13 December 2021, 17:57:40 UTC gfx: Implement remaining resource commands on D3D12. (#2054) Includes: `textureBarrier`, `copyTexture`, `uploadTextureData`, `copyTextureToBuffer`, and `textureSubresourceBarrier`. Co-authored-by: Yong He <yhe@nvidia.com> 13 December 2021, 17:57:40 UTC
62161b8 lucy96chen 09 December 2021, 23:15:16 UTC Implement instanced and indirect draw calls (#2053) * Added implementations of drawInstanced, drawIndexedInstanced, drawIndirect, and drawIndexedIndirect to both render-d3d12 and render-vk * drawInstanced test WIP * Draw calls implemented, working on debugging test * Added new test and accompanying shader file * Fixes. * Fixes. Co-authored-by: Yong He <yhe@nvidia.com> Co-authored-by: Yong He <yonghe@outlook.com> 09 December 2021, 23:15:16 UTC
1c99a98 Yong He 09 December 2021, 18:46:41 UTC Remove `PipelineType` from gfx header. (#2051) Co-authored-by: Yong He <yhe@nvidia.com> 09 December 2021, 18:46:41 UTC
4ca37fe Yong He 09 December 2021, 17:35:26 UTC gfx: Implement `claerResourceView`. (#2052) 09 December 2021, 17:35:26 UTC
6c08cd9 Yong He 08 December 2021, 23:20:14 UTC gfx Fence implementation improvements. (#2049) 08 December 2021, 23:20:14 UTC
9606401 lucy96chen 08 December 2021, 19:38:14 UTC D3D12 and Vulkan to CUDA Texture Sharing (#2038) 08 December 2021, 19:38:14 UTC
90d8af8 Yong He 07 December 2021, 21:45:49 UTC gfx: D3D12 and VK Fence implementation. (#2048) * gfx: D3D12 and VK Fence implementation. * Fix. * Update project files. * Revert project file changes. * Remove project files Co-authored-by: Yong He <yhe@nvidia.com> 07 December 2021, 21:45:49 UTC
646eecc jsmall-nvidia 07 December 2021, 19:16:28 UTC Check g++ version compatibility (#2044) * #include an absolute path didn't work - because paths were taken to always be relative. * Test gcc >= 5.0 * Disable codegen for reflection tests. * Add parsing options. * Small comment changes to kick CI build. 07 December 2021, 19:16:28 UTC
8b3df74 jsmall-nvidia 07 December 2021, 18:15:23 UTC Output of IR ids as command line option (#2043) * #include an absolute path didn't work - because paths were taken to always be relative. * WIP control of dump options. * Removed SourceManager for IRDumpOptions * Arm aarch64 debug connection timeout - as CI timed out. 07 December 2021, 18:15:23 UTC
9e666a3 jsmall-nvidia 07 December 2021, 17:26:48 UTC Support array args for JSON-RPC (#2046) * #include an absolute path didn't work - because paths were taken to always be relative. * Add ability to send/receive JSON-RPC params optionally as an array. * More array conversions to json-native. * Simplified setting up of 'CallStyle' on JSONRPCConnection. * Small simplification in JSONRPCConnection. * Small improvements around JSON-RPC connection. * Improve some comments. Kick CI build. 07 December 2021, 17:26:48 UTC
ef4ee0b Yong He 06 December 2021, 22:50:42 UTC Fix shader-toy example. (#2047) Co-authored-by: Yong He <yhe@nvidia.com> 06 December 2021, 22:50:42 UTC
5cbd617 Yong He 06 December 2021, 17:06:16 UTC gfx Mutable Root shader object implementation. (#2042) * gfx Mutable Root shader object implementation. * Fix x86 build. Co-authored-by: Yong He <yhe@nvidia.com> 06 December 2021, 17:06:16 UTC
da6be80 jsmall-nvidia 03 December 2021, 16:45:01 UTC Split out ExecutableLocation (#2041) * #include an absolute path didn't work - because paths were taken to always be relative. * Split out ExecutableLocation. * Fixes for changes to ExecutableLocation. * Fix issues around Process on windows. * Improve comments. Kick CI. 03 December 2021, 16:45:01 UTC
f4b86ff jsmall-nvidia 03 December 2021, 15:19:19 UTC U/Int64 coverage on slang-llvm JIT (#2040) * #include an absolute path didn't work - because paths were taken to always be relative. * Update slang-llvm dependencies. 03 December 2021, 15:19:19 UTC
80ff45f jsmall-nvidia 03 December 2021, 14:46:08 UTC Improvements to repro diagnostics (#2039) * #include an absolute path didn't work - because paths were taken to always be relative. * Improvements to repro diagnostics. * Fix typo. 03 December 2021, 14:46:08 UTC
fb6cf46 Yong He 01 December 2021, 22:35:16 UTC Fix reinterpret lowering not done for generic functions. (#2037) Co-authored-by: Yong He <yhe@nvidia.com> 01 December 2021, 22:35:16 UTC
d10ca98 jsmall-nvidia 30 November 2021, 22:09:52 UTC Fix issue around constant folding/bool (#2036) * #include an absolute path didn't work - because paths were taken to always be relative. * Fix bool handling in constant folding for generic parameters. 30 November 2021, 22:09:52 UTC
ce12e1d jsmall-nvidia 30 November 2021, 21:50:05 UTC Auto flush for streams for stdin/out in slang-test (#2035) * #include an absolute path didn't work - because paths were taken to always be relative. * Move StreamType from Process to StdStreamType in slang-stream.h * Disable buffering for stdout/stderr for slang-test. * Improve comment. 30 November 2021, 21:50:05 UTC
ace4e33 jsmall-nvidia 30 November 2021, 21:34:52 UTC Use test-server on CI (#2034) * #include an absolute path didn't work - because paths were taken to always be relative. * Vary what SpawnType is used, if one isn't explicitly set. * Terminate on linux if exec fails. * Use a more sophisticated sleeping mechanism. * Attempt to make CI tests to work on aarch64 debug. Small fixes. 30 November 2021, 21:34:52 UTC
dd18f2b jsmall-nvidia 25 November 2021, 00:37:24 UTC JSON-RPC make id explicit (#2030) * #include an absolute path didn't work - because paths were taken to always be relative. * Use PersistantJSONValue for id storage. * Make id handling explicit - so can make message processing disjoint from receiving order. * Fix some issues on linux with templates. * Fix typo. * Fix call not passing id for JSON-RPC. * Simplify getting persistent id from JSONRPCConnection. 25 November 2021, 00:37:24 UTC
233635c jsmall-nvidia 24 November 2021, 22:07:22 UTC Fix for slang-test memory leak (#2029) * #include an absolute path didn't work - because paths were taken to always be relative. * Fix memory leak due to Rtti usage. 24 November 2021, 22:07:22 UTC
7db340b jsmall-nvidia 24 November 2021, 15:52:11 UTC PersistentJSONValue (#2028) * #include an absolute path didn't work - because paths were taken to always be relative. * Use 'Process' to communicate with an command line tool. * Remove slang-win-stream * Tidy up windows ProcessUtil. * First version of BufferedReadStream. * Windows working IPC for steams. * Test proxy count option. * Split Process/ProcessUtil. Process is platform dependant. ProcessUtil are functions that are platform independent. * First implementation of Unix Process interface. * Unix process compiles on cygwin. * Fix typo in unix process. * Separate unix pipe stream error of invalid access, from pipe availability. * Fix in standard line extraction. * Make fd non blocking. * Fix issues with Windows Process streams. * Added UnixPipe. * Some fixes around UnixPipeStream. * Make a unix stream closed explicit. * Hack to debug linux process/stream. * Revert to old linux pipe handling. * Pass executable path for unit tests. Split out CommandLine into own source. * Small improvements in process/command line. * Check process behavior with crash. * Make stderr and stdout unbuffered for crash testing. * Only turn disable buffering in crash test. * Disable crash test on CI. * Fix crash on clang/linux. * Enable crash test. Remove _appendBuffer as can use StreamUtil functionality. * Added inital processing for http headers. * Small improvements to HttpHeader. * First pass HTTPPacketConnection working on windows. * Enable other Process communication tests. * Update comments. * WIP JSON RPC. * Add terminate to Process. Made JSONRPC a Util. * Small tidy up around HTTPPacketConnection. * Improve process termination options. * WIP for test-server. * Add diagnostics error handling to test-server. * Improved JSON support. Parsing/creating JSON-RPC messages. * WIP JSONRPC parsing. * First pass RttiInfo support. * WIP converting between JSON/native types. * Project files. * Split out RttiUtil. Made RttiInfo constuction thread safe. * WIP RTTI<->JSON. * Add diagnostics to JSON<->native conversions. * Make RttiInfo for structs globals. Avoids problem around derived types (like pointers), being able to cause an abort. * Add pointer support to RTTI. Fixed some compilation issues on linux. * Add fixed array support. * Added Rtti unit test. * Add rtti unit test. * Split out quoted/unquoted key handling. Fix bugs in JSON value/container. Added JSON native test. * Make default array allocator use malloc/free. Remove the new[] handler (doesn't work on visuals studio). * Fix for linux warning. * Remove some test code. * Fix issues on x86 win. * Fix warning on aarch64. * Fix some bugs in JSON parsing/handling. Make Rtti work copy/dtor/ctor struct types. * Testing JSON<->native with fixed array. Make makeArrayView explicit if it's just a single value. Added array type. * Fix getting arrayView. * Improve JSON diagnostic name. * First pass refactor using Rtti for JSON RPC. * First pass of test server using RTTI/JSON-RPC. * Added JSONRPCConnection. * Fix some naming issues. * First pass of test-server working. * Added unit test support for JSON-RPC test server. * Fix compilation issues on linux around template handling. * Typo fix. * Fix a bug around SourceLoc lookup with JSONContainer. * Set the console type to console for ISlangWriters. * Small improvements to test-server. * Small improvements in test-server. * Small fix. * Remove test-proxy. Make test-process a process that can be used to unit test 'Process'. Adding mechanism to control spawning that will create a new process for every test. * Ability to remove source manager for JSONValue. * WIP SimpleJSONValue. * Add PersistentJSONValue * Testing around PersistentJSONValue. Bug fixes. * Small code improvements. 24 November 2021, 15:52:11 UTC
bdc61bb jsmall-nvidia 24 November 2021, 14:34:05 UTC Remove test-proxy (#2027) * #include an absolute path didn't work - because paths were taken to always be relative. * Use 'Process' to communicate with an command line tool. * Remove slang-win-stream * Tidy up windows ProcessUtil. * First version of BufferedReadStream. * Windows working IPC for steams. * Test proxy count option. * Split Process/ProcessUtil. Process is platform dependant. ProcessUtil are functions that are platform independent. * First implementation of Unix Process interface. * Unix process compiles on cygwin. * Fix typo in unix process. * Separate unix pipe stream error of invalid access, from pipe availability. * Fix in standard line extraction. * Make fd non blocking. * Fix issues with Windows Process streams. * Added UnixPipe. * Some fixes around UnixPipeStream. * Make a unix stream closed explicit. * Hack to debug linux process/stream. * Revert to old linux pipe handling. * Pass executable path for unit tests. Split out CommandLine into own source. * Small improvements in process/command line. * Check process behavior with crash. * Make stderr and stdout unbuffered for crash testing. * Only turn disable buffering in crash test. * Disable crash test on CI. * Fix crash on clang/linux. * Enable crash test. Remove _appendBuffer as can use StreamUtil functionality. * Added inital processing for http headers. * Small improvements to HttpHeader. * First pass HTTPPacketConnection working on windows. * Enable other Process communication tests. * Update comments. * WIP JSON RPC. * Add terminate to Process. Made JSONRPC a Util. * Small tidy up around HTTPPacketConnection. * Improve process termination options. * WIP for test-server. * Add diagnostics error handling to test-server. * Improved JSON support. Parsing/creating JSON-RPC messages. * WIP JSONRPC parsing. * First pass RttiInfo support. * WIP converting between JSON/native types. * Project files. * Split out RttiUtil. Made RttiInfo constuction thread safe. * WIP RTTI<->JSON. * Add diagnostics to JSON<->native conversions. * Make RttiInfo for structs globals. Avoids problem around derived types (like pointers), being able to cause an abort. * Add pointer support to RTTI. Fixed some compilation issues on linux. * Add fixed array support. * Added Rtti unit test. * Add rtti unit test. * Split out quoted/unquoted key handling. Fix bugs in JSON value/container. Added JSON native test. * Make default array allocator use malloc/free. Remove the new[] handler (doesn't work on visuals studio). * Fix for linux warning. * Remove some test code. * Fix issues on x86 win. * Fix warning on aarch64. * Fix some bugs in JSON parsing/handling. Make Rtti work copy/dtor/ctor struct types. * Testing JSON<->native with fixed array. Make makeArrayView explicit if it's just a single value. Added array type. * Fix getting arrayView. * Improve JSON diagnostic name. * First pass refactor using Rtti for JSON RPC. * First pass of test server using RTTI/JSON-RPC. * Added JSONRPCConnection. * Fix some naming issues. * First pass of test-server working. * Added unit test support for JSON-RPC test server. * Fix compilation issues on linux around template handling. * Typo fix. * Fix a bug around SourceLoc lookup with JSONContainer. * Set the console type to console for ISlangWriters. * Small improvements to test-server. * Small improvements in test-server. * Small fix. * Remove test-proxy. Make test-process a process that can be used to unit test 'Process'. Adding mechanism to control spawning that will create a new process for every test. * Ability to remove source manager for JSONValue. 24 November 2021, 14:34:05 UTC
9e084ff jsmall-nvidia 23 November 2021, 21:23:15 UTC JSON-RPC test server (#2026) * #include an absolute path didn't work - because paths were taken to always be relative. * Use 'Process' to communicate with an command line tool. * Remove slang-win-stream * Tidy up windows ProcessUtil. * First version of BufferedReadStream. * Windows working IPC for steams. * Test proxy count option. * Split Process/ProcessUtil. Process is platform dependant. ProcessUtil are functions that are platform independent. * First implementation of Unix Process interface. * Unix process compiles on cygwin. * Fix typo in unix process. * Separate unix pipe stream error of invalid access, from pipe availability. * Fix in standard line extraction. * Make fd non blocking. * Fix issues with Windows Process streams. * Added UnixPipe. * Some fixes around UnixPipeStream. * Make a unix stream closed explicit. * Hack to debug linux process/stream. * Revert to old linux pipe handling. * Pass executable path for unit tests. Split out CommandLine into own source. * Small improvements in process/command line. * Check process behavior with crash. * Make stderr and stdout unbuffered for crash testing. * Only turn disable buffering in crash test. * Disable crash test on CI. * Fix crash on clang/linux. * Enable crash test. Remove _appendBuffer as can use StreamUtil functionality. * Added inital processing for http headers. * Small improvements to HttpHeader. * First pass HTTPPacketConnection working on windows. * Enable other Process communication tests. * Update comments. * WIP JSON RPC. * Add terminate to Process. Made JSONRPC a Util. * Small tidy up around HTTPPacketConnection. * Improve process termination options. * WIP for test-server. * Add diagnostics error handling to test-server. * Improved JSON support. Parsing/creating JSON-RPC messages. * WIP JSONRPC parsing. * First pass RttiInfo support. * WIP converting between JSON/native types. * Project files. * Split out RttiUtil. Made RttiInfo constuction thread safe. * WIP RTTI<->JSON. * Add diagnostics to JSON<->native conversions. * Make RttiInfo for structs globals. Avoids problem around derived types (like pointers), being able to cause an abort. * Add pointer support to RTTI. Fixed some compilation issues on linux. * Add fixed array support. * Added Rtti unit test. * Add rtti unit test. * Split out quoted/unquoted key handling. Fix bugs in JSON value/container. Added JSON native test. * Make default array allocator use malloc/free. Remove the new[] handler (doesn't work on visuals studio). * Fix for linux warning. * Remove some test code. * Fix issues on x86 win. * Fix warning on aarch64. * Fix some bugs in JSON parsing/handling. Make Rtti work copy/dtor/ctor struct types. * Testing JSON<->native with fixed array. Make makeArrayView explicit if it's just a single value. Added array type. * Fix getting arrayView. * Improve JSON diagnostic name. * First pass refactor using Rtti for JSON RPC. * First pass of test server using RTTI/JSON-RPC. * Added JSONRPCConnection. * Fix some naming issues. * First pass of test-server working. * Added unit test support for JSON-RPC test server. * Fix compilation issues on linux around template handling. * Typo fix. * Fix a bug around SourceLoc lookup with JSONContainer. * Set the console type to console for ISlangWriters. * Small improvements to test-server. * Small improvements in test-server. * Small fix. 23 November 2021, 21:23:15 UTC
fd46034 Yong He 22 November 2021, 19:24:25 UTC gfx: Add more fixed function states and instancing draw calls. (#2023) * gfx: Add more fixed function states and instancing draw calls. * Fix clang error. * Fix clang. * Fixes. * Add `AccelerationStructureCurrentSize` enum. Co-authored-by: Yong He <yhe@nvidia.com> 22 November 2021, 19:24:25 UTC
87eb789 jsmall-nvidia 19 November 2021, 15:22:56 UTC Improvements to JSON/RTTI (#2022) * #include an absolute path didn't work - because paths were taken to always be relative. * Use 'Process' to communicate with an command line tool. * Remove slang-win-stream * Tidy up windows ProcessUtil. * First version of BufferedReadStream. * Windows working IPC for steams. * Test proxy count option. * Split Process/ProcessUtil. Process is platform dependant. ProcessUtil are functions that are platform independent. * First implementation of Unix Process interface. * Unix process compiles on cygwin. * Fix typo in unix process. * Separate unix pipe stream error of invalid access, from pipe availability. * Fix in standard line extraction. * Make fd non blocking. * Fix issues with Windows Process streams. * Added UnixPipe. * Some fixes around UnixPipeStream. * Make a unix stream closed explicit. * Hack to debug linux process/stream. * Revert to old linux pipe handling. * Pass executable path for unit tests. Split out CommandLine into own source. * Small improvements in process/command line. * Check process behavior with crash. * Make stderr and stdout unbuffered for crash testing. * Only turn disable buffering in crash test. * Disable crash test on CI. * Fix crash on clang/linux. * Enable crash test. Remove _appendBuffer as can use StreamUtil functionality. * Added inital processing for http headers. * Small improvements to HttpHeader. * First pass HTTPPacketConnection working on windows. * Enable other Process communication tests. * Update comments. * WIP JSON RPC. * Add terminate to Process. Made JSONRPC a Util. * Small tidy up around HTTPPacketConnection. * Improve process termination options. * WIP for test-server. * Add diagnostics error handling to test-server. * Improved JSON support. Parsing/creating JSON-RPC messages. * WIP JSONRPC parsing. * First pass RttiInfo support. * WIP converting between JSON/native types. * Project files. * Split out RttiUtil. Made RttiInfo constuction thread safe. * WIP RTTI<->JSON. * Add diagnostics to JSON<->native conversions. * Make RttiInfo for structs globals. Avoids problem around derived types (like pointers), being able to cause an abort. * Add pointer support to RTTI. Fixed some compilation issues on linux. * Add fixed array support. * Added Rtti unit test. * Add rtti unit test. * Split out quoted/unquoted key handling. Fix bugs in JSON value/container. Added JSON native test. * Make default array allocator use malloc/free. Remove the new[] handler (doesn't work on visuals studio). * Fix for linux warning. * Remove some test code. * Fix issues on x86 win. * Fix warning on aarch64. * Fix some bugs in JSON parsing/handling. Make Rtti work copy/dtor/ctor struct types. * Testing JSON<->native with fixed array. Make makeArrayView explicit if it's just a single value. Added array type. * Fix getting arrayView. * Improve JSON diagnostic name. 19 November 2021, 15:22:56 UTC
1d5f815 jsmall-nvidia 18 November 2021, 20:58:12 UTC RTTI/JSON (#2021) * #include an absolute path didn't work - because paths were taken to always be relative. * Use 'Process' to communicate with an command line tool. * Remove slang-win-stream * Tidy up windows ProcessUtil. * First version of BufferedReadStream. * Windows working IPC for steams. * Test proxy count option. * Split Process/ProcessUtil. Process is platform dependant. ProcessUtil are functions that are platform independent. * First implementation of Unix Process interface. * Unix process compiles on cygwin. * Fix typo in unix process. * Separate unix pipe stream error of invalid access, from pipe availability. * Fix in standard line extraction. * Make fd non blocking. * Fix issues with Windows Process streams. * Added UnixPipe. * Some fixes around UnixPipeStream. * Make a unix stream closed explicit. * Hack to debug linux process/stream. * Revert to old linux pipe handling. * Pass executable path for unit tests. Split out CommandLine into own source. * Small improvements in process/command line. * Check process behavior with crash. * Make stderr and stdout unbuffered for crash testing. * Only turn disable buffering in crash test. * Disable crash test on CI. * Fix crash on clang/linux. * Enable crash test. Remove _appendBuffer as can use StreamUtil functionality. * Added inital processing for http headers. * Small improvements to HttpHeader. * First pass HTTPPacketConnection working on windows. * Enable other Process communication tests. * Update comments. * WIP JSON RPC. * Add terminate to Process. Made JSONRPC a Util. * Small tidy up around HTTPPacketConnection. * Improve process termination options. * WIP for test-server. * Add diagnostics error handling to test-server. * Improved JSON support. Parsing/creating JSON-RPC messages. * WIP JSONRPC parsing. * First pass RttiInfo support. * WIP converting between JSON/native types. * Project files. * Split out RttiUtil. Made RttiInfo constuction thread safe. * WIP RTTI<->JSON. * Add diagnostics to JSON<->native conversions. * Make RttiInfo for structs globals. Avoids problem around derived types (like pointers), being able to cause an abort. * Add pointer support to RTTI. Fixed some compilation issues on linux. * Add fixed array support. * Added Rtti unit test. * Add rtti unit test. * Split out quoted/unquoted key handling. Fix bugs in JSON value/container. Added JSON native test. * Make default array allocator use malloc/free. Remove the new[] handler (doesn't work on visuals studio). * Fix for linux warning. * Remove some test code. * Fix issues on x86 win. * Fix warning on aarch64. 18 November 2021, 20:58:12 UTC
b482844 Yong He 18 November 2021, 19:14:29 UTC gfx: add coverage for more resource commands. (#2020) * gfx: specify SubresourceRange for resource view creation. * Add gfx method for `resolveResource`. * Fix compile error. * `copyTextureToBuffer` and `textureSubresourceBarrier`. * Fix vulkan bug. * Fix test cras;h. Co-authored-by: Yong He <yhe@nvidia.com> 18 November 2021, 19:14:29 UTC
efebfad Yong He 17 November 2021, 21:20:13 UTC gfx ShaderObject interface update, getTextureAllocationInfo() (#2019) * gfx ShaderObject interface update, getTextureAllocationInfo() * Fix render-vk compiler warnings and errors. Co-authored-by: Yong He <yhe@nvidia.com> 17 November 2021, 21:20:13 UTC
763d0b5 Alexander Weinrauch 17 November 2021, 07:28:59 UTC Added 8/16/64 bit int/unsigned int texture formats support for glsl shaders (#2017) Co-authored-by: Yong He <yonghe@outlook.com> 17 November 2021, 07:28:59 UTC
5042867 Yong He 16 November 2021, 20:59:42 UTC gfx: setSamplePositions and clearResourceView. (#2018) Co-authored-by: Yong He <yhe@nvidia.com> 16 November 2021, 20:59:42 UTC
c51f1e2 jsmall-nvidia 16 November 2021, 13:54:55 UTC Support around JSON-RPC (#2014) * #include an absolute path didn't work - because paths were taken to always be relative. * Use 'Process' to communicate with an command line tool. * Remove slang-win-stream * Tidy up windows ProcessUtil. * First version of BufferedReadStream. * Windows working IPC for steams. * Test proxy count option. * Split Process/ProcessUtil. Process is platform dependant. ProcessUtil are functions that are platform independent. * First implementation of Unix Process interface. * Unix process compiles on cygwin. * Fix typo in unix process. * Separate unix pipe stream error of invalid access, from pipe availability. * Fix in standard line extraction. * Make fd non blocking. * Fix issues with Windows Process streams. * Added UnixPipe. * Some fixes around UnixPipeStream. * Make a unix stream closed explicit. * Hack to debug linux process/stream. * Revert to old linux pipe handling. * Pass executable path for unit tests. Split out CommandLine into own source. * Small improvements in process/command line. * Check process behavior with crash. * Make stderr and stdout unbuffered for crash testing. * Only turn disable buffering in crash test. * Disable crash test on CI. * Fix crash on clang/linux. * Enable crash test. Remove _appendBuffer as can use StreamUtil functionality. * Added inital processing for http headers. * Small improvements to HttpHeader. * First pass HTTPPacketConnection working on windows. * Enable other Process communication tests. * Update comments. * WIP JSON RPC. * Add terminate to Process. Made JSONRPC a Util. * Small tidy up around HTTPPacketConnection. * Improve process termination options. Co-authored-by: Yong He <yonghe@outlook.com> 16 November 2021, 13:54:55 UTC
5a29c15 lucy96chen 16 November 2021, 04:42:39 UTC Add BGRA8 Unorm SRGB to the list of supported formats (#2016) Co-authored-by: Yong He <yonghe@outlook.com> 16 November 2021, 04:42:39 UTC
914a380 jsmall-nvidia 16 November 2021, 01:45:21 UTC Http protocol support (#2012) * #include an absolute path didn't work - because paths were taken to always be relative. * Use 'Process' to communicate with an command line tool. * Remove slang-win-stream * Tidy up windows ProcessUtil. * First version of BufferedReadStream. * Windows working IPC for steams. * Test proxy count option. * Split Process/ProcessUtil. Process is platform dependant. ProcessUtil are functions that are platform independent. * First implementation of Unix Process interface. * Unix process compiles on cygwin. * Fix typo in unix process. * Separate unix pipe stream error of invalid access, from pipe availability. * Fix in standard line extraction. * Make fd non blocking. * Fix issues with Windows Process streams. * Added UnixPipe. * Some fixes around UnixPipeStream. * Make a unix stream closed explicit. * Hack to debug linux process/stream. * Revert to old linux pipe handling. * Pass executable path for unit tests. Split out CommandLine into own source. * Small improvements in process/command line. * Check process behavior with crash. * Make stderr and stdout unbuffered for crash testing. * Only turn disable buffering in crash test. * Disable crash test on CI. * Fix crash on clang/linux. * Enable crash test. Remove _appendBuffer as can use StreamUtil functionality. * Added inital processing for http headers. * Small improvements to HttpHeader. * First pass HTTPPacketConnection working on windows. * Enable other Process communication tests. * Update comments. 16 November 2021, 01:45:21 UTC
ae9df74 Yong He 15 November 2021, 21:52:37 UTC Update gfx interface. (#2015) 15 November 2021, 21:52:37 UTC
7a4f08e lucy96chen 12 November 2021, 23:43:23 UTC Add support for buffer sharing from Vulkan to CUDA (#2008) * Added getSharedHandle() and additional code to handle shareable buffer creation to Buffer::init() and initVulkanInstanceAndDevice() for Vulkan; Modified createBufferFromSharedHandle() in CUDA to assign externalMemoryHandleDesc.type based on the type of handle being provided; Added an additional test case to get-shared-handle.cpp testing Vulkan to CUDA * Added createBufferFromNativeHandle() to Vulkan and enabled corresponding test * disable cuda * Fixed getSharedHandle() for D3D12 buffers assigning Win32 as the handle's source * Removed a dangling comment inside Buffer::init() * Added a missing override; Added code to check that a physical device supports the necessary external memory extensions before adding them to the deviceExtensions list; Added #if SLANG_WINDOWS_FAMILY guards around all Windows-specific code and sharedHandleVulkanToCUDA test (which uses said platform-specific code) * Added Windows check around vkGetMemoryWin32HandleKHR in vk-api.h * Added missing Windows check around BufferResourceImpl destructor * Added a temporary hack to ensure synchronization between devices, which solves an issue with buffer sharing resulting in incorrect values being read back; Added #if SLANG_WIN64 around all CUDA tests as the backend currently only supports running CUDA on 64-bit (despite devices being created successfully in a 32-bit config) 12 November 2021, 23:43:23 UTC
6f523dd Theresa Foley 11 November 2021, 21:45:23 UTC Fix two test cases that were causing problems (#2011) First, we have a CUDA-only test that simply needed a format name to be changed to match the new conventions in `gfx`. Second, we have one of the "active mask" tests that seems to produce different results locally for developers (under Vulkan) than it does on CI. This is almost certainly down to differences in GPUs and/or drivers. The inconsistency ultimately proves the point that I was trying to make when I wrote those tests - the "active mask" concept is effectively meaningless as exposed in D3D and Vulkan because it has not been specified in a way that allows programmers to reason about its value, and drivers have implemented wildly different interpretations of its supposed semantics for so long that there is no real hope of turning `WaveGetActiveMask()` into something that returns a well-defined value in any but the most trivial cases. TLDR: I disabled that test for Vulkan, which means it is completely disabled. 11 November 2021, 21:45:23 UTC
ab8eeea Theresa Foley 11 November 2021, 20:07:07 UTC Fix a bug in CUDA source emit (#2010) It seems that we were missing a `template<>` prefix when emitting explicit specializations for the `Matrix<...>` type in the CUDA prelude. This might not have caused errors in some versions of the CUDA compiler, but it at least seems to fail on the 11.5 SDK, which I am using to test locally. 11 November 2021, 20:07:07 UTC
8a9e518 jsmall-nvidia 10 November 2021, 22:33:22 UTC Interprocess communication via pipes (#2009) * #include an absolute path didn't work - because paths were taken to always be relative. * Use 'Process' to communicate with an command line tool. * Remove slang-win-stream * Tidy up windows ProcessUtil. * First version of BufferedReadStream. * Windows working IPC for steams. * Test proxy count option. * Split Process/ProcessUtil. Process is platform dependant. ProcessUtil are functions that are platform independent. * First implementation of Unix Process interface. * Unix process compiles on cygwin. * Fix typo in unix process. * Separate unix pipe stream error of invalid access, from pipe availability. * Fix in standard line extraction. * Make fd non blocking. * Fix issues with Windows Process streams. * Added UnixPipe. * Some fixes around UnixPipeStream. * Make a unix stream closed explicit. * Hack to debug linux process/stream. * Revert to old linux pipe handling. * Pass executable path for unit tests. Split out CommandLine into own source. * Small improvements in process/command line. * Check process behavior with crash. * Make stderr and stdout unbuffered for crash testing. * Only turn disable buffering in crash test. * Disable crash test on CI. * Fix crash on clang/linux. * Enable crash test. Remove _appendBuffer as can use StreamUtil functionality. 10 November 2021, 22:33:22 UTC
95e82ac Theresa Foley 10 November 2021, 18:48:44 UTC Fix a bug with CUDA entry-point params (#2007) Recent work that added support for translating DXR-style ray tracing shaders to work with OptiX seems to have accidentally applied its transformations even when compute shaders are translated for CUDA. As a result, compute entry points with `uniform` parameters at entry-point scope would be miscompiled to use OptiX calls that are not available for non-OptiX compiles. This change fixes the relevant pass so that it correctly opts-out on compute entry points, and also unifies some pieces of code that were being shared between a few different IR passes but that had gotten copy-pasted for the OptiX case. The fix has been confirmed by running relevant CUDA tests locally, but CUDA is still disabled in the default CI builds, so this change is not yet actively being tested to avoid further regression. 10 November 2021, 18:48:44 UTC
4d4cd56 lucy96chen 09 November 2021, 19:59:43 UTC Allow buffers to be shared between D3D12 and CUDA (#2005) * Added both the SharedHandle struct containing a handle and the API the handle originated from and the getSharedHandle() method to IResource, which returns a Windows system handle for the resource that can then be shared between multiple APIs (currently only fully implemented for D3D12); Added createTextureFromNativeHandle() and createBufferFromNativeHandle() to IDevice, which creates a buffer or texture resource using the provided handle (currently only fully implemented for D3D12); Added createBufferFromSharedHandle() to IDevice, which creates a BufferResource using the provided system handle (currently only fully implemented for the D3D12 to CUDA interface); Provided a proper implementation for CUDADevice::getNativeHandle(); Added several new tests testing the aforementioned implementations; Moved NativeHandle and getNativeHandle() for IBufferResource and ITextureResource up a layer into IResource and renamed to NativeResourceHandle; Modified NativeResourceHandle to be a struct containing the handle and the API it originated from and propagated these changes where appropriate * Combined all native and shared handle representations into a unified InteropHandle struct which tracks the handle's value and source API; Modified all getNativeHandle() and getSharedHandle() variants to operate on InteropHandle and modified all affected files * D3D12 buffers and textures are now responsible for closing their shared handles if they exist; Renamed IDevice::getNativeHandle() to getNativeDeviceHandles() * Fixed getNativeDeviceHandles() in render-cuda to match updated method elsewhere * Temporarily disabling existingDeviceHandleCUDA and sharedHandleD3D12ToCUDA due to currently unreproducable test failures on TC 09 November 2021, 19:59:43 UTC
8fe3f9c Yong He 04 November 2021, 16:40:56 UTC Add interface for new gfx features. (#2003) * Add interface for new gfx features. * Add cuda implementation. * Code review fixes. * Fix. Co-authored-by: Yong He <yhe@nvidia.com> 04 November 2021, 16:40:56 UTC
af0f26d Theresa Foley 04 November 2021, 05:44:21 UTC Fix an infinite-recursion bug in type-checking (#2004) Fixes #1990 The underlying problem here is in the `ExtractExistentialType` AST node class. An "existential" in current Slang is typically a value of interface type. When such a value is used in an operation, the type-checker "opens" the extistential so that subsequent type-checking steps can work with the (statically unknown) specific type of the value stored inside. The `ExtractExistentialType` AST node represents the type of an existential that has been "opened" in this way. When the front-end performs lookup "into" a value with one of these types, it nees to use a reference to the original interface declaration with a "this-type substitution" that refers to the "opened" type (a this-type substitution tells the compiler the concrete type it should use in place of `This` in signatures within the interface; it allows compiler to "see" the right associated type definitions to use in a context). Prior to this change, the implementation would store the specialized reference to the original interface declaration in the `ExtractExistentialType` node as part of its state. The catch there is that the specialized interface reference indirectly refers to the `ExtractExistentialType` AST node itself, creating a circularity. As soon as the front-end performs any operation that tries to recurse over that structure, it would go into an infinite loop. The fix here sounds kind of like a hack, but seems to be pretty nice in practice. Instead of always storing the specialized interface reference, we instead store the few values that are needed to construct it, and then create and cache the actual reference on-demand. The on-demand created fields are not considered part of the state of the AST node for any kind of recursion or serialization, so they avoid the original problem. A single test case was added that represents the original bug, and confirms the fix. 04 November 2021, 05:44:21 UTC
4b7e674 Yong He 02 November 2021, 17:40:35 UTC Include gfx.lib in release. (#2002) Co-authored-by: Yong He <yhe@nvidia.com> 02 November 2021, 17:40:35 UTC
f9a7b7d Yong He 02 November 2021, 17:16:50 UTC Include gfx header in releases (#2001) Co-authored-by: Yong He <yhe@nvidia.com> 02 November 2021, 17:16:50 UTC
45ae0eb Yong He 01 November 2021, 17:53:42 UTC Disable aarch64 build for releases. (#2000) Co-authored-by: Yong He <yhe@nvidia.com> 01 November 2021, 17:53:42 UTC
b1384a9 jsmall-nvidia 29 October 2021, 21:19:12 UTC Update slang-llvm deps (#1999) * #include an absolute path didn't work - because paths were taken to always be relative. * Update slang-llvm deps. 29 October 2021, 21:19:12 UTC
22e7f3f Yong He 29 October 2021, 20:03:14 UTC [gfx/d3d12] Add a constant for microsoft vendor ID. (#1998) 29 October 2021, 20:03:14 UTC
4ebbf35 Yong He 29 October 2021, 16:36:07 UTC Use detected shader model in gfx/d3d12. (#1996) * Use detected shader model in gfx/d3d12. * Enable all d3d12 tests on Github. * Improve d3d12 software device detection. * Disable d3d12 tests on github for now. Co-authored-by: Yong He <yhe@nvidia.com> 29 October 2021, 16:36:07 UTC
d406e72 jsmall-nvidia 28 October 2021, 21:27:40 UTC Re-enable gfx unit tests (#1993) * #include an absolute path didn't work - because paths were taken to always be relative. * Reenable gfx unit tests * Kick. 28 October 2021, 21:27:40 UTC
a9bbcb9 jsmall-nvidia 28 October 2021, 19:03:29 UTC Upgrade glslang (#1995) * #include an absolute path didn't work - because paths were taken to always be relative. * Update glslang. * Update slang-binaries to have glslang 11.16.0-3 28 October 2021, 19:03:29 UTC
203b557 jsmall-nvidia 28 October 2021, 14:54:18 UTC Hotfix/osx fix (#1994) * #include an absolute path didn't work - because paths were taken to always be relative. * OSX 10.15 produced error/warning for not using ref. 28 October 2021, 14:54:18 UTC
8f6450c jsmall-nvidia 27 October 2021, 23:44:34 UTC Update glslang binaries (#1991) * #include an absolute path didn't work - because paths were taken to always be relative. * Use updated slang-binaries that have SPIR-V diagnostics improvements. * Re-enable nv-ray-tracing-motion-blur, because with SPIR-V diagnostic fixes in glslang - there shouldn't be spurious errors from glslang compilation. * If optimization fails use the SPIR-V we have. * Update slang binaries. * Hack to disable gfx unit tests for now to try and get CI pass for this PR. 27 October 2021, 23:44:34 UTC
95654c3 jsmall-nvidia 27 October 2021, 19:21:09 UTC SPIR-V fixes (#1992) * #include an absolute path didn't work - because paths were taken to always be relative. * Use updated slang-binaries that have SPIR-V diagnostics improvements. * Re-enable nv-ray-tracing-motion-blur, because with SPIR-V diagnostic fixes in glslang - there shouldn't be spurious errors from glslang compilation. * If optimization fails use the SPIR-V we have. * Update SPIR-V headers and generated files. Updated documentation. * Update spirv-headers/tools. Revert slang-binaries. * Remove hack around spir-v optimization as no longer needed. disable nv-ray-tracing-motion-blur.slang 27 October 2021, 19:21:09 UTC
76a1ec5 jsmall-nvidia 27 October 2021, 15:50:44 UTC Small improvements around invoking unit-tests (#1978) * #include an absolute path didn't work - because paths were taken to always be relative. * Attempt to check if a crash is occurring in a unit test * Small update really to just kick another build. 27 October 2021, 15:50:44 UTC
62b1e58 jsmall-nvidia 27 October 2021, 01:42:11 UTC Runs all gfx unit tests through a 'test proxy' (#1981) * #include an absolute path didn't work - because paths were taken to always be relative. * Support for test proxy. * Turn on testing using proxy. * Don't pass sink into check of downstream compiler. * Small change to kick off build. * Remove register specification on transcendental. * Increase poll timeout. Small improvements to proxy. * Disable gfx unit tests. * Put test runner in shared library mode by default. * Change comment. Kick off another CI test. * Small edit to kick off builds. * Run unit tests on proxy. * Turn on using proxy for now. * Enable swift shader. * Fix typo. Add exception support. * Make the default spwan type SharedLibrary Use isolation for gfx unit tests. * Update slang-binaries. * Fix typo. * Report unit test output information. 27 October 2021, 01:42:11 UTC
dcc2b85 lucy96chen 26 October 2021, 23:30:59 UTC Expanded gfx::Format to include additional formats (#1982) * Format list updated with additional formats supported by both D3D and Vulkan; D3DUtil::getMapFormat() and VkUtil::getVkFormat() updated to include additional formats; GFX_FORMAT() updated with all additional formats (BC compression unfinished) * Finished updating GFX_FORMAT with newly added formats and sizes; Pixel size is now tracked using the FormatPixelSize struct containing the values for bytes per block and pixels per block to accomodate BC formats; Updated gfxGetFormatSize and associated sub-calls to return FormatPixelSize instead of uint8_t; Most calls to gfxGetFormatSize() updated to reflect changes, a couple calls still unupdated * Changes to accommodate new formats finished, debugging slang-literal unit test * First format unit test working * One test added for BC1Unorm and RGBA8Unorm_SRGB, both passing * Refactored format testing code to merge BC1Unorm and RGBA8Unorm SRGB into a single file * All unit tests added for BC and Srgb formats * Most tests added and working; Added five additional formats (still need tests) and made the appropriate changes to support these; createTextureView() modified for D3D11, D3D12, and Vulkan to take into account the format specified in the texture view desc when the texture's format is typeless * Format enums renamed to more closely match their D3D counterparts; Added a universal float and uint buffer and buffer view for use across all Format tests * Remaining tests added; D3D12 tests pass, but Vulkan crashes in BC1_UNORM and D3D11 spits out a bunch of D3D11 Errors (but supposedly passes) * re-run premake * Added Sint versions of test shaders; Vulkan and D3D11 tests also pass * Size struct for format unit tests no longer use initializer lists * Fixed a Size struct missed in the previous pass * Fixed minor bugs causing tests to fail * Added documentation detailing all currently unsupported formats * Skip tests causing unsupported format warnings due to swiftshader * updated several test using old Format enum names * Revert change to compareComputeResult() that was added for debugging purposes * DEBUGGING: Added prints to identify which formats are failing on CI * Reverted attempted debugging changes; Fixed texture2d-gather.hlsl to use updated Format enums * Fixed incorrect array sizes in d3d11 _initSrvDesc() * Commented out further tests that produce unexpected results when tested for Vulkan with swiftshader * Revert "Merge branch 'expanded-format-support' of https://github.com/lucy96chen/slang into expanded-format-support" This reverts commit 20008f0d3ecc3b1405ecac8c138edaa3cd37ed6b, reversing changes made to 6081e95827315fee50e18409394d5abd62fac787. * Added a fuzzy comparison function for use with floats * submodule update * Revert messed up changes caused by previous revert after automatically merging on github 26 October 2021, 23:30:59 UTC
fe6d5f1 jsmall-nvidia 26 October 2021, 21:40:36 UTC SPIR-V optimization diagnostics (#1989) * #include an absolute path didn't work - because paths were taken to always be relative. * WIP fixing glslang error reporting. * SPIR-V opt diagnostics. * Small formatting tidy up. * Disable nv-ray-tracing-motion-blur.slang for now, until merged and can rebuild slang-glslang. * Formatting improvements. * Some small improvements around optimizing in glslang. * Make clear if calling optimize with 'non' in glslang no work is done (such as a pointless copy). * Small formatting change - mainly to kick off a build. * Upgrade slang-binaries to fix git fetch issue on TC. * Small formatting edits. Kick build. 26 October 2021, 21:40:36 UTC
499e676 jsmall-nvidia 25 October 2021, 19:02:17 UTC Enabling slang-llvm for host-callable (#1975) * #include an absolute path didn't work - because paths were taken to always be relative. * First integration of slang-pack. * Use .os * Add optional dependency support. * Update github actions/scripts to update deps. aarch64 needs special handling. * Upgrade to latest slang-pack for ignore-deps support. * Fix linux build issues. * Copying slang-llvm from dependencies. * Add support for LLVM for host callable. Added CodeGenTransitionMap. * Remove hack to enable host callable for LLVM. * Small improvements around transitions/downstream compiler. * Fix typo in method name. * Fix comment. * Update visual studio project. * Updage slang-llvm to include initialization fix. * Fix handling extraction of clang version number. * Fix some formatting problems. * hack - to see if there is a version problem on CI. * Remove progress on github action linux. * Allow version lines to have text before 'prefix'. * Update slang-binaries to include centos-7 premake binaries. * Upgrade slang-binaries. * Upgrade slang-binaries. * Update slang binaries to have certificates. * Fix handling of dependency path. * Update README to include LLVM Update building to include --deps and --arch * Include slang-llvm in packages. * Update building docs. 25 October 2021, 19:02:17 UTC
8fb8459 jsmall-nvidia 22 October 2021, 18:10:51 UTC Feature/update slang binaries (#1988) * #include an absolute path didn't work - because paths were taken to always be relative. * Add slang-binaries to have certificates. 22 October 2021, 18:10:51 UTC
dc991f7 Yong He 21 October 2021, 23:27:40 UTC Passing associated type arguments to existential parameters + packing for `bool`. (#1987) * Passing associated type arguments to existential parameters + packing for `bool`. * fix typo Co-authored-by: Yong He <yhe@nvidia.com> 21 October 2021, 23:27:40 UTC
9304c2d Yong He 21 October 2021, 22:51:18 UTC Diagnostic for no type conformance + bug fix. (#1985) * Diagnostic for no type conformance + bug fix. * Fixes. * Fix. * Include heterogeneous example only with --enable-experimental-projects premake flag Co-authored-by: Yong He <yhe@nvidia.com> Co-authored-by: jsmall-nvidia <jsmall@nvidia.com> 21 October 2021, 22:51:18 UTC
66e319e jsmall-nvidia 21 October 2021, 17:13:25 UTC Upgrade slang-binaries (#1986) * #include an absolute path didn't work - because paths were taken to always be relative. * Update slang-binaries to have centos binaries. 21 October 2021, 17:13:25 UTC
edade2e jsmall-nvidia 20 October 2021, 15:26:52 UTC Initial `slang-pack` integration (#1974) * #include an absolute path didn't work - because paths were taken to always be relative. * First integration of slang-pack. * Use .os * Add optional dependency support. * Update github actions/scripts to update deps. aarch64 needs special handling. * Upgrade to latest slang-pack for ignore-deps support. * Fix linux build issues. 20 October 2021, 15:26:52 UTC
deb638f jsmall-nvidia 20 October 2021, 14:32:50 UTC Selecting downstream compiler on code gen transition (#1980) * #include an absolute path didn't work - because paths were taken to always be relative. * Add support for LLVM for host callable. Added CodeGenTransitionMap. * Remove hack to enable host callable for LLVM. * Small improvements around transitions/downstream compiler. * Fix typo in method name. * Fix comment. 20 October 2021, 14:32:50 UTC
8406244 Yong He 20 October 2021, 14:18:03 UTC Re-enable swiftshader. (#1984) Ignore some gfx unit tests when using swiftshader. Co-authored-by: Yong He <yhe@nvidia.com> 20 October 2021, 14:18:03 UTC
ba081c8 David Siher 19 October 2021, 20:07:29 UTC Generalize heterogenous code emit (#1968) * Bring heterogeneous-hello-world back up to date. * Reintroduced heterogeneous-hello-world into the premake * No longer uses compiled bytecode for entry point, instead a loadModule call is hardocoded with the slang file name. * Entry point is, similarly, hardcoded for now. * Added a bypass to slang-legalize-types for an unneeded GPUForeach check * Run premake and change to relative path * Removed experimental and added README * Add prebuild command to premake for heterogeneous example * Pass in entry point as parameter (also remove shader bytecode) * Pass in module name as parameter * Squashed commit of the following: commit 5b13b57fe600724344c556fe4309a5d6bb3d39ab Author: Kai Yao <kyao@nvidia.com> Date: Thu Oct 7 23:38:50 2021 -0700 Return diagnostics data when encountering module load error by exception (#1966) commit 112e1515c30fa972ff56f91514b70946153c718c Author: jsmall-nvidia <jsmall@nvidia.com> Date: Thu Oct 7 16:12:29 2021 -0400 Disable test crashing CI (#1965) * #include an absolute path didn't work - because paths were taken to always be relative. * Disable test that appears to be crashing. commit da32069a0c1c8c723d7ef45100049a8f0dd5d9c4 Author: Kai Yao <kyao@nvidia.com> Date: Mon Oct 4 13:58:51 2021 -0700 Modified barrier API to accept multiple resources per call (#1959) Co-authored-by: Yong He <yonghe@outlook.com> commit 97bb82ebcdf8f1391b9d93b5a8d7b1dfc4e88e52 Author: jsmall-nvidia <jsmall@nvidia.com> Date: Mon Oct 4 14:15:51 2021 -0400 Removing exceptions from core/compiler-core (#1953) * #include an absolute path didn't work - because paths were taken to always be relative. * Refactor Stream. Working on all tests. * Split out CharEncode. * Make method names lower camel. m_prefix in Writer/Reader * Tidy up around CharEncode interface. * Small improvements around encode/decode. * Better use of types. * Remove readLine from TextReader. * Remove exceptions from Stream/Text handling. * Fix some typos. * Fix tabbing. * Fix missing override. * Remove remaining exception throw/catch via using signal mechanism. * Remove exceptions that are not used anymore. * Document the Stream interface. * Remove index for decoding 'get byte' function. * Fix CharReader -> ByteReader. commit b3dfe383c6d31ff3dbd76dcfb32de8d536382f3e Author: lucy96chen <47800040+lucy96chen@users.noreply.github.com> Date: Mon Oct 4 09:46:33 2021 -0700 Get native handles for TextureResource and BufferResource (#1960) * Added getNativeHandle() to TextureResource and BufferResource; Implemented getNativeHandle() in Vulkan and D3D12; Added new unit test files for the aforementioned implementation * Added missing getNativeHandle() implementations to renderer-shared.cpp and CUDA * Finished new getNativeHandle() unit tests for ITextureResource and IBufferResource; Modified ICommandQueue and ICommandBuffer unit tests to call QueryInterface to convert to IUnknown then back and compare resulting pointers for equality * Unit tests updated and pass locally * Cast m_buffer.m_buffer and m_image to uint64_t commit 35bca4cc432613af3926da3bed217a6baa9cbd26 Author: lucy96chen <47800040+lucy96chen@users.noreply.github.com> Date: Fri Oct 1 13:08:25 2021 -0700 Add getNativeHandle() to ICommandQueue and ICommandBuffer (#1952) * Added support for getting command buffer and command queue handles to ICommandBuffer and ICommandQueue; D3D12Device, VkDevice, and DebugDevice modifieid to implement this new functionality; immediate-renderer-base.cpp also modified to implement the new functions * Removed excess boilerplate * Changed readRef() to get() in D3D12 getNativeHandle() implementation for ICommandBuffer and ICommandQueue * Added unit tests for new getNativeHandle() implementations, unfinished * Queue test added; Minor cleanup changes * getBufferHandleTestImpl() now closes the command buffer before returning * Added getNativeHandle() implementations to CUDADevice * Added comment clarifying that the Vulkan check is checking for a null handle, which is defined to be 0 commit 6c6200f547c7387598743b23bb3c8f0d375d9494 Author: Kai Yao <kyao@nvidia.com> Date: Thu Sep 30 20:25:34 2021 -0700 VK Resource Barrier (#1955) * Resource barrier API and VK implementation * Stub implementations * Handle VK Acceleration Structure flag * Add a couple more cases to pipeline barrier stages commit 627fc976bac5c2381dbace9c7925cb6a68b8de12 Author: Yong He <yonghe@outlook.com> Date: Thu Sep 30 19:48:47 2021 -0700 Fix aarch64 build on github (#1957) Co-authored-by: Yong He <yhe@nvidia.com> commit 122d701513e116856bd59c999221ce36a373d7db Author: Yong He <yonghe@outlook.com> Date: Thu Sep 30 17:51:56 2021 -0700 Fix GitHub release (#1956) * Fix aarch64 release build config. * Fix for WinAarch64 build. * Update premake for embed-std-lib build on aarch64. * `platform` fix for aarach64 build. * Try revert back to use absolute output path for slang-stdlib-generated.h * Fix * fix Co-authored-by: Yong He <yhe@nvidia.com> commit aa8f7b899b7b562b3d3c6e25c3da41569505e70c Author: Chad Engler <englercj@live.com> Date: Wed Sep 29 13:02:47 2021 -0700 Fix ARM64 detection for MSVC (#1951) commit 6736b0c1c5fa3e89bc561eb7965a1a0d17af3466 Author: Yong He <yonghe@outlook.com> Date: Wed Sep 29 11:29:46 2021 -0700 Add ISession::loadModuleFromSource. (#1950) Co-authored-by: Yong He <yhe@nvidia.com> commit d8e452412e14a6a8ba137f2adcae13b398e5cecb Author: Yong He <yonghe@outlook.com> Date: Tue Sep 28 15:03:03 2021 -0700 Fix AbortCompilationException leaking through loadModule API. (#1949) * Fix AbortCompilationException leaking through loadModule API. * Update. * Fix. Co-authored-by: Yong He <yhe@nvidia.com> commit cdf1b2c007fefdca128584d2a9f63dec3d350e16 Author: Yong He <yonghe@outlook.com> Date: Tue Sep 28 11:54:24 2021 -0700 Improvements to the unit test framework. (#1948) commit af788b62e18bbd55cd748ad60400a74cf1bc93ee Author: lucy96chen <47800040+lucy96chen@users.noreply.github.com> Date: Fri Sep 24 16:53:41 2021 -0700 Add existing device handle support unit test (#1946) commit bec8e6aec85b6e3f875c58bdd59eb15613978358 Author: Yong He <yonghe@outlook.com> Date: Fri Sep 24 11:33:44 2021 -0700 Move existing unit tests to a standalone dll. (#1945) commit f2a3c933bc11a498c622fa18694c84beca8ca031 Author: lucy96chen <47800040+lucy96chen@users.noreply.github.com> Date: Thu Sep 23 12:19:49 2021 -0700 Add method to retrieve native handles (#1944) * Added a getNativeHandle() method that retrieves the natively created handles; Modified RendererBase, VKDevice, D3D12Device, and DebugDevice to implement this new method * Moved ExistingDeviceHandles out of Desc directly inside IDevice and renamed to NativeHandles; Modified calls accessing the struct accordingly in RendererBase, DebugDevice, VKDevice, and D3D12Device * Minor cleanup changes (renames, etc.) commit b9b398d038b524f15a86ff27cd6888d54e8754e0 Author: Yong He <yonghe@outlook.com> Date: Wed Sep 22 10:06:59 2021 -0700 Add gfx unit testing framework. (#1943) * Add gfx unit testing framework. * Fix compilation error. * Reset gfxDebugCallback after render_test. * Pass enabledApi flags through. * Fix for code review suggestions. Co-authored-by: Yong He <yhe@nvidia.com> commit 6e9cee69b3588ddae09b08b9f580f59ad899983f Author: lucy96chen <47800040+lucy96chen@users.noreply.github.com> Date: Tue Sep 21 18:46:32 2021 -0700 Support for existing device/instance handles in Vulkan (#1942) commit b1f04c8544c650de3947955ca68f679535d249aa Author: lucy96chen <47800040+lucy96chen@users.noreply.github.com> Date: Wed Sep 15 20:22:45 2021 -0700 Allow D3D12Device to use an existing device handle (#1940) * Added a new field for an existing device handle to IDevice::Desc; Modified D3D12Device::initialize to set the device stored in desc if it already exists instead of creating a new one * Turned existingDeviceHandle into a struct containing an array of two elements; Updated D3D12Device::initialize to match changes to existingDeviceHandle; Updated comments * Fixed style error for ExistingDeviceHandles struct commit 2f7b9f5ae8be21c6c1d75ae9caefbc7b3f8986a9 Author: Pablo Delgado <private@pablode.com> Date: Thu Sep 16 01:17:57 2021 +0200 Fix incorrect WIN32 macros and missing Windows.h inclusion (#1939) * Replace WIN32 preprocessor macros with _WIN32 * Add missing Windows.h include for InterlockedIncrement commit 11d43642008905ac69a3832eb8a9b2ae7b785f86 Author: Yong He <yonghe@outlook.com> Date: Tue Sep 14 11:36:44 2021 -0700 Avoid upcasting to f32 in 16bit float-uint bit cast. (#1938) Co-authored-by: Yong He <yhe@nvidia.com> commit 502aa3812a82cf0d091cff0c67804e4ee448ac78 Author: David Siher <32305650+dsiher@users.noreply.github.com> Date: Tue Sep 14 12:59:55 2021 -0400 Bring heterogeneous-hello-world back up to date. (#1935) * Bring heterogeneous-hello-world back up to date. * Reintroduced heterogeneous-hello-world into the premake * No longer uses compiled bytecode for entry point, instead a loadModule call is hardocoded with the slang file name. * Entry point is, similarly, hardcoded for now. * Added a bypass to slang-legalize-types for an unneeded GPUForeach check * Run premake and change to relative path * Removed experimental and added README Co-authored-by: Yong He <yonghe@outlook.com> * Revert "Squashed commit of the following:" This reverts commit 4f665858d65f7c332c616ef6db9fdafa1c5e0b9f. * Run premake * Remove prebuild command (only works on Windows?) * Rerun premake * Fix heterogeneous prebuild command * Remove linux specific prebuild command * Fix prebuild command (again) * Change target from dxbc to hlsl to see if that fixes linux issues * Use Path::getFileNameWithoutExt * Change string-literal.slang.expected to have extra filename in decoration Co-authored-by: Yong He <yonghe@outlook.com> 19 October 2021, 20:07:29 UTC
back to top