| Age | Commit message (Collapse) | Author |
|---|
|
* `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.
|
|
* Enable lower-generics pass universally.
* Exclude builtin interfaces and functions from lower-generics pass.
* Update stdlib.
* Fixup.
* Fixes handling of nested intrinsic generic functions.
* Fixes.
* Fixes.
|
|
* Support initializing an existential value from a generic value.
* Remove trailing spaces and clean up debugging code.
|
|
* GPU Foreach Loop
This PR introduces the completed GPU foreach loop and updates the
heterogeneous-hello-world example to use it. This PR builds on the
previous introduction of the GPU Foreach loop parsing and semantic
checking PR (#1482) by introducing IR lowering and emmitting. THe new
feature can be used by having a GPU_Foreach loop interacting with a
named non-CPP entry point, and using the -heterogeneous flag.
* Fix to path
Co-authored-by: Tim Foley <tfoleyNV@users.noreply.github.com>
|
|
* Tuple types.
* Fix x86 warning
* Improved deduplication
Co-authored-by: Tim Foley <tfoleyNV@users.noreply.github.com>
|
|
* AnyValue packing/unpacking pass.
* Add diagnostic for types that does not fit in required AnyValueSize.
* Add expected test result
* Fix warnings.
|
|
* AnyValue based dynamic code gen
* Fix aarch64 build error
|
|
Co-authored-by: Tim Foley <tfoleyNV@users.noreply.github.com>
|
|
This change includes a few different fixes for issues that arose in a user shader that made use of DXR 1.1.
The existing solution we had for handling the DXR 1.1 `RayQuery` type relied on the fact that a declaration like:
```hlsl
RayQuery<0> myRayQuery;
```
Looks like an undefined variable to existing Slang, while to dxc it is a variable declaration that runs an implicit default constructor (sneaking a bit of C++ into HLSL, but only in a way the standard library can use).
Slang was getting away with the fact that this maps to an undefined variable because it turns out that our emit logic would output the exact same declaration for an undefined value (since declaring a variable without initializing it is the simplest way to get an undefined value of a given type in a C-like language).
The main bug that arose here was that if the `RayQuery<...>`-typed variable was declared under control flow, then the `undefined` instructions introduced by our SSA pass would actually get inserted into the wrong block. Basically, when a block was trying to read a variable, and there was no preceding `store` to that variable in the block, we'd start looking for incoming values from its predecessor block(s). In the case where the variable *never* gets stored to, this search would eventually reach the first block of the function, where we'd realize the value must be `undefined`. The result was that we might insert an `undefined` instruction of some `T` into the first block of a function, but the type `T` might be the result of a lookup operation performed later in that function. This ends up creating a use of `T` that isn't dominated by the definition, which violates the SSA property. This violation of the SSA properties lead to us generating incorrect code in a later pass that deals with scoping differences between SSA form and our structured output statements; that code would end up creating a local variable to hold a *type* instead of a value.
The main fix is in `slang-ir-ssa.cpp`, where we catch the case of trying to read a variable in the block that declares it, if there we no preceding `store`s. We simply insert an `undefined` instruction before the first such read and write that out as the value of the variable to be used for subsequent instructions (up to the next `store`). This fixes the SSA dominance property for the `undefined` values that get introduced and thus technically fixes the output code for the user shader.
A secondary issue is that it is kind of gross to be relying on the behavior of `undefined` instructions in the IR for the semantics of an important standard library type like `RayQuery<...>`. A preceding change already added basic support for Slang to run default initializers (declared as `__init()`) on variables that are declared without an initial-value expression. This change adds such a default initializer to `RayQuery<...>` and maps it to a dedicated IR instruction that is intended to represent the idea of running a C++-style default constructor to produce a value. It turns out that the code we need to emit in that cse is identical to what we currently emit for `undefined` instructions, so that is helpful.
A tertiary issue is that when trying to run the user shader in debug mode, I ran into an assertion because our type layout logic for reflection had never dealt with the issue of user-defined `enum` types being used in constant buffers or other memory that needs layout. I added a quick fix that lays out any `enum` types as their "tag type" (which defaults to `int`).
Unfortunately, there is no easy way to check in a regression test for the user issue, because official `dxcompiler` versions with support for DXR 1.1 are not yet released (at least as of last time I checked).
|
|
* Ensure labels are dumped in `lower-to-ir`.
There is a `dumpIR` function that accepts a label parameter already in slang-emit.cpp. This change moves it to slang-ir.cpp so it may be called from other files.
* update expected test result
Co-authored-by: Yong He <yhe@nvidia.com>
Co-authored-by: Tim Foley <tfoleyNV@users.noreply.github.com>
|
|
* Run SSA pass to clean up generic temporary variables during lowering.
* Fix `undefined` emitting logic.
* revert dumpir control flag
* Defer fold decision of `undefined` values after special case logic for GLSL and HLSL.
* Update expected test result.
* Manually update raygen.slang.glsl to minimize change.
* fix formatting
Co-authored-by: Tim Foley <tfoleyNV@users.noreply.github.com>
|
|
|
|
* Dynamic code gen for functions returning generic types.
* Add expected test result.
|
|
The main change here is that the CPU and CUDA C++ emit paths now rely on an earlier IR pass to legalize the varying parameter list of a kernel and translate references to varying parameters with semantics like `SV_DispatchThreadID`. Doing so removes a lot of special-case logic from the emit passes.
This work moves us even closer to being able to eliminate `KernelContext` from the CPU/CUDA emit logic, because it removes the issue of state related to varying inputs being stored in `KernelContext`.
The new pass that handles the legalization is in `slang-ir-legalize-varying-params.cpp`, and it borrows heavily from the existing `slang-ir-glsl-legalize.cpp` pass. The new pass factors out the target-independent and target-dependent logic, so that both CPU and CUDA can share much of the same code despite having very different rules for how the system-value parameters are being provided.
An eventual goal is to have the new pass also handle the GLSL case, but doing so requires copying even more logic out of the GLSL-specific pass, and doing so seemed like a step to far for what was meant to be a stepping-stone change as part of other work. As a result of the incomplete nature of the pass, certain cases don't work for compute shader inputs for CPU/CUDA (e.g., wrapping your varying inputs in a `struct` type parameter), but those were cases that also didn't work in the existing `emit`-based logic.
One major consequence of this change is that the logic for emitting the various different functions that represent an entry point for our CPU back-end has been streamlined and simplified. The original logic had a fair bit of cleverness built in to try and avoid unnecessary math ops when computing the various IDs/indices, while the new logic is much more simplistic (the main dispatch function loops over threadgroups with a triply-nested `for` and then delegates to the group-level function loops over threads with its own nested `for`s).
Longer term, it will be important to simplify the CPU functions we emit further, by eliminating things like the `_Thread` function that should never really be exposed to users (the minimum granularity of invoking a CPU compute kernel should be a single threadgroup). We may eventually decide to synthesize all of the extra code that is being generated in the `emit` pass as IR instead.
|
|
* Dynamic code gen for generic local variables.
* Fixes to function calls with generic typed `in` argument.
* Fixes per code review comments
|
|
* Adding support for global uniform shader parameters
This change adds support for Slang programmers to declare shader parameters of "ordinary" types at global scope:
```hlsl
uniform float gScaleFactor;
void main() { ... *= gScaleFactor; ... }
```
The generated HLSL/GLSL/DXIL/SPIR-V output will be something along the lines of:
```hlsl
struct GlobalParams
{
float gScaleFactor;
}
cbuffer globalParams
{
GlobalParams globalParams;
}
void main() { ... *= globalParams.gScaleFactor; ... }
```
The binding information used for the implicit `globalParams` constant buffer will be determined by the existing implicit parameter binding logic (which already had support for this kind of transformation).
The reason this change is being pursued right now is because it is one step toward removing the implicit `KernelContext` type that is used to wrap the generated code for our CPU and CUDA C++ targets. Handling global-scope parameters of ordinary type requires an IR pass that synthesizes the `GlobalParams` structure type above, and that step ends up removing the need for the similar `UniformState` structure that was being used in the CPU/CUDA emit logic.
A more detailed guide to the changes included follows:
* The diagnostic for a global-scope variable that is implicitly a shader parameter was kept, but changed to a warning. Users can opt out of the warning by decorating their parameter as a `uniform` (since that keyword is already being used to mark entry-point parameters that should be treated as uniform shader parameters).
* To simplify the task of finding the global shader parameters, the `CLikeSourceEmitter` type has been given an `m_irModule` member. The previous emit logic for `UniformState` was having to do a roundabout solution involving the `EmitAction`s to deal with not having direct access to the module.
* Removed a few dead declarations in the emit logic (related to a much earlier point where emit was based on the AST instead of the IR).
* Made the computation of type names in C++ emit take into account `ConstantBuffer<T>` and `ParameterBlock<T>`. As far as I can tell, these were being handled with some special-case hacks in the emit logic instead of being supported more fundamentally. It might actually be good to pass these through as `ConstantBuffer<T>` and `ParameterBlock<T>` in the C++ output, and allow the prelude to customize their translation (defaulting to defining them as `T*`).
* Removed the special-case C++ emit logic for references to global shader parameters. There are now at most two global shader parameters to deal with, and the default emit logic (referring to them by name) does the Right Thing.
* Changed the handling of entry points for C++ (both CPU and CUDA) so that it handles the bundled-up shader paameters for the global and entry-point scopes the same way. The main complication here is OptiX, where parameter data is passed very differently than it is for CUDA compute kernels.
* Reverted changes to `ir-entry-point-uniforms` that had made its logic depend on the compilation target. The parameter binding logic was already responsible for deciding if a given target needed to wrap up its entry-point parameters in a constant buffer, and the IR pass was respecting that layout information. The current workaround had been removing the `ConstantBuffer<T>` indirection from this IR pass for CPU/CUDA, but then reintroducing the same indirection later on in the emit step.
* Added an explicit IR pass with the task of collecting global-scope parameters of uniform/ordinary type and packaging them up into a `struct`, and then optionally packaging that `struct` up in a constant buffer. This pass bases its decisions on the IR layout information that was already computed, so it should match whatever policy choices were made at the layout level.
* Changed the "key" operand on IR `struct` layout information to not assume an `IRStructKey`. The problem here is that the global scope gets a `StructTypeLayout` to represent its members, and this is convenient (rather than having to always special-case logic that handles the global scope), but the "fields" of that struct are global variables which do not have `IRStructKey`s associated with them. The simplest solution is to use the variables themselves as the keys, which required removing the assumption in the IR encoding.
* Updated the IR layout process to compute a layout for the global scope of an entire program, and to attach that to the `IRModule` via a decoration. Updated the IR linking process to carry through that decoration to the linked output. This is necessary so that the IR pass that transforms global parameters can access the global-scope layout information.
An important concern with this approach is that the contents and layout of the monolithic `GlobalParams` structure depends on the exact set of modules that were linked (and the order in which they were specified, in some cases). This isn't really a new thing with this change, but it becomes more important as we start to think of how to generalize things to better support separate compilation and linking.
There are changes that can (and should) be made to the way that IR layouts are computed for programs (e.g., so that we compute layout per-module and then combine them rather than as a whole-program step). In this case, the problem of forming the combined/linked global layout can be moved down the IR level and not be reliant on AST-level information.
Just changing the way layout and linking interact would not change the fundamental problem that global shader parameters as they currently exist in Slang/HLSL/GLSL are not readily compatible with true separate compilation. We either need to find a solution strategy that we can apply to allow existing shaders to work with separate compilation *or* we need to incrementally work toward removing support for global-scope shader parameters in favor of explicit entry-point parameters in all cases.
* fixup: missing files
* fixup: comment the new code
|
|
|
|
|
|
|
|
-Lower interfaces into actual `IRInterfaceType` insts.
-Lower `DeclRef<AssocTypeDecl>` into `IRAssociatedType`
-Generate proper IRType for generic functions.
-Add a test case exercising dynamic dispatching a generic static function through an associated type.
-Bug fixes for the test case.
|
|
|
|
* Add IR pass to lower generics into ordinary functions.
* Fix project files
* Emit dynamic C++ code for simple generics and witness tables.
Fixes #1386.
* Remove -dump-ir flag.
* Fixups.
|
|
IRWitnessTable values (#1387)
* Generate IRType for interfaces, and use them as the type of IRWitnessTable values.
This results the following IR for the included test case:
```
[export("_S3tu010IInterface7Computep1pii")]
let %1 : _ = key
[export("_ST3tu010IInterface")]
[nameHint("IInterface")]
interface %IInterface : _(%1);
[export("_S3tu04Impl7Computep1pii")]
[nameHint("Impl.Compute")]
func %Implx5FCompute : Func(Int, Int)
{
block %2(
[nameHint("inVal")]
param %inVal : Int):
let %3 : Int = mul(%inVal, %inVal)
return_val(%3)
}
[export("_SW3tu04Impl3tu010IInterface")]
witness_table %4 : %IInterface
{
witness_table_entry(%1,%Implx5FCompute)
}
```
* Fixes per code review comments.
Moved interface type reference in IRWitnessTable from their type to operand[0].
* Fix typo in comment.
|
|
* Add a ASTBuilder to a Module
Only construct on valid ASTBuilder (was being called on nullptr on occassion)
* Add nodes to ASTBuilder.
* Compiles with RefPtr removed from AST node types.
* Initialize all AST node pointer variables in headers to nullptr;
* Initialize AST node variables as nullptr.
Make ASTBuilder keep a ref on node types.
Make SyntaxParseCallback returns a NodeBase
* Don't release canonicalType on dtor (managed by ASTBuilder).
* Give ASTBuilders a name and id, to help in debugging.
For now destroy the session TypeCache, to stop it holding things released when the compile request destroys ASTBuilders.
* Moved the TypeCheckingCache over to Linkage from Session.
* NodeBase no longer derived from RefObject.
* Only add/dtor nodes that need destruction.
First pass compile on linux.
|
|
* Small improvements to documentation and code around DiagnosticSink
* Made methods/functions in slang-syntax.h be lowerCamel
Removed some commented out source (was placed elsewhere in code)
* Making AST related methods and function lowerCamel.
Made IsLeftValue -> isLeftValue.
|
|
* Synthesize "active mask" for CUDA
The Big Picture
===============
The most important change here is to `hlsl.meta.slang`, where the declaration of `WaveGetActiveMask()` is changed so that instead of mapping to `__activemask()` on CUDA (which is semantically incorrect) it maps to a dedicated IR instruction.
The other `WaveActive*()` intrinsics that make use of the implicit "active mask" concept had already been changed in #1336 so that they explicitly translate to call the equivalent `WaveMask*()` intrinsic with the result of `WaveGetActiveMask()`. As a result, all of the `WaveActive*()` functions are now no different from a user-defined function that uses `WaveGetActiveMask()`.
The bulk of the work in this change goes into an IR pass to replace the new instruction for getting the active mask gets replaced with appropriately computed values before we generate output CUDA code. That work is in `slang-ir-synthesize-active-mask.{h,cpp}`.
Utilities
=========
There are a few pieces of code that were helpful in writing the main pass but that can be explained separately:
* IR instructions were added corresponding to the Slang `WaveMaskBallot()` and `WaveMaskMatch()` functions, which map to the CUDA `__ballot_sync()` and `__match_any_sync()` operations, respectively. These are only implemented for the CUDA target because they are only being generated as part of our CUDA-only pass.
* The `IRDominatorTree` type was updated to make it a bit more robust in the presence of unreachable blocks in the CFG. It is possible that the same ends could be achieved more efficiently by folding the corner cases into the main logic, but I went ahead and made things very explicit for now.
* I added an `IREdge` utility type to better encapsulate the way that certain code operating on the predecessors/successors of an `IRBlock` were using an `IRUse*` to represent a control-flow edge. The `IREdge` type makes the logic of those operations more explicit. A future change should proably change it so that `IRBlock::getPredecessors()` and `getSuccessors()` are instead `getIncomingEdges()` and `getOutgoingEdges()` and work as iterators over `IREdge` values, given the way that the predecessor and successor lists today can contain duplicates.
* Using the above `IREdge` type, the logic for detecting and break critical edges was broken down into something that is a bit more clear (I hope), and that also factors out the breaking of an edge (by inserting a block along it) into a reusable subroutine.
The Main Pass
=============
The implementation of the new pass is in `slang-ir-synthesize-active-mask.cpp`, and that file attempts to include enough comments to make the logic clear. A brief summary for the benefit of the commit history:
* The first order of business is to identify functions that need to have the active mask value piped into them, and to add an additional parameter to them so that the active mask is passed down explicitly. Call sites are adjusted to pass down the active mask which can then result in new functions being identified as needing the active mask.
* The next challenge is for a function that uses the active mask, to compute the active mask value to use in each basic block. The entry block can easily use the active mask value that was passed in, while other blocks need more work.
* When doing a conditional branch, we can compute the new mask for the block we branch to as a function of the existing mask and the branch condition. E.g., the value `WaveMaskBallot(existingMask, condition)` can be used as the mask for the "then" block of an `if` statement.
* When control flow paths need to "reconverge" at a point after a structured control-flow statement, we need to insert logic to synchronize and re-build the mask that will execute after the statement, while also excluding any lanes/threads that exited the statement in other ways (e.g., an early `return` from the function).
The explanation here is fairly hand-wavy, but the actual pass uses much more crisp definitions, so the code itself should be inspected if you care about the details.
Tests
=====
The tests for the new feature are all under `tests/hlsl-intrinsic/active-mask/`. Most of them stress a single control-flow construct (`if`, `switch`, or loop) and write out the value of `WaveGetActiveMask()` at various points in the code.
In practice, our definition of the active mask doesn't always agree with what D3D/Vulkan implementations seem to produce in practice, and as a result a certain amount of effort has gone into adding tweaks to the tests that force them to produce the expected output on existing graphics APIs. These tweaks usually amount to introducing conditional branches that aren't actually conditional in practice (the branch condition is always `true` or always `false` at runtime), in order to trick some simplistic analysis approaches that downstream compilers seem to employ.
One test case currently fails on our CUDA target (`switch-trivial-fallthrough.slang`) and has been disabled. This is an expected failure, because making it produce the expected value requires a bit of detailed/careful coding that would add a lot of additional complexity to this change. It seemed better to leave that as future work.
Future Work
===========
* As discussed under "Tests" above, the handling of simple `switch` statements in the current pass is incomplete.
* There's an entire can of worms to be dealt with around the handling of fall-through for `switch`.
* The current work also doesn't handle `discard` statements, which is unimportant right now (CUDA doesn't have fragment shaders), but might matter if we decide to synthesize masks for other targets. Similar work would probably be needed if we ever have `throw` or other non-local control flow that crosses function boundaries.
* An important optimization opportunity is being left on the floor in this change. When block that comes "after" a structured control-flow region (which is encoded explicitly in Slang IR and SPIR-V) post-dominates the entry block of the region, then we know that the active mask when exiting the region must be the same as the mask when entering the region, and there is no need to insert explicit code to cause "re-convergence." This should be addressed in a follow-on change once we add code to Slang for computing a post-dominator tree from a function CFG.
* Related to the above, the decision-making around whether a basic block "needs" the active mask is perhaps too conservative, since it decides that any block that precedes one needing the active mask also needs it. This isn't true in cases where the active mask for a merge block can be inferred by post-dominance (as described above), so that the blocks that branch to it don't need to compute an active mask at all.
* If/when we extend the CPU target to support these operations (along with SIMD code generation, I assume), we will also need to synthesize an active mask on that platform, but the approach taken here (which pretty much relies on support for CUDA "cooperative groups") wouldn't seem to apply in the SIMD case.
* Similarly, the approach taken to computing the active mask here requires a new enough CUDA SM architecture version to support explicit cooperative groups. If we want to run on older CUDA-supporting architectures, we will need a new and potentially very different strategy.
* Because the new pass here changes the signature of functions that require the active mask (and not those that don't), it creates possible problems for generating code that uses dynamic dispatch (via function pointers). In principle, we need to know at a call site whether or not the callee uses the active mask. There are multiple possible solutions to this problem, and they'd need to be worked through before we can make the implicit active mask and dynamic dispatch be mutually compatible.
* Related to changing function signatures: no effort is made in this pass to clean up the IR type of the functions it modifies, so there could technically be mismatches between the IR type of a function and its actual signature. If/when this causes problems for downstream passes we probably need to do some cleanup.
* fixup: backslash-escaped lines
I did some "ASCII art" sorts of diagrams to explain cases in the CFG, and some of those diagrams used backslash (`\`) characters as the last character on the line, causing them to count as escaped newlines for C/C++.
The gcc compiler apparently balked at those lines, since they made some of the single-line comments into multi-line comments.
I solved the problem by adding a terminating column of `|` characters at the end of each line that was part of an ASCII art diagram.
* fixup: typos
Co-authored-by: jsmall-nvidia <jsmall@nvidia.com>
|
|
* Fields from upper to lower case in slang-ast-decl.h
* Lower camel field names in slang-ast-stmt.h
* Fix fields in slang-ast-expr.h
* slang-ast-type.h make fields lowerCamel.
* slang-ast-base.h members functions lowerCamel.
* Method names in slang-ast-type.h to lowerCamel.
* GetCanonicalType -> getCanonicalType
* Substitute -> substitute
* Equals -> equals
ToString -> toString
* ParentDecl -> parentDecl
Members -> members
* * Make hash code types explicit
* Use HashCode as return type of GetHashCode
* Added conversion from double to int64_t
* Split Stable from other hash functions
* toHash32/64 to convert a HashCode to the other styles.
GetHashCode32/64 -> getHashCode32/64
GetStableHashCode32/64 -> getStableHashCode32/64
* Other Get/Stable/HashCode32/64 fixes
* GetHashCode -> getHashCode
* Equals -> equals
* CreateCanonicalType -> createCanonicalType
* Catches of polymorphic types should be through references otherwise slicing can occur.
* Fixes for newer verison of gcc.
Fix hashing problem on gcc for Dictionary.
* Another fix for GetHashPos
* Fix signed issue around GetHashPos
|
|
* Extractor builds without any reference to syntax (as it will be helping to produce this!).
* Change macros to include the super class.
* Added indexOf(const UnownedSubString& in) to UnownedSubString.
Refactored extractor
* Output a macro for each type with the extracted info - can be used during injection in class
* Simplify the header file - as can get super type and last from macro now
* Store the 'origin' of a definition
* Some small tidy ups to the extractor.
* Improve comments on the extractor options.
* Made CPPExtractor own SourceOrigins
* Small fixes around SourceOrigin.
* Small tidy up around macroOrign
|
