| Age | Commit message (Collapse) | Author |
|---|
|
|
|
(#2388)
|
|
|
|
* `is` and `as` operator and `Optional<T>`.
* Fix.
Co-authored-by: Yong He <yhe@nvidia.com>
|
|
* Merge slang-ir-diff-jvp.cpp
* Added support and tests for other float vector types
* Added swizzle test and code to handle it (tests failing currently)
* Fixed one test, the other is still pending
* Fixed instruction cloning logic to avoid modifying original function
* Fixed an issue with custom 'pow_jvp' and added support for vector contructor
* Minor update to comments
* Fixed support for division
* Fixed an issue with uninitialized diagnostic sink
* Moved derivative processing to after mandatory inlining.
Skip instructions that don't have side-effects and aren't used by anything.
* WIP: Handling unconditional control flow and multi-block functions
* Support for unconditional multi-block functions
* Added a dead code elimination step to the derivative pass
* Changed name of 'hasNoSideEffects()'
* Refactored variable names
* Added initial IR defs for new type system
* Added necessary logic for semantic checking
* Overhauled type system to use builtin pair types and conform to the IDifferentiable interface
* Automatically replace IRDifferentiablePairType to a custom IRStructType
* Added generics handling by expanding the conformance context functionality and allowing for type parameters
* Minor fix: early return in processPairTypes()
* Minor fixes to differentiable resolution on generic types
* Added new instructions for differential pairs. Basic tests work now.
Looking into generic types.
* Adjusted most tests to the new type system. OutType and InOutType are still not properly working.
* Updated __jvp to produce both primal and differential output
* Moved autodiff related declarations to diff.meta.slang
* Refactored variable names
* Added initial IR defs for new type system
* Added necessary logic for semantic checking
* Overhauled type system to use builtin pair types and conform to the IDifferentiable interface
* Automatically replace IRDifferentiablePairType to a custom IRStructType
* Added generics handling by expanding the conformance context functionality and allowing for type parameters
* Minor fix: early return in processPairTypes()
* Minor fixes to differentiable resolution on generic types
* Added new instructions for differential pairs. Basic tests work now.
Looking into generic types.
* Adjusted most tests to the new type system. OutType and InOutType are still not properly working.
* Updated __jvp to produce both primal and differential output
* Moved autodiff related declarations to diff.meta.slang
* Removed external changes
* Cleanup the transcription logic: each case returns a pair of insts for the primal and differential computation.
|
|
* Allow `class` to implement COM interface, [DLLExport]
* Fix [COM] usage in tests and examples with UUIDs.
Co-authored-by: Yong He <yhe@nvidia.com>
|
|
* Support `class` types.
* Ignore class-keyword test
* Fix codereview comments and warnings.
Co-authored-by: Yong He <yhe@nvidia.com>
|
|
Co-authored-by: Yong He <yhe@nvidia.com>
|
|
arithmetic. Also added type-checking during the semantic stage. (#2303)
* Added JVPTranscriber to handle differentiation of load, store, var, param and return instructions, as well as conversion of data and function types
* Changed class names to be more in line with convention. Added correct type checking for __jvp() and verified that simple calls with only loads and stores are processed correctly
* Added logic to differentiate basic arithmetic and literals inside IRConstruct and fixed the way parameters are differentiated
Co-authored-by: Yong He <yonghe@outlook.com>
|
|
framework for passes to process them. (#2297)
* Added a decorator to mark functions for forward-mode differentiation
* Fill out support for calls to non-decl values
The existing compiler logic has a few places (semantic checking plus AST-to-IR lowering) where it assumes that function calls (`InvokeExpr`) are only ever made to expressions that resolve to a specific `Decl` (`DeclRefExpr`). This assumption allows semantic checking and lowering code to inspect things like the parameter list of an actual declaration, rather than just the type signature of the callee, and that infrastructure is used to support various features (e.g., default argument values on parameters).
The AST and IR representations themselves have no matching requirement, and the places where the more general case of call expressions would need to be supported were relatively clear in the code. This change attempts to add suitable logic into each of those places.
Note that this change does *not* surface any valid way to form input code that would cause these new code paths to be executed, so it is entirely possible that there are bugs in the logic as written here. The primary goal of this change is simply to get a sketch of the correct code checked in so that we have something to build on once we have language features that will require this support.
* fixup: warnings-as-errors
* Added parser logic for '__jvp(<fn-name>)' operator
* Fixed issue with missing overload candidate item and added basic parsing test for the __jvp syntax
* Added a blank JVP Auto-diff pass and a pass that replaces 'JVPDerivativeOf' calls with the differentiated function
* Added a couple comments
* Added parameter handling for the JVP pass
Co-authored-by: Theresa Foley <tfoley@nvidia.com>
|
|
* Lower throwing COM interface method.
* Fix.
* Fix warnings.
Co-authored-by: Yong He <yhe@nvidia.com>
|
|
* #include an absolute path didn't work - because paths were taken to always be relative.
* Use TerminatedUnownedStringSlice for literals in output C++.
* Remove Escape/Unescape functions used in slang-token-reader.cpp
Add target type of 'host-cpp' etc to map to the target types.
* Fix some corner cases around string encoding.
* Added unit test for string escaping.
Fixed some assorted escaping bugs.
* Updated test output.
* Added decode test.
* Stop using hex output, to get around 'greedy' aspect. Use octal instead.
* Added HostHostCallable
Small changes to use ArtifactDesc/Info instead of large switches.
* Fix C++ emit to handle arbitrary function export.
* Add options handling for callable without an output being specified.
* Can compile with COM interface. Added example using com interface.
* Use the IR Ptr type instead of hack in C++ emit for interfaces.
* Fix issue with outputting the COM call when ptr is used.
* Fix crash issue on compilation failure.
* Add support for __global.
* Added `ActualGlobalRate`
Added special handling around globals and COM interfaces.
Tested out in cpu-com-example.
* Fix typo in NodeBase.
* Support for accessing globals by name working.
* Check that actual global initialization is working.
* Refactor the com replacement such that it doesn't need a cache or do anything special with GlobalVar.
* Remove context.
Only create replacement if needed.
* Split out COM host-callable into a unit-test.
* host-callable com testing on C++and llvm.
* Comment around the COM ptr replacement.
* Disable com test on vs 32 bit.
Fix C++ prelude
* Disable 32 bit targets testing com host-callable.
* Use JSON parsing to locate VS version.
* Need platform detection in C++prelude.
* Fix com host callable test for LLVM.
* Work around for not being able to include "targetConditionals.h"
|
|
* Clean up `IRReturnVoid`.
* Update gitignore.
Co-authored-by: Yong He <yhe@nvidia.com>
|
|
* New language feature: basic error handling.
* Fix.
* Fix `tryCall` encoding according to code review.
Co-authored-by: Yong He <yhe@nvidia.com>
|
|
* #include an absolute path didn't work - because paths were taken to always be relative.
* Use TerminatedUnownedStringSlice for literals in output C++.
* Remove Escape/Unescape functions used in slang-token-reader.cpp
Add target type of 'host-cpp' etc to map to the target types.
* Fix some corner cases around string encoding.
* Added unit test for string escaping.
Fixed some assorted escaping bugs.
* Updated test output.
* Added decode test.
* Stop using hex output, to get around 'greedy' aspect. Use octal instead.
|
|
* #include an absolute path didn't work - because paths were taken to always be relative.
* Refactor Liveness pass, such that locations can be found independently of setting up ranges.
* Refactor around different stages of liveness span analysis.
* WIP Take into account PHI temporaries in liveness tracking.
* WIP First pass of PHI liveness refactor.
* Add BlockIndex.
* WIP Refactor phi liveness around inst runs.
* More improvements around liveness tracking.
* Bug fixes.
Special handling to not add multiple ends, at starts of blocks and after accesses.
* Fix test output.
* Use IRInsertLoc to track insertion point.
* Liveness markers don't have side effects.
* Fix typo in liveness test.
* Small improvements around setting SuccessorResult.
* Fix memory issue around reallocation and RAIIStackArray.
Update test output.
* Update test output for liveness.slang.
* Fix typo in SuccessorResult blockIndex.
* Small tidy up.
* Handle the root start block, correctly scoping the run.
* Split BlockInfo into 'Root' and 'Function'.
Store successors as BlockIndices.
* Tidy up around liveness tracking.
* Add head/tail support to ArrayViews.
Use Count where appropriate.
Use head/tail in liveness impl.
|
|
Co-authored-by: Yong He <yhe@nvidia.com>
Co-authored-by: Theresa Foley <10618364+tangent-vector@users.noreply.github.com>
|
|
* #include an absolute path didn't work - because paths were taken to always be relative.
* Add SPIRVLiteralType, to mark types that have spirv_literal in function parameter output.
* Update test result.
Co-authored-by: Theresa Foley <10618364+tangent-vector@users.noreply.github.com>
|
|
* #include an absolute path didn't work - because paths were taken to always be relative.
* Add support for HLSL `export`.
* Test for using `export` keyword.
|
|
* #include an absolute path didn't work - because paths were taken to always be relative.
* WIP tracking liveness.
* Skeleton around adding liveness instructions.
* Calling into liveness tracking logic.
Adds live start to var insts.
* Liveness macros have initial output.
* Looking at different initialization scenarios.
* Some discussion around liveness.
* WIP for working out liveness end.
* WIP Updated liveness using use lists.
* Is now adding liveness information
* Some small fixes.
* WIP around liveness.
* Seems to output liveness correctly for current scenario.
* Tidy up liveness code.
* Update comment arounds liveness to current status.
* Small fixes to liveness test.
* Add support for call in liveness analysis.
* Improve liveness example with array access.
* Small updates to comments.
* Disable liveness test because inconsistencies with output on CI system.
* Fix some issues brought up in PR.
* Rename liveness instructions.
|
|
* #include an absolute path didn't work - because paths were taken to always be relative.
* Compile to a dxil library.
* Added CompileProduct.
* Support handling of ModuleLibrary.
* CacheBehavior -> Cache
* Use CompileProduct for -r references.
* CompileProduct -> Artifact.
* Determining an artifact type on binding.
* Determine binary linkability.
* Added Artifact::exists.
* Added ArtifactKeep.
* Small fixes.
* Small improvements to Artifact.
* Add zip extension.
* Fix some comments.
* Fix multiple adding of PublicDecoration.
Make public output export for DXIL/lib.
Add checking for simpleDecorations such that only added once.
* Use 'whole program' to identify library build.
* Add -target dxil so test infrastructure knows it needs DXC.
|
|
* Support `[DllImport]`
* Fix.
* Fix.
* Fix array type emit in cpp.
* Fix.
* Fix.
* Fix
Co-authored-by: Yong He <yhe@nvidia.com>
|
|
|
|
`ImageSubscript` for GLSL (#2146)
|
|
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).
|
|
* 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
|
|
* #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.
|
|
* `reinterpret` and 16-bit value packing.
* Update `half-texture` cross-compile test reference result.
* Revert inadvertent reformatting of slang-ir-inst-defs.h
Co-authored-by: Yong He <yhe@nvidia.com>
|
|
* Further implementation of SPIRV direct emit.
This change implements:
- Struct, Vector, Matrix and Unsized Array types.
- Basic arithmetic opcodes, vector construct, swizzle etc.
- getElementPtr, getElement, fieldAddress, extractField.
- SPIRV target intrinsics with SPIRV asm code in stdlib.
- RWStructuredBuffer and StructuredBuffer.
- Pointer storage class propagation.
- Control flow.
* Fix.
|
|
And rename debug symbols for navis.
Co-authored-by: Yong He <yhe@nvidia.com>
|
|
- Fix emitting `StructuredBuffer<ISomething>::Load`, which triggers emitting for `IROp_WrapExistential` that is previously unhandled.
- Fix cuda layout around vectors, they should be aligned to 1,2,4,8,16 bytes instead of just using element type's alignment. That means `float4` has alignment of 16 instead of 4.
- Fix `SLANG_CUDA_HANDLE_ERROR` macro definition.
- Fix navis sometimes fail to find `Slang::kIROp_*` enum values when debugging external projects.
Co-authored-by: Yong He <yhe@nvidia.com>
Co-authored-by: jsmall-nvidia <jsmall@nvidia.com>
|
|
|
|
* Append proper suffixes to 16-bit literals for GLSL
The GLSL output path wasn't putting suffixes on literals of 16-bit types, and that was leading to compilation errors in downstream `glslang`. This change adds the suffixes defined by `GL_EXT_shader_explicit_arithmetic_types`.
This change also wraps up 8-bit literals so that they are emitted as, e.g., `int8_t(1)` instead of just `1`, to make sure we don't have implicit conversions in the output GLSL that weren't implicit in the Slang IR. We similarly wrap floating-point special values like infinities in their desired types when the type is `float` (e.g., `double(1.0 / 0.0)` for a double-precision infinity).
Note: Standad IEEE 754 half-precision doesn't provide an encoding for infinite or not-a-number values, so it might be considered an error if we emit `half(1.0 / 0.0)` but there really isn't a significantly better alternative for us to emit.
* fixup
|
|
* #include an absolute path didn't work - because paths were taken to always be relative.
* WIP: First pass in supporting output of line error information.
* Add support for lexing to better be able to indicate SourceLocation information.
* Fix lexer usage in DiagnosticSink in C++ extractor.
* Update diagnostics tests to have line location info.
* Fixed test expected output that now have source location information in them.
* Better handling of tab.
* Fix test expected results for tabbing change.
* DiagnosticLexer -> DiagnosticSink::SourceLocationLexer
Added line continuation tests.
* Fix typo.
* Added String::appendRepeatedChar
* Change to rerun tests.
* Added source locations to IR dumping.
* Output column for IR dump source loc.
* Add support for closing brace location to AST.
Use closing brace location in lowering when adding return void.
* Set the source location through SourceLoc - simplifies identifying if current loc is valid.
* Copy terminator sloc.
* Test for improved #line handling.
* Made writer the last parameter for dumpIR.
Small improvements to comments.
* Disable sloc output on dump IR by default.
* Fix issue with #line and inlining.
* Fix for output with improved #line output.
* Small comment change - mainly to kick off TC build.
Co-authored-by: Tim Foley <tfoleyNV@users.noreply.github.com>
|
|
* Add an accessor for IRInst opcode
This main changing is renaming `IRInst::op` over to `IRInst::m_op` and then adds an accessor `IRInst::getOp()` to read it. The rest of the changes are just changing use sites to `getOp` (or to `m_op` in the limited cases where we write to it).
This work is in anticipation of a future change that might need to store an extra bit in the same field as the opcode. It seemed better to do this massive refactoring as a separate PR.
* fixup
|
|
The basic feature here is the ability to use the `&` operator to produce the conjunction/intersection of two interfaces. That is, you can have interfaces:
interface IFirst { int getFirst(); }
interface ISecond { int getSecoond(); }
and if you need a generic function where the type parameter `T` must conform to *both* of these interfaces, you express that by constraining the parameter to the intersection of the interfaces:
void someFunction<T : IFirst & ISecond>(T value) { ... }
Without this feature, the main alternative an application would have is to define an intermediate interface, like:
interface IBoth : IFirst, ISecond {}
Forcing users to deal with an intermediate interface creates more work for type authors (they need to remember to inherit from the right combined interface(s)), or for `extension` authors (when you add `ISecond` to a type that used to just support `IFirst`, you had better also add `IBoth`). In the worst case, a family of N related "leaf" interfaces would give rise to an exponential number of intermediate interfaces to represnt the possible combinations.
A conjunction like `IFirst & ISecond` is officially its own type, and can be used to declare a type alias:
typealias IBoth = IFirst & ISecond;
This change only includes the first pass of work on this feature, so there are several caveats to be aware of:
* Using a conjunction as part of an inheritance clause is not yet supported (e.g., `struct X : IFirst & ISecond`). This is true even if the conjunction was introduced by an intermediate `typealias`
* The `&` syntax introduced here is only parsed in places where only a type (not an expression) is possible. This means you cannot do things like cast to a conjunction with `(IFirst & ISecond)(someValue)`.
* This work *should* apply to conjunctions of more than two interfaces (like `IA & IB & IC`) but that has not yet been tested
* In the long run it may be sensible to allow conjunctions that use concrete types, but we really ought to have the semantic checking logic rule that out for now.
* During testing, I encountered compiler crashes when trying to use this feature together with `property` declarations. Further investigation and debugging is called for.
* The handling of conjunction types is currently incomplete, in that there are many equivalences the compiler does not yet understand. For example, it is clear that `IA & IB` is equivalent to `IB & IA`, but the compiler currently does not understand this and will treat them as different types. A deeper implementation approach is called for.
* Conjunctions are currently only supported for generic type parameter constraints, when performing full specialization. Use of conjunctions for existential-type value parameters or with dynamic dispatch is not yet supported.
|
|
* Use "capability" system to select VKRT extension
Slang currently supports translation of ray tracing shader code to Vulkan GLSL code that uses the `GL_NV_ray_tracing` extension. A multi-vendor equivalent of that extension has been released as `GL_EXT_ray_tracing` and we want Slang to support that extension as well.
At the simplest, making the change from one extension to the other is just a matter of changing a few strings, since it does not appear that anything of significance was changed at the GLSL level (or even in SPIR-V). Where this gets trickier is when we have users who want us to support *both* extensions, and to be able to switch between them.
The solution we've implemented here more or less amounts to:
* If you don't tell the compiler which extension to use, it will default to `GL_EXT_ray_tracing` (the newer multi-vendor one).
* If you explicitly want the older extension, you can opt into it using the `-profile` option or via a new API for explicitly adding capabilities to your target.
Making that work required a few different kinds of changes:
* The options parsing and public API needed ways to add optional capabilities to a target.
* During GLSL code emit, we can check the capabilities that were added to the target to see if the `GL_NV_ray_tracing` extension was explicitly enabled and, if not, default to using the `GL_EXT_ray_tracing` names for things. This step is needed because some of the modifiers/attributes involved in the extension have to be handled explicitly in the code generator rather than implicitly as part of mapping intrinsic functions.
* We add two different translations to the relevant operatiosn in the stdlib, one marked with each of the extensions. If profile/capability-based overload resolution can be relied on to pick the right one, this should Just Work.
* Next, a bunch of work had to go into making capability-based overloading Just Work for the purposes of this change. There's been a nearly complete reworking of the implementation of `CapabilitySet` here to make it more suitable for our needs.
* The tests that were using ray tracing translation for Vulkan needed to be updated. For some of them I updated their baselines to use `GL_EXT_ray_tracing` so that they can test the new path. For others, I updated the command line for the test case so that it explicitly opts into using `GL_NV_ray_tracing`. The result is that we have some coverage of each extension. I would have liked to have each test run in both modes, but our pass-through glslang support doesn't support `-D` options, so I couldn't take that step easily.
This change does *not* add support for `GL_EXT_ray_query`, the extension that supports "DXR 1.1" style queries under Vulkan. Adding support for that extension should hopefully be a smaller step because it doesn't have the same multiple-extensions issue.
This change does *not* address a lot of possible avenues for improvement or cleanup around the capability system. It focuses only on those changes that are necessary to make the ray tracing feature work and leaves the rest for future work.
* fixup: infinite loop
* Comment-only change to retrigger TC build
|
|
* Add first steps toward a "capability" system
We already have cases in the stdlib where we mark declarations as being specific to certain targets, e.g.:
```
// My ordinary function to add two numbers.
// Works everywhere.
//
void myFunc(int a, int b) { return a + b; }
// On the "coolgpu" target, we can use a secret intrinsic
// that adds numbers even faster!
//
__specialized_for_target(coolgpu)
void myFunc(int a, int b) { return __secretIntrinsic(a, b); }
```
The existing logic for dealing with these modifiers (`__specialized_for_target` and `__target_intrinsic`) was almost entirely string-based. We would turn the chosen compilation target into a string, and then use that to try and search for the "best" definition of a function at a few steps:
* During IR linking, we always pick one definition of an `[import]`ed function, and that definition will be the one with the "best" target-specialization modifier (if any)
* During final code generation, we always look up the "best" target-intrinsic modifier, and use it as the template for the code we output.
This change preserves the basic flow there, but replaces the ad hoc string-based logic with something a bit more principled, in terms of a new `CapabilitySet` type.
A `CapabilitySet` represents a set of zero or more atomic features (here represented as `CapabilityAtom`s). What a `CapabilitySet` means depends on how and where it is used:
* A compilation target implies a `CapabilitySet` where the contents of the set are the features the target *supports*.
* A `CapabilitySet` attached to a declaration (or a modifier on that declaration) describes a set of feature that declaration *requires*.
The current implementation of `CapabilitySet` is wasteful and inefficient, but that is something we can iterate on over time.
In practice, most of the current code only ever uses capability sets that are either empty (because they represent a function with no specific requirements) or singleton (because they represent asingle atomic capability like "is a GLSL target," "is an HLSL target," etc.).
The main goal here was to put in the skeleton of a new system, including some of the features it might need down the line, and then to leave changes that eventually use the greater flexibility for later. Eventually, the capability system should encompass:
* Differences between shader model versions, GLSL versions, SPIR-V versions, etc. (currently tracked with other modifiers)
* Optional extensions, and functions that are made available only with certain extensions (currently tracked with other modifiers)
* Front-end checking that the call graph of a program doesn't violate any capability-requirements (e.g., having a GLSL+HLSL portable function call a GLSL-only subroutine)
* Hypothetically we can also try to fold stage-specific (vertex-only, fragment-only, etc.) functions into this system, but doing so would require more linker cleverness if we allow overloading on stages (since we might have to clone a caller if it calls through to a callee with multiple stage-specific versions)
One important complication that the system has to deal with just because of the "do what I mean" nature of the current compiler is that somethings a current Slang user might compile for target X and specify version N, but then use a function that actually requires version N+1 of that target. Currently the Slang compiler silently "upgrades" the version(s) used by user code in these cases, because it is often what users want in cross-compilation scenarios.
Dealing with the "silent upgrade" situation requires us to be a little careful and sometimes pick a "best" capability set that doesn't appear to be supported on our target. Refining that system and potentially getting rid of the "do what I mean" behavior over time could be a goal for future changes.
* fixup: handle case where value is incompatible during linking
|
|
Overview
========
Prior to this change, we had two different code generation strategies for interface/existential types in Slang, that didn't always play nicely together:
* The "legacy" static specialization approach could handle plugging in an arbitrary concrete type for an existential type parameter (including types with resources, etc.), but wouldn't work well with things like a `StructuredBuffer<>` of an interface type, and requires somewhat counter-intuitive layout rules to make work.
* The new dynamic dispatch approach produces simpler, more easily understood layouts by assuming that values of interface type can fit into a fixed number of bytes. The tradeoff there is that it cannot handle types that include resources (only POD types).
The goal of this change is to make it so that the two strategies can co-exist. In particular, in cases where a shader is amenable to both static specialization and dynamic dispatch, the type layouts should agree.
In order to make the type layouts agree, we:
* Declare that *all* values of existential type reserve storage according to the dynamic-dispatch rules (so 16 bytes for the RTTI and witness-table information, plus whatever bytes are needed to story "any value" of a conforming type).
* Then we modify the "legacy" layout rules so that if a value of concrete type can fit in the reserved "any value" space for a given interface, then it is laid out there exactly like the dynamic dispatch rules would do. Otherwise, we fall back to the previous legacy rules (since we don't need to agree with the dynamic-dispatch layout on types that can't be used with dynamic dispatch).
Details
=======
* Renamed `ExistentialBox` to `BoundInterfaceType` to better clarify how it relates to `BindExistentialsType`
* Unconditionally apply the `lowerGenerics` pass during emit, since it is now responsible for aspects of the lowering of existential types when specialization is used.
* Made IR type layout take the target into account, so that the layout of resource types can vary by target (e.g., being POD on some targets, and invalid on others)
* Cleaned up some issues around using global shader parameters as the "key" for their layout information in the global-scope layout (only comes up when there are global-scope `uniform` parameters)
* Made there be a default any-value size (16) instead of making it be an error to leave out. This was the simplest option; we could try to go back to having an error, but we'd need to only issue it if we are sure a type/interface is being used with dynamic dispatch, since static dispatch doesn't have to obey the restrictions.
* Changed lowering of existential types to tuples so that bound interfaces where the concrete type won't fit use a "pseudo-pointer" instead of an "any-value" to hold the payload
* Changed IR type legalization to handle the "pseudo-pointer" case and apply layout information from an interface type over to the payload part when static specialization was used.
* Changed some details of how witness tables were being lowered, so that we didn't have to create "proxy" witness tables for the constraints on associated types (just use the actual requirement entries we generate)
* Changed witness tables so that they know the subtype doing the conforming
* Added logic so that we don't generate pack/unpack logic and witness table wrapper functions for types that are incompatible with any-value/dynamic dispatch for a given interface.
* Changed the core AST-level type layout logic to use the dynamic-dispatch layout in case things fit, and the legacy static specialization case when things don't (while also reserving space for the dynamic-dispatch fields)
* Changed a bunch of test cases for static specialization to properly use the new layout (which introduces new buffers in some cases, and moves data around in others).
Future Work
===========
The experience of trying to reconcile our older way of handling interface-type specialization with our newer model (that supports dynamic dispatch) makes it clear that we really need to make similar changes to our handling of generic type parameters on entry points and at the global scope.
A future change should make it so that a global type parameter is lowered with a type layout similar to a value parameter of interface type, including the RTTI and witness-table pieces, and just leaving out the "any value" piece. A similar translation strategy should apply to entry-point generic parameters (mirroring how we lower generic functions for dynamic dispatch already), and value specialization parameters.
Co-authored-by: Yong He <yonghe@outlook.com>
|
|
* Use integer RTTI/witness handles in existential tuples.
* Fix clang error.
* Fix IR serialization to use 16bits for opcode.
* Undo accidental comment change.
* Use variable length encoding for opcode.
* Fix compile error.
* Fixing issues
* Fix code review issues.
|
|
* Specialize witness table lookups.
* Remove generated files from vcxproj
* Fix call to generic interface methods.
|
|
The Slang IR builder has a notion of "hoistable" instructions, which are basically those instructions that represent a pure side-effect-free operation on their operands, and which can and should be deduplicated. Most types are "hoistable" instructions.
In order to make deduplication of hoistable instructions work, we need to emit them at the right location. Consider if we had:
```hlsl
void myFunc<T>(...)
{
if(someCondition) {
vector<T, 4> a = ...;
...
} else {
vector<T, 4> b = ...;
}
}
```
The IR instruction that represents `vector<T,4>` can't be inserted at the global scope, because then the parameter `T` would not be visible to it. That instruction also shouldn't be inserted into the same block that declares `a`, because then the instruction itself wouldn't be visible at the point where `b` is declared.
The IR builder already has logic to pick the right parent instruction. In the example given, the IR instruction for `vector<T,4>` should be inserted into the body of the IR generic, but outside of the IR function that represents `myFunc`.
The problem this change fixes is that while the logic was picking the *parent* for a hoistable instruction correctly, it wasn't putting much care into pick the insertion *location*. The existing strategy amounted to:
* If the IR builder was set with an insertion location inside the chosen parent, then use that insertion location
* Otherwise, insert at the end of the chosen parent
Neither of those options is perfect. Either could lead to an instruction being inserted after one of its uses, and the second option could even lead to a type being inserted *after* the `return` instruction in a function/generic, which violates another structural invariant of our IR (that every block must end with a terminator, and terminators must only appear at the end of blocks).
This change updates the rules as follows:
* If the type of the instruction being created, or any of its operands are in the chosen parent, then insert immediately after whichever of those instructions is last in that parent.
* Otherwise, insert before the first non-decoration, non-parameter child of the chosen parent
The combined effect of these two rules is now that we insert any hoistable instruction as early as we can in its parent, without violating the structural validity rules.
(One small exception to these rules is that if the parent is the module then we don't worry about ordering and just insert at the end, since order-of-declaration isn't significant at module scope in our IR)
All of our existing tests work with this new behavior, although there could conceivably be future cases that lead to complicated breakage. For example, if a pass looks at the first "ordinary" instruction in a block and saves it to use as an insertion point for parameter, and then proceeds to manipulate code in the block before going back and inserting parameters at the chosen location, there is a chance that a hoistable instruction might have been inserted before the chosen insertion point, leading to a parameter being inserted after an ordinary instruction. In general, though, code that works like that would already be playing a dangerous game in that it is manipulating instructions in a block while assuming the first instruction will remain fixed.
This change is currently just a refactor, but the underlying issue surfaced as a bug when I made other changes in a feature branch.
|
|
Co-authored-by: Tim Foley <tim.foley.is@gmail.com>
|
|
|
|
|
|
* Specialize exsitentials parameters in struct fields.
* Cleanup.
* Handle partial existential parameter type specialization.
Co-authored-by: Yong He <yhe@nvidia.com>
|
|
* Enable all dynamic dispatch tests on CUDA.
* Fix expected cross-compile test results.
|
|
* Allow existential types in `StructuredBuffer` element type.
* Handle StructuredBuffer.Load/.Consume methods
* Clean up unnecessary changes
* Code cleanup
* Update test comment
|
|
A long-standing problem for the Slang implementation has been that some targets (notably GLSL/SPIR-V) do not support treating resources (textures, buffers, samplers, etc.) as first-class types. Resource types on such platforms are restricted so that they may not be used as the type of:
1. fields of aggregate types (`struct`s)
2. local variables
3. function results or `out`/`inout` parameters
Issue (1) is handled by our "type legalization" pass today, by splitting aggregates that contain resources into separate fields/variables/parameters. Issue (2) is worked around by putting code into SSA form and promoting local variables to SSA temporaries when possible; the net result is that many local variables of texture type are eliminated (that pass is not perfect, though, and it is possible for users to get errors when it doesn't fully clean up local variables of texture type).
Issue (3) is a much more complicated matter, and it is what this change is concerned with.
A typical solution to issue (3) is to simply inline all of the code in a program, at which point function results and `out`/`inout` parameters will no longer exist to cause problems. We reject such solutions for two reasons. First, there are limitations on control-flow structure in HLSL/GLSL/SPIR-V that mean they cannot express certain programs after inlining has been performed. Second, and more importantly, the philosophy of the Slang compiler is to perform as little duplication of code as possible, so that we do not accidentally contribute to binary size bloat.
Instead, this change tackles the problem of functions that output resource types by adding a new specialization pass. The pass detects functions that ought to be specialized (because they have resource-type outputs), and inspects their bodies to see if the values they output have a predicatable structure that can be replicated outside of the function body. The same logic that inspects the function body also rewrites (a copy of) the function to not have the offending outputs. Finally, all the call sites to a function that is rewritten in this way also get rewritten so that instead of using output values from the function itself, they reproduce the expected output value(s) in their own code.
The pass as presented here is intentionally limited in the scope of what it can optimize away (and the test case only touches on that specific functionality). The goal is to get a basic version of this pass in place and evaluated, and then to expand on its functionality incrementally over time.
|
|
* Allow mixing unspecialized and specialized existential parameters.
* Fixes.
|
