| Age | Commit message (Collapse) | Author |
|---|
|
* Remove KernelContext wrapper from CPU/CUDA emit
Currently, the CPU and CUDA C++ targets rely on a `KernelContext` type that is generated during emit, as a way to provide implicit access to things that were global in the input Slang code, but that can't actually be emitted as globals in the target language (because the semantics of global declarations differ).
For example, input like:
```hlsl
ConstantBuffer<Stuff> gStuff; // shader parameter
groupshared int gData[1024]; // thread-group shared variable
static int gCounter = 0; // "thread-local" global-scope variable
void subroutine() { ... }
[shader("compute")] void computeMain() { ... }
```
would translate to output C++ for CPU a bit like:
```c++
struct KernelContext
{
ConstantBuffer<Stuff> gStuff;
int gData[1024];
int gCounter = 0;
void subroutine() { ... }
void computeMain() { ... }
};
```
Note that both `computeMain()` and `subroutine()` are non-`static` members functions on `KernelContext`, so they have an implicit `this` parameter of type `KernelContext`, which allows the bodies of those functions to implicitly reference `gStuff`, etc. by name in their bodies.
Because `KernelContext::computeMain()` is a member function, we end up emitting an additional global-scope function to expose the entry point to the outside world, and that function is responsible for declaring a local `KernelContext` and invoking the generated entry point on it.
This approach has several important drawbacks:
* It complicates the emit logic for CPU and CUDA, with many special cases around when/how things get emitted
* It complicates the implementation of dynamic dispatch, because what seems like a function pointer in Slang IR needs to be a pointer-to-member-function in C++.
* It makes it difficult to have a non-kernel-oriented mode of compilation for CPU where a Slang function with a given signature gets output as a C++ CPU function with the "same" signature (not wrapped up as a member function of `KernelContext`.
This change makes a step toward addressing these issues by making the introducing of the `KernelContext` type be something that is done in an explicit IR pass instead of being handled as part of the last-mile emit logic.
The most important change is the removal of code related to `KernelContext` from the `slang-emit-{cpp,cuda}.{h,cpp}` files, with the equivalent logic instead being handled in a new pass in `slang-ir-explicit-global-context.{h,cpp}`. It should be noted that further cleanups to the emit logic should now be possible; in particular, both the CPU and CUDA emit paths are manually sequencing the `EmitAction`s instead of relying on the default logic, but at this point they should be able to just use the default. The additional cleanups are left for future work.
The explicit IR pass does more or less what one would expect: it identifies global-scope entities (global variables and parameters) that need to be wrapped and turns them into fields of a `KernelContext` type. It then modifies all entry points to initialize a `KernelContext` as part of their startup. Finally, any code that used to refer to the global entities is changed to refer to a field of the context, with the context passed via new function parameters (the new parameter is only added to functions that need it for now).
Transforming global variables into fields of a `KernelContext` type in the IR pass ends up dropping their initial-value expressions (since those were attached as basic blocks on the `IRGlobalVar`). To avoid breaking code that relies on global-scope (but thread-local) variables, this change also adds an explicit pass that takes the initialization logic on all global variables and moves it to explicit logic that runs at the start of every entry point in a linked module (`slang-ir-explicit-global-init.{h,cpp}`). This pass would also be useful when we get back to direct SPIR-V emit, since SPIR-V also requires initialization logic for globals to be emitted into entry points.
One complication that arises when the IR is introducing the types for entry-point parameters, global-scope parameters, and the `KernelContext` type is that it becomes harder for the emit logic to utter the names of those types (they might not even have names, since `IRNameHint`s might get stripped). This created a problem since the wrapper operations that were being generated for CPU were taking `void*` parameters and casting them to the appropriate type. To work around this issue, we have added an explicit IR pass (`slang-ir-entry-point-raw-ptr-params.{h,cpp}`) that transforms the signature of entry points so that any pointer parameters instead become raw pointer (`void*`) parameters, with the casting being handled inside the entry point itself.
One consequence of all the above changes is that for the CUDA target we no longer need a wrapper function to invoke the generated entry point any more, because the IR function for the entry point ends up having the correct/expected signature already. This is also the case for CPU when it comes to the `*_Thread` wrapper function, but this change doesn't try to eliminate the wrapper because of a belief that the `*_Thread`-level interface is going away anyway.
Because the IR is now responsible for ensuring the signature of the IR entry point for CUDA and CPU is what is expected, I needed to modify the `slang-ir-entry-point-uniforms` pass to always create an explicit parameter for the entry point uniforms when compiling for CUDA/CPU, even if there were no `uniform` parameters on the entry point as written. This also ended up requiring some tweaks to the parameter layout logic to ensure that CPU/CUDA targets always treat `ConstantBuffer<T>` as a `T*` even in the case where `T` is an empty `struct` type (which happens when we construct a `struct` type to represent the uniform parameters of an entry point with no uniform parameters...).
There are several future changes that can/should build on this work:
* We should change the generated signatures for CUDA kernels, so that they don't rely on `KernelContext` for global-scope parameters. At that point we can avoid generating a `KernelContext` at all for CUDA, except when a program uses global-scope thread-local variables.
* We should figure out how to make the "ABI" for dynamic-dispatch calls ensure that the kernel context is either always passed, or always *not* passed. Making a hard-and-fast rule as part of the calling convention for dynamic calls would ensure that they access through the context continues to work with dynamic calls (this change might break it in some cases).
* We should figure out how to handle the layout for the `KernelContext` in cases where a program is composed of multiple separately-compiled modules. Right now the layout of the `KernelContext` requires global knowledge (as does the pass that introduces explicit initialization for global-scope thread-locals).
* We should try to further clean up the CPU/CUDA C++ emit logic to fall back on the default emit behavior more, now that the various special-case approaches that were taken are no longer needed
* fixup: restore build files to default configuration
|
|
* Dynamic code gen for generic local variables.
* Fixes to function calls with generic typed `in` argument.
* Fixes per code review comments
|
|
Co-authored-by: Yong He <yhe@nvidia.com>
|
|
|
|
|
|
|
|
-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.
|
|
|
|
* Associate a downstream compiler for prelude lookup even if output is source.
* Remove LanguageStyle and just use SourceLanguage instread.
* Added set/getPrelude.
Made prelude work on source language.
* Fix typo in method name replacement.
get/SetPrelude get/setLanguagePrelude
* Fix issue because of method name change.
* Remove getPreludeDownstreamCompilerForTarget
|
|
* 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.
|
|
* Added GLSL_460 if ray tracing is used on fragment shader.
Moved GLSL specific setup init function.
* Split out _requireRayTracing method.
|
|
* 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.
|
|
* 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
|
|
* Add support for generic load/store on byte-addressed buffers
Introduction
============
The HLSL `*ByteAddressBuffer` types originaly only supported loading/storing `uint` values or vectors of the same, using `Load`/`Load2`/`Load3`/`Load4` or `Store`/`Store2`/`Store3`/`Store4`. More recent versions of dxc have added support for generic `Load<T>` and `Store<T>`, which adds a two main pieces of functionality for users.
The first and more fundamental feature is that `T` can be a type that isn't 32 bits in size (or a vector with elements of such a type), thus exposing a capability that is difficult or impossible to emulate on top of 32-bit load/store (depending on what guarantees `*StructuredBuffer` makes about the atomicity of loads/stores).
The secondary benefit of having a generic `Load<T>` and `Store<T>` is that it becomes possible to load/store types like `float` without manual bit-casting, and also becomes possible to load/store `struct` types so long as all the fields are loadable/storable.
This change adds generic `Load<T>` and `Store<T>` to the Slang standard library definition of byte-address buffers, and tries to bring those same benefits to as many targets as possible. In particular, the secondary benefits become available on all targets, including DXBC: byte-address buffers can be used to directly load/store types other than `uint`, including user-defined `struct` types, so long as all of the fields of those types can be loaded/stored.
The ability to load/store non-32-bit types depends on target capabilities, and so is only available where direct support for those types is available. For 16-bit types like `half` this includes both Vulkan and D3D12 DXIL with appropriate extensions or shader models.
The implementation is somewhat involved, so I will try to explain the pieces here.
Standard Library
================
The changes to the Slang standard library in `hlsl.meta.slang` are pretty simple. We add new `Load<T>` and `Store<T>` generic methods to `*ByteAddressBuffer`, and route them through to a new IR opcode.
Right now the generic `Load<T>` and `Store<T>` do *not* place any constraints on the type `T`, although in practice they should only work when `T` is a fixed-size type that only contains "first class"
uniform/ordinary data (so no resources, unless the target makes resource types first class). Our front-end checking cannot currently represent first-class-ness and validate it (nor can it represent fixed-size-ness), so these gaps will have to do for now.
Rather than directly translate `Load<T>` or `Store<T>` calls into a single instruction, we instead bottleneck them through internal-use-only subroutines. The design choice here is intended to ensure that for some large user-defined type like `MassiveMaterialStruct` we only emit code for loading all of its fields *once* in the output HLSL/GLSL rather than once per load site. While downstream compilers are likely to inline all of this logic anyway, we are doing what we can to avoid generating bloated code.
Emit and C++/CUDA
=================
Over in `slang-emit-c-like.cpp` we translate the new ops into output code in a straightforward way. A call like `obj.Load<Foo>(offset)` will eventually output as a call like `obj.Load<Foo>(offset)` in the generated code, by default.
For the CPU C++ and CUDA C++ codegen paths, this is enough to make a workable implementation, and we add suitable templated `Load<T>` and `Store<T>` declarations to the prelude for those targets.
Legalization
============
For targets like DXBC and GLSL there is no way to emit a load operation for an aggregate type like a `struct`, so we introduce a legalization pass on the IR that will translate our byte-address-buffer load/store ops into multiple ops that are legal for the target.
Scalarization
-------------
The big picture here is easy enough to understand: when we see a load of a `struct` type from a byte-address buffer, we translate that into loads for each of the fields, and then assemble a new `struct` value from the results. We do similar things for arrays, matrices, and optionally for vectors (depending on the target).
Bit Casting
-----------
After scalarization alone, we might have a load of a `float` or a `float3` that isn't legal for D3D11/DXBC, but that *would* be legal if we just loaded a `uint` or `uint3` and then bit-casted it. The legalization pass thus includes an option to allow for loads/stores to be translated to operate on a same-size unsigned integer type and then to bit-cast.
To make this work actually usable, I had to add some more details to the implementation of the bit-cast op during HLSL emit and, more importantly, I had to customize the way that the byte-address buffer load/store ops get emitted to HLSL so that it prefers to use the existing operations like `Load`/`Load2`/`Load3`/`Load4` instead of the generic one, whenever operating on `uint`s or vectors of `uint`.
Translation to Structured Buffers
---------------------------------
Even after scalarizing all byte-address-buffer loads/stores, we still have a problem for GLSL targets, because a single global `buffer` declaration used to back a byte-address buffer can only have a single element type (currently always `uint`), so the granularity of loads/stores it can express is fixed at declaration time. If we want to load a `half` from a byte-address buffer, we need a dedicated `buffer` declaration in the output GLSL with an element type of `half`.
The solution we employ here is to translate all byte-address buffer loads into "equivalent" structured-buffer ops when targetting GLSL. We add logic to find the underlying global shader parameter that was used for a load/store and introduce a new structured-buffer parameter with the desired element type (e.g., `half`) and then rewrite the load/store op to use that buffer instead. We copy layout information from the original buffer to the new one, so that in the output GLSL all the various `buffer`s will use a single `binding` and thus alias "for free."
We don't want to create a new global buffer for every load/store, so we try to cache these "equivalent" structured buffers as best as we can. For the caching I ended up needing a pair to use as a key, so I tweaked the `KeyValuePair<K,V>` type in `core` so that it could actually work for that purpose.
Because we are working at the level of IR instructions instead of stdlib functions at this work I had to add new IR opcodes to represent structured-buffer load/store that only (currently) apply to GLSL.
Layout
======
In order to translate a load/store of a `struct` type into per-field load/store we need a way to access layout information for the types of the fields. Previously layout information has been an AST-level concern that then gets passed down to the IR only when needed and only on global parameters, so layout information isn't always available in cases like this, at the actual load/store point.
As an expedient move for now I've introduced a dedicated module that does IR-level layout and caches its results on the IR types themselves. This approach *only* supports the "natural" layout of a type, and thus is usable for structured buffers and byte-address buffers (or general pointer load/store on targets that support it), but which is *not* usable for things like constant buffer layout.
We've known for a while that the Right Way to do layout going forward is to have an IR-based layout system, and this could either be seen as a first step toward it, or else as a gross short-term hack. YMMV.
Details
=======
The GLSL "extension tracker" stuff around type support needed to be tweaked to recognize that types like `int16_t` aren't actually available by default. I switched it from using a "black list" of unavailable types at initialization time over to using a "white list" of types that are known to always be available without any extensions.
Tests
=====
There are two tests checked in here: one for the basic case of a `struct` type that has fields that should all be natively loadable, and one that stresses 16-bit types. Each test uses both load and store operations.
Future Directions
=================
Right now we translate vector load/store to GLSL as load/store of individual scalars, which means the assumed alignment is just that of the scalars (consistent with HLSL byte-address buffer rules). We could conceivably introduce some controls to allow outputting the vector load/store ops more directly to GLSL (e.g., declaring a `buffer` of `float4`s), which might enable more efficient load/store based on the alignment rules for `buffer`s.
The IR layout work has a number of rough edges, but the most worrying is probably the assumption that all matrices are laid out in row-major order. Slang really needs an overhaul of its handling of matrices and matrix layout, so I don't know if we can do much better in the near term.
At some point the IR-based layout system needs to be reconciled with our current AST-base layout, and we need to figure out how "natural" layout and the currently computed layouts co-exist (in particular, we need to make sure that the IR-based layout and the existing layout logic for structured buffers will agree). This probably needs to come along once we have moved the core layout logic to operate on IR types instead of AST types (a change we keep talking about).
As part of this work I had to touch the implementation of bit-casting for HLSL, and it seems like that logic has some serious gaps. We really ought to consider a separate legalization pass that can turn IR bitcast instructions into the separate ops that a target actually supports so that we can implement `uint64_t`<->`double` and other conersions that are technically achievable, but which are hard to express in HLSL today.
* fixup: missing files
|
|
For a long time the various source-to-source back-ends have been emitted the text `/* unhandled */` when they encounter an IR instruction opcode that didn't have any emit logic implemented. This choice had two apparent benefits:
* In most common cases, emitting `/* unhandled */` in place of an expression would lead to downstream compilation failure, so most catastrophic cases seemed to work as desired (e.g., if we emit `int i = /* unhandled */;` we get a downstream parse error and we know something is wrong.
* In a few cases, if a dead-but-harmless instruction slips through (e.g., a type formed in the body of a specialized generic function), we would emit `/* unhandled */;`, which is a valid empty statement.
It is already clear from the above write-up that the benefits of the policy aren't really that compelling, and where it has recently turned out to be a big liability is that there are actually plenty of cases where emitting `/* unhandled */` instead of a sub-expression won't cause downstream compilation failure, and will instead silently compute incorrect results:
* Emitting `/* unhandled */ + b` instead of `a + b`
* Emitting `/* unhandled */(a)` instead of `f(a)`, or even `/* unhandled */(a, b)` instead of `f(a,b)`
* Emitting `f(/*unhandled*/)` instead of `f(a)` in cases where `f` is a built-in with both zero-argument and one-argument overloads
The right fix here is simple: where we would have emitted `/* unhandled */` to the output we should instead diagnose an internal compiler error, thus leading to compilation failure.
This change appears to pass all our current tests, but it is possible that there are going to be complicated cases in user code that were relying on the previous lax behavior. I know from experience that we sometimes see `/* unhandled */` in output for generics, and while we have eliminated many of those cases I don't have confidence we've dealt with them all.
When this change lands we should make sure that the first release that incorporates it is marked as potentially breaking for clients, and we should make sure to test the changes in the context of the client codebases before those codebases integrate the new release.
|
|
The Slang compiler was bit by a known issue when translating from SSA form back to straight-line code. Give code like the following:
int x = 0;
int y = 1;
while(...)
{
...
int t = x;
x = y;
y = t;
}
...
The SSA construction pass will eliminate the temporary `t` and yield code something like:
br(b, 0, 1);
block b(param x : Int, param y : Int):
...
br(b, y, x);
The loop-dependent variables have become parameters of the loop block, and the branchs to the top of the loop pass the appropriate values for the next iteration (e.g., the jump that starts the loop sends in `0` and `1`).
The problem comes up when translating the back-edge the continues the loop out of SSA form. Our generated code will re-introduce temporaries for `x` and `y`:
int x;
int y;
// jump into loop becomes:
x = 0;
y = 1;
for(;;)
{
...
// back-edge becomes
x = y;
y = x;
continue;
}
The problem there is that we've naively translated a branch like `br(b, <a>, <b>)` into `x = <a>; y = <b>;` but that doesn't work correctly in the case where `<b>` is `x`, because we will have already clobbered the value of `x` with `<a>`.
The simplest fix is to introduce a temporary (just like the input code had), and generate:
// back-edge becomes
int t = x;
x = y;
y = t;
This change modifies the `emitPhiVarAssignments()` function so that it detects bad cases like the above and emits temporaries to work around the problem. A new test case is included that produced incorrect output before the change, and now produces the expected results.
A secondary change is folded in here that tries to guard against a more subtle version of the problem:
for(...)
{
...
int x1 = x + 1;
int y1 = y + 1;
x = y1;
y = x1;
}
In this more complicated case, each of `x` and `y` is being assigned to a value derived from the other, but neither is being set using a block parameter directly, so the changes to `emitPhiVarAssignments()` do not apply.
The problem in this case would be if the `shouldFoldInstIntoUseSites()` logic decided to fold the computation of `x1` or `y1` into the branch instruction, resulting in:
x = y + 1;
y = x + 1;
which would again violate the semantics of the original code, because now there is an assignment to `x` before the computation of `x + 1`.
Right now it seems impossible to force this case to arise in practice, due to implementation details in how we generate IR code for loops. In particular, the block that computes the `x+1` and `y+1` values is currently always distinct from the block that branches back to the top of the loop, and we do not allow "folding" of sub-expressions from different blocks. It is possible, however, that future changes to the compiler could change the form of the IR we generate and make it possible for this problem to arise.
The right fix for this issue would be to say that we should introduce a temporary for any branch argument that "involves" a block parameter (whether directly using it or using it as a sub-expression). Unfortunately, the ad hoc approach we use for folding sub-expressions today means that testing if an operand "involves" something would be both expensive and unwieldy.
A more expedient fix is to disallow *all* folding of sub-expressions into unconditional branch instructions (the ones that can pass arguments to the target block), which is what I ended up implementing in this change. Making that defensive change alters the GLSL we output for some of our cross-compilation tests, in a way that required me to update the baseline/gold GLSL.
A better long-term fix for this whole space of issues would be to have the "de-SSA" operation be something we do explicitly on the IR. Such an IR pass would still need to be careful about the first issue addressed in this change, but the second one should (in principle) be a non-issue given that our emit/folding logic already handles code with explicit mutable local variables correctly.
|
|
* Add unroll support for CUDA, and preliminary for C++.
Document [unroll] support.
* Fix loop-unroll to run on CPU, and test on CPU and elsewhere.
Fix bug in emitting loop unroll condition.
* Improved comment.
* Added support for vk/glsl loop unrolling.
|
|
This change continues the work already started in moving the definitions of many built-in functions to the standard library.
The main focus in this change was reducing the number of operations that had to be special-cased on the CPU and CUDA targets by making sure that the scalar cases of built-in functions map to the proper names in the prelude (e.g., `F32_sin()`) via the ordinary `__target_intrinsic` mechanism. In some cases this cleanup meant that special-case logic that was constructing definitions for those functions using C++ code could be scrapped.
Additional changes made along the way:
* A few scalar functions that were missing in the CPU/CUDA preludes got added: `round`, hyperbolic trigonometric functions, `frexp`, `modf`, and `fma`
* The floating-point `min()` and `max()` definitions in the preludes were changed to use intrinsic operations on the target (which are likely to follow IEEE semantics, while our definitions did not)
* For the CUDA target, many of the functions had their names translated during code emit from, e.g., `sin` to `sinf`. This change makes the CUDA target more closely match the C++/CPU target in using names like `F32_sin` consistently.
* For the CUDA target, a few additional functions have intrinsics that don't exist (portably) on CPU: `sincos()` and `rsqrt()`.
* For the Slang stdlib definitions to work, a new `$P` replacement was defined for `__targert_intrinsic` that expands to a type based on the first operand of the function (e.g., `F32` for `float`).
* I removed the dedicated opcodes for matrix-matrix, matrix-vector, and vector-matrix multiplication, and instead turned them into ordinary functions with definitions and `__target_intrinsic` modifiers to map them appropriately for HLSL and GLSL. This is realistically how we would have implemented these if we'd had `__target_intrinsic` from the start.
Notes about possible follow-on work:
* The `ldexp` function is still left in the Slang stdlib because it has to account for a floating-point exponent and the `math.h` version only handles integers for the exponent. It is possible that we can/should define another overload for `ldexp` (and `frexp`) that uses an integer for exponent, and then have that one be a built-in on CPU/CUDA, with the HLSL `frexp` being defined in the stdlib to delegate to the correct `frexp` for those targets.
* The `firstbithigh` and related functions are missing for our CPU and CUDA targets, and will need to be added. It is worth nothing that `firstbithigh` apparently has some very odd functionality around signed integer arguments (which are supported, despite MSDN being unclear on that point). General cleanup will be required for those functions.
* Maxing the various matrix and vector products no longer be intrinsic ops might affect how we emit code for them as sub-expressions (both whether we fold them into use sites and how we parenthize them). This doesn't seem to affect any of our existing tests, but we could consider marking these functions with `[__readNone]` to ensure they can be folded, and then also adding whatever modifier(s) we might invent to control precdence and parentheses insertion during emit.
|
|
Some of the functions declared in the Slang standard library are built in on some targets (almost always the case for HLSL) but aren't available on other targets (often the case for GLSL, CUDA, and CPU). To date, the CUDA and CPU targets have worked around this issue by synthesizing definitions of the missing functions on the fly as part of output code generation, at the cost of some amount of code complexity in the emit pass.
This change adds definitions inside the stdlib itself for a large number of built-in HLSL functions that act element-wise over both vectors and matrices (e.g., `sin()`, `sqrt()`, etc.), and changes the CPU/CUDA codegen path to *not* synthesize C++ code for those functions (instead relying on code generated from the Slang definitions).
The element-wise vector/matrix function bodies are being defined using macros in the stdlib, so that we can more easily swap out the definitions en masse if we find an implementation strategy we like better. This could involve defining special-case syntax just for vector/matrix "map" operations that can lower directly to the IR and theoretically generate cleaner code after specialization is complete.
As a byproduct of this change, the matrix versions of these functions should in principle now be available to GLSL (GLSL only defines vector versions of functions like `sin()`, and leaves out matrix ones). No testing has been done to confirm this fix.
In some cases builtins were being declared with multiple declarations to split out the HLSL and GLSL cases, and this change tries to unify these as much as possible into single declarations to keep the stdlib as small as possible.
Two functions -- `sincos()` and `saturate()` -- were simple enough that their full definitions could be given in the stdlib so that even the scalar cases wouldn't need to be synthesized, so the corresponding enumerants were removed in `slang-hlsl-intrinsic-set.h`. In the case of `saturate()` the pre-existing definition used for GLSL codegen could have been used for CPU/CUDA all along.
In some cases functions that can and should be defined in the future have had commented-out bodies added as an outline for what should be inserted in the future. Most of these functions cannot be implemented directly in the stdlib today because basic operations like `operator+` are currently not defined for `T : __BuiltinArithmeticType`, etc. Adding such declarations should be straightforward, but brings risks of creating unexpected breakage, so it seemed best to leave for a future change.
This change does not try to address making vector or matrix versions of builtin functions that map to single `IROp`s, because the existing mechanisms for target-based specialization, etc., do not apply for such cases. In the future we will either have to make those operations into ordinary functions (eliminating many `IROp`s) so that stdlib definitions can apply, or add an explicit IR pass to deal with legalizing vector/matrix ops for targets that don't support them natively. The right path for this is not yet clear, so this change doesn't wade into it.
This change does not touch the `Wave*` functions added in Shader Model 6, despite many of these having vector/matrix versions that could benefit from the same default mapping. It is expected that these functions will have GLSL/Vulkan translation added soon, and it probably makes sense to know what cases are directly supported on Vulkan before adding the hand-written definitions.
Because of the limitations on what could be ported into the stdlib, it is not yet possible to remove any of the infrastructure for synthesizing builtin function definitions in the CPU and CUDA back-ends.
|
|
|
|
* Fix for GCC C++ target - remove warning for int literal value.
* Comment around hack to negate -0x8000 0000
* Special case negation of literals in parser - to fix problems with errors on gcc.
* Apply the literal integer 'fix' when doing negation of a literal.
|
|
* Test to check fix
|
|
* Launch CUDA test taking into account dispatch size.
* Enable isCPUOnly hack to work on CUDA.
* Rename 'isCPUOnly' hack to 'onlyCPULikeBinding'.
* Add $T special type.
Support SampleLevel on CUDA.
* Fix typo.
|
|
* WIP: 64 literal diagnostic and truncation.
* Improve how integer truncation is handled/supported.
Added literal-int64.slang test.
Set a suffix on all literals.
Fixed problem on C++ based targets where l suffix was not the same as int() cast. So on C++ derived emitters, int() is used instead of l suffix to have same behavior across targets.
* Add literal diagnostic testing.
* Allow lexer to lex - in front of literals.
* Fix lexing and converting int literal with -.
* Too large small values of floats become inf.
Handling writing inf types out on different targets.
Add function to deterimine if a float literals kind.
* Roll back the support of lexer lexing negative literals.
* Fixed tests broken because of diagnostics numbers.
Improved _isFinite
* Fix compilation on linux.
* Fix problem with abs on linux - use Math::Abs.
* Fix typo.
* * Improve warnings for float literals zeroed
* Improved 64 bit type documentation
* Handle half
* Improved comments
* Fixed tests broken
* Use capital letters for suffixes.
* Make default behavior on outputting a int literal that is an 'int32_t' is cast (not suffix) to avoid platform inconsistencies.
Improve documentation for 64 bit types.
Make tests cover material in docs.
* Fixed tests.
* Rename FloatKind::Normal -> Finite
* Fix half zero check.
|
|
* * For integer literals add postfix, and use unsigned/signed output appropriately
* Extend GLSL extension handling by type, and for adding 64 bit int extensions
* Added tests for int/uint64 types
* Add explicit Int/UInt64 emit functions to avoid ambiguity.
|
|
* When checking if an instruction can be folded, take into account if it's called by a target intrinsic, because if it is we need to check if the parameter is accessed multiple times to see if it's worth allowing to fold.
* Tidy up code around folding/target intrinsics.
* Fix texture-load.slang .
* Fix typo in assert.
|
|
* Fix tests/compute/global-init.slang by handling some other cases where functions are emitted.
* Fix comment.
|
|
* WIP: Trying to figure out how texturing will work with CUDA.
* WIP: Fixes for CUDA layout. Initial CUDA texture test.
* WIP: Outputs something compilable by CUDA for TextureND.Sample
* 2d texture working with CUDA.
* Fix how binding for SamplerState occurs in CUDA.
* Small tidy up of comments.
|
|
* Added hlsl-intrinsic test folder.
Enabled ceil as works across targets.
* log10 support.
* Fix float % on CPU/CUDA to match HLSL which is fmod (not fremainder).
* Added log10 tests back to scalar-float.slang
* Don't add the ( for $Sx - it's clearer what's going on without it.
* Works on CUDA/CPU. Problem with asint/asuint do not seem to be found.
* Only asuint exists for double.
* Support countbits on CUDA and C++.
* Fix typo in C++ population count.
* First pass at int vector intrinsic tests.
* Swizzle for int.
* Bit cast tests on CUDA.
* Fix warning on gcc.
* Fix bit-cast-double execution on CUDA.
* scalar-int test working on gcc release.
* GetAt working on CUDA/C++
* Split out runtime index into it's own test.
* Removed SetAt, as can use assignment with GetAt.
* Allowing getAt to be used on matrices.
* Don't need [] on matrix type any longer because use getAt.
* Enable clamp on matrix-int.
* Fix matrix-int.slang test - because clamp behavior varied if min and max were say inverted.
Added runtime indexing version of matrix-int.
|
|
* Added hlsl-intrinsic test folder.
Enabled ceil as works across targets.
* log10 support.
* Fix float % on CPU/CUDA to match HLSL which is fmod (not fremainder).
* Added log10 tests back to scalar-float.slang
* Don't add the ( for $Sx - it's clearer what's going on without it.
|
|
* CPPCompiler -> DownstreamCompiler
* Added DownstreamCompileResult to start abstraction such that we don't need files.
* * Split out slang-blob.cpp
* Made CompileResult hold a DownstreamCompileResult - for access to binary or ISlangSharedLibrary
* Keep temporary files in scope.
* Add a hash to the hex dump stream.
* Move all file tracking into DownstreamCompiler.
* WIP support for nvrtc.
* WIP: Adding support for nvrtc compiler.
Adding enum types, wiring up the nvrtc into slang.
* Fix remaining CPPCompiler references.
* Fix order issue on target string matching.
* Use ISlangSharedLibrary for nvrtc.
* Use DownstreamCompiler for nvrtc.
* WIP first pass at compilation win nvrtc.
* Added testing if file is on file system into CommandLineDownstreamCompiler.
Added sourceContentsPath.
* Make test cuda-compile.cu work by just compiling not comparing output.
* Genearlize DownstreamCompiler usage.
* Fix warning on clang.
* Remove CompilerType from DownstreamCompiler.
* Use DownstreamCompiler interface for all compilers.
NOTE for FXC, DXC and GLSLANG this doesn't mean using 'compile' - it's still extracting functions from shared library.
* Replace DownstreamCompiler::SourceType -> SlangSourceLanguage
* Replace _canCompile with something data driven.
* Fix compiling on gcc/clang for DownstreamCompiler.
* Moved some text conversions into DownstreamCompiler.
* Fix problem on non-vc builds with not having return on locateCompilers for VS.
* Change so no warning for code not reachable on locateCompilers for vs.
* WIP: CUDA code generation - currently just using CPU layout and HLSL.
* emitXXXForEntryPoint -> emitEntryPointSource
emitSourceForEntryPoint -> emitEntryPointSourceFromIR
Fix up generating cuda to get PTX.
* WIP emitting cuda for IR.
* Small improvements to CUDA ouput.
* Disable the CUDA emit test, as output not currently compilable.
|
|
* WIP getStringHash
* Have a use.
* Add slang-string-hash.h/.cpp
* Use StringSlicePool for holding strings for StringHash.
Add outputBuffer to string-literal-hash.slang so value can be tested.
Ignore the GlobalHashedStringLiterals instruction on emit.
* Add all the hashed string literals to ProgramLayout.
* Add reflection support for hashed string literals to reflection test.
* Fix string literal hash test.
* Small fixes to pass test suite.
* Fix issue in serialization where IRUse is not correctly initialized.
* Fix problem initializing IRUse for string hash pass.
Remove hack from slang-ir-specialize - specially handling if user is not null.
* * Use shared builder when replacing getStringHash
* Comments for functions in slang-ir-string-hash
* Do not allow zero length string literals. Could be allowed, but doing so would require StringSlicePool to have a special case (or some other mechanism)
|
|
* Added RiffReadHelper
* Move type to fourCC in Chunk simplifies some code.
* Make MemoryArena able to track external blocks.
Allow ownership of Data to vary.
Changed IR serialization to use moved allocations to avoid copies.
As it turns out all of the array writes could use unowned data, but doing so requires the IRData to stay in scope longer than IRSerialData, which it does at the moment - but perhaps needs better naming or a control for the feature.
* Write out slang-module container.
* WIP on -r option.
Loading modules - with -r.
* Making the serialized-module run (without using imported module).
* Split compiling module from the test.
* Separate module compilation with a function working.
* Remove serialization test as not used.
* Fix warning on gcc.
* Updated test to have types across module boundary.
* Allow entry point declaration.
A test that tries to build with just an entry point declaration and a module.
* Try to make link work with multiple modules.
* Multi module linking first pass working.
* Multi module test working with -module-name option
* Added feature to repro manifest of approximation of command line that was used.
* Use isDefinition - for determining to add decorations to entry point lowering.
* Added support for repo-file-system.h
More precise control of CacheFileSystem.
Allow RelativeFileSystem to strip paths optionally.
Use canonical paths in PathInfo cache.
Fix bug in -D options for command line output of StateSerailizeUtil
* Add missing slang-options.h
* Fix bug in bit slang-state-serialize.cpp with bit removal.
* Added documentation around -repro-file-system
Added spLoadReproAsFileSystem function.
* Fix warning.
* spAddLibraryReference
* * Add support for slang-lib extension
* Container output when using -no-codegen option
* Use the m_containerFormat to determine if the module container is constructed.
Store the result in a blob. This allows for potential access via the API.
Write the blob if a filename is set.
Use m_ prefix for container variables.
* Added spGetContainerCode.
Made spGetCompileRequestCode work.
* * Put obfuscateCode on linkage
* Remove obfuscation from variable names - as can be achieved by either stripping and/or removing NameHintDecorations at lowering
* Remove name hints being added during lowering
* Add stripping of SourceLoc location in strip phase
* Hashing of linkage import/export names.
* Do final strip in emitEntryPoint, removes any remaining SourceLoc.
|
|
* Small cleanup to how standard-library code gets marked
Declarations in the standard library used to be individualy marked with `FromStdLibModifier` so that downstream passes (notably IR lowering) could treat them specially.
At some point we simplified things by just looking for `FromStdLibModifier` in the ancestor list of a declaration, so that we could handle nested declarations without having to recursively attach the modifier to everything.
This change simplifies things a bit further in that the `FromStdLibModifier` now only get attached to the `ModuleDecl`, since that will be on the ancestor chain of all the declarations anyway.
The second change here is that we attach this marker modifier *before* we parse and do semantic checking on the module, rather than after.
There is no code in this change that relies on that difference, but changing the timing of when we attach the modifier means that the semantic checking logic can now reliably detect if something represents a declaration in the standard library.
* Change implementation of `sign()` for GLSL target
Issue #602 related to the `sign()` function having a different return type between GLSL and HLSL/SLang.
At the time, the issue was fixed by adding special-case logic in the emit pass that looked for the `sign()` function by name.
This change cleans up that logic to work using the existing `__target_intrinsic` mechanism instead.
* Make sure code emit doesn't depend on mangled names
There was existing logic in the HLSL/GLSL emit path (that got copy-pasted into the C/C++ path) where if a builtin function didn't have an application `__target_intrinsic` modifier (`IRTargetIntrinsicDecoration` at the IR level), we used some fairly fragile logic to take the mangled symbol name for a function and unmangle it to recover the original name, as well as the number of parameters (to detect member functions). This approach had always been a kludge, and we have reached a point where it actively hinders forward progress on some valuable new features.
The big idea of this change is fairly simple: in the IR lowering pass we detect when we have a stdlib function that ought to have an intrinsic definition but doesn't have a "catch-all" `__target_intrinsic` modifier applied. If we discover such a function, we go ahead and emit it to the IR with just such a catch-all modifier.
(The main alternative here would be to require that all functions in the stdlib be manually marked `__target_intrinsic` and make this a front-end issue where we get errors if we write stdlib functions wrong, but this workaround is more expedient for now)
Some of the logic around target intrinsics already handled the idea of having catch-all intrinsic modifiers/decorations with an empty target name, but that support wasn't 100% complete so this change strives to get it working.
One detail that was only being handled along this mangled-name path was support for emitting calls using member-function syntax. I updated the generic handling of `__target_intrinsic`-based calls to support specifying a member function name using a prefix `.` on the definition string.
With this work in place, it is possible to clean up a bunch of logic in the `emit` code that had to assume that any function without a body/definition must be an intrinsic. Instead, with this change we can now use the presence of a suitable `IRTargetIntrinsicDecoration` as the indicator that a function is an intrinsic. This should make the emit logic a bit more robust.
One wrinkle that remains with this change is the special handling of functions that get the name `operator[]` (that is, the `get`/`set`/etc. accessors for `__subscript` declarations). The logic no longer depends on the mangled name, but it still feels a bit gross to have this kind of string-based special case (especially since it is our main/only special case like this).
Another wrinkle is that the C/C++ back-end logic for handling intrinsic functions largely mirrors the HLSL/GLSL case, but can't just use the same exact code because it has to be intercepted by the logic that generates some of the required functions on-demand. There's still a net cleanup with this change, but there is probably an opportunity to remove even more duplication down the road.
|
|
These were meant to be merged into #1084, but I failed to push the changes to the server.
|
|
This change builds on previous work that moves toward a more IR-based representation of layout.
Those steps added some instructions for representing layout in the IR (initially just proxies for the AST layout objects), and an explicit lowering pass that could build a target-specific IR module that binds parameters and entry points to layout information.
This change aims to complete that work, in the sense that the IR representation of layout is now self-contained and does not rely on having pointers back into the AST-level representation.
Achieving this requires two main kinds of work:
1. Update any code that used layout information derived from the IR (most notably all the `slang-emit-*` code) to use the new IR representation and its accessors.
2. Update any code that *constructs* layouts using information derived from the IR to construct IR layouts instead.
The biggest new infrastructure feature in this change is support for "attributes" in the IR (I'd welcome feedback on the naming).
An attribute can either be thought of like key/value arguments that can be added to certain instructions to encode optional data, or alternatively like a decoration that is referenced as an operand instead of a child.
The value of attributes over decorations is that they can affect the hash/identity of an instruction (which decorations can't), while the advantage of decorations is that they can easily be added/removed over the lifetime of an instruction (which attributes can't).
We mostly use them here to represent operands that are logically optional.
Once attributes are available, the encoding of layout information into the IR is mostly straightforward:
* An `IRVarLayout` has a fixed operand for its type layout, and can accept a few different attributes
* Zero or more `IRVarOffsetAttr`s that specify the offset of the variable for a given resource kind. These are equivalent to the `VarLayout::ResourceInfo`s at the AST level.
* An optional `IRUserSemanticAttr` and `IRSystemValueSemanticAttr` to represent the (possibly derived) semantic of a varying input/output parameter.
* An option `IRStageAttr` to represent the known stage for a parameter.
* An `IREntryPointLayout` has a var layout for the entry point parameters (logically grouped in to a struct) and another var layout for the result parameter.
* There is a small type hierarchy rooted at `IRTypeLayout` where each subtype can add fixed operands and attributes that are expected to appear. It also supports `IRTypeSizeAttr`s that serve a similar role to the `IRVarOffsetAttr`s.
* Structure types maintain the mapping of fields to their var layouts using `IRStructFieldLayoutAttr`s.
With the encoding in place, most of the changes in category (1) (code that just *uses* rather than *creates* layouts) was straightforward. The biggest different beyond name changes was that everything needs to be fetched using accessors instead of bare fields. It would have been possible to stage this commit and make the diffs smaller by first introducing mandatory acessors to the AST layout types.
The changes in category (2) were more involved. There were a lot of places in the existing code where a `TypeLayout` or `VarLayout` would be created, and then initialized piecemeal over several lines of code (and sometimes even across functions). Because of the way that layouts need to support many optional properties, it did not seem practical to just have monolithic factory functions that took all the options as arguments, so I instead opted for a builder approach.
The builders for `IRVarLayout` and `IREntryPointLayout` are both straightforward, and honestly there is no realy need for a builder for entry point layouts right now, but I was trying to future-proof in case we decidd to add some optional attributes to them.
The builders for type layouts are more involved because of the inheritance hierarchy. Each concrete sub-type of type layout needs to define its own builder type that customizes the opcode, operands, and attributes of the final instruction.
The refactoring that had to go into this change was a nice excuse to clean up a few ugly warts in the AST layout code that were largely there to support IR use cases. While this change adds a lot of new infrastructure code to the IR, most of the client code has stayed the same or gotten simpler.
One annoying wart that remains with this change is the notion of an "offset element type layout" for parameter group types. That idea was added to deal with a legacy feature in the reflection API that we realized was a mistake, but unfortunately having that "offset" layout handy made writing a few other pieces of code simpler so that there are use cases of the feature even in the IR. Removing those uses is do-able, but requires careful refactoring so it is best left to a follow-on change.
Another thing that could be considered for a follow-on change is how much information should be specified when constructing a `Builder` for an IR type layout, and how much should be allowed to be specified statefully/piecemeal. It would be nice to force all the required operands to be specified up front, but `IRParameterGroupTypeLayout::Builder` doesn't currently work that way because so much of the client code that needs it involved a lot of stateful setting and would need to be refactored heavily to provide the necessary information up front.
|
|
* Initial work on representing layout at IR level
This change starts the process of making the back-end of the compiler independent of the AST-level layout information (`TypeLayout`, `VarLayout`, etc.) so that it instead only relies on layout information that is embedded into IR modules. This brings us incrementally closer to a world in which the back-end could be run without the AST-level structures even existing (e.g., for an application that just wants to ship IR without any AST information for IP protection, while still supporting some amount of linking and specialization).
The main parts of the change are:
* There is a bunch of incidental churn related to specifying entry points by index instead of the `EntryPoint` object for certain operations. This ends up being a better choice because we can use the index to look up side-band information about the entry point that might not be stored on the `EntryPoint` object itself. In particular...
* We expand the `ComponentType` interface to support looking up the mangled name of an entry point by index. In common cases (no generic/interface specialization) this would be the same as asking the `EntryPoint` for its mangled name, but in cases where we have specialized a generic entry point, the mangled name would include speicalization arguments that are only available on the `SpecializedComponentType` that wraps the entry point. This part of the change isn't ideal and there might be a better solution waiting to be invented. Note that we store mangled entry point names as strings rather than using `DeclRef`s because that ensures that the information could be serialized and deserialized without a dependence on the AST.
* The `TargetProgram` type (which represents binding a specific `ComponentType` for a shader program to a specific `TargetRequest` that represents the target platform) is expanded to include an `IRModule` that represents layout information, in addition to the AST-level `ProgramLayout` it already contained. We create both of these objects at the same time (on-demand) to simplify the overall flow (so that any code that triggers creation of the AST-level layout will also ensure that the IR-level layout exists).
* A bunch of code in the emit passes that was passing down layout-related objects has been eliminated. It appears that most of those objects weren't actually being used, so this is just a cleanup, but it helps ensure that the back-end steps are "clean" and don't depend on the AST-level information. The one big exception here is that the emit logic needs to know the stage for the entry point being emitted (to deal with one wrinkle in translating DXR to VKRT).
* A big change (actually introduced by @jsmall-nvidia in a branch that this change copied and then built from) is to introduce some more explicit IR instructions to represent layout information, notably an `IRTypeLayout` and an `IRVarLayout`. For now these objects still reference their AST equivalents, but the separation gives us an incremental path to move information from the AST-level objects over to the IR ones. This work includes logic in `IRBuilder` to construct the IR-level layout objects from the AST-level ones on-demand, so that the existing code paths that try to attach AST-level layout will continue to work for now.
* Because layout information is now embedded in the IR, the `slang-ir-link.cpp` logic loses a lot of cases that used to deal with attaching AST-level layout objects to IR-level instructions during the linking process. Instead, the linker now assumes that one (or more) of the input IR modules will have layout information associated with it, and the linker makes sure to copy layout decorations (and the instructions they reference) from the input IR module(s) to the output using its more ordinary mechanisms.
* Inside `slang-lower-to-ir.cpp`, we add logic to construct an IR module in a `TargetProgram` that simply references the global shader parameters, entry points, etc. and attaches IR layout decorations to them. This is akin to the existing pass in the same file that constructs IR to represent specialization information, and both of these passes share infrastructure with the main AST->IR lowering pass. Eventually, it is expected that this pass will encompass more of the logic for copying AST-level layout information over to IR-level equivalents.
* One small wrinkle with this change was that the output for an HLSL generation test case changed some of its `#line` directives. The old code was actually more inaccurate than the new, so this change just updated the baseline. It also added some logic in the linker to make sure that when an IR instruction has multiple definitions, we try to pick up a source location from any of them, in case the "main" one somehow didn't get a location.
* Another small fix was that the key/value map in `StructTypeLayout` for mapping fields/members to their layouts was keyed on `Decl*` when it really should have been `VarDeclBase*`.
This change should in principle be a pure refactoring with no functionality changes, so no new tests were added. It is unfortunately also a change that has a high probability of breaking at least *some* client code, so we may want to be defensive and mark this with a new major version number (well, a new *minor* version number since we are pre-`1.0`) to give us some room for releasing hotfixes to the old version if needed.
* fixup: infinite recursion bug detected by clang
* fixup: remove commented-out code
|
|
|
|
* Use name hint on EntryPoint naming.
* Placed the entry point name on the EntryPointDecoration.
|
|
* Split out EntryPointParamDecoration.
* Add profile to EntryPointDecoration.
* WIP for GS handling for GLSL.
* WIP for StreamOut GLSL
* Fixed GLSL geometry output.
* Clean up - remove unneeded/commented out code from the entry point change.
* Use Op nums to identify GeometryTypeDecorations (as opposed to contained enum).
* Remove setSampleRateFlag & doSampleRateInputCheck
* Remove EntryPointLayout from emit.
* Change to force CI.
|
|
* Split out EntryPointParamDecoration.
* Add profile to EntryPointDecoration.
* WIP for GS handling for GLSL.
* WIP for StreamOut GLSL
* Fixed GLSL geometry output.
* Clean up - remove unneeded/commented out code from the entry point change.
* Use Op nums to identify GeometryTypeDecorations (as opposed to contained enum).
* Remove setSampleRateFlag & doSampleRateInputCheck
|
|
* IROutputControlPointsDecoration
* IROutputTopologyDecoration
* IRPartitioningDecoration
* IRDomainDecoration
* Use IRPatchConstantDecoration alone for hlsl output.
* IRMaxVertexCountDecoration
* IRInstanceDecoration
* Removed _emitHLSLAttributeSingleString and _emitHLSLAttributeSingleInt
Removed GLSLBindingAttribute and just use NumThreadsAttribute
* Added IRNumThreadsDecoration.
* Added IRNumThreadsDecoration
* Fix build problem on x86.
Improve diagnostic text based on review.
|
|
Work on #1059
The `%` operator in the Slang implementation had several issues, and this change tries to address some of them:
* Renamed most occurences of "mod" describing this operator to be "rem" for "remainder" to better match its semantics in HLSL
* Split the operator into distinct integer and floating-point variants (`IRem` and `FRem`) to simplify having different codegen for the two
* Added floating-point variants of `operator%` and `operator%=` to the stdlib.
* Added custom C++ codegen for `kIROp_FRem` such that it maps to the standard C/C++ `remainder()` function
* Added custom GLSL codegen so that `kIROp_FRem` maps to the GLSL `mod()` function (which isn't correct...)
* Added a test case to confirm that D3D11, D3D12, and CPU targets all agree on the definition of floating-point `%`
* Fixed `render-test-tool` to allow a negative integer in a `data=...` specification. This didn't end up being used in the final test, but still seems like a good fix.
* Added a customized baseline for the Vulkan flavor of that test to confirm that we are *not* compiling correctly to SPIR-V just yet
Addressing the correctness of the output for GLSL/SPIR-V will have to come as a later change given that the operation we want is not exposed directly by unextended GLSL.
|
|
* Expanded prelude for some other resource types. Disable C++ output for ParameterGroup.
* WIP: Layout for CPU.
* Fixes to CPU layout.
* WIP: The uniform is output, but the variable definition is not.
* WIP: Entry point parameters to global scope in C++.
Handling of resource types (in so far as outputting)
* Some discussion of ABI and different input types.
* WIP: More C++ support around resource types.
* WIP: Split up variables into different structures on emit.
* WIP: Emitting C++ with wrapping up of 'Context'
* WIP: C++ code has access to semantic values.
Wrap in struct so can use method calls to pass shared state.
Disable legalizeResourceTypes and legalizeExistentialTypeLayout
* Fix structured buffer layout for CPU.
* Remove testing/handling of global uniforms on CPU path.
Typo fix.
Changed CPU tests to use new CPU calling convention.
* Check globals are working. Initalize context to zero globals.
* Order the global parameters for C++ ouput by their layout.
Note - that layout isn't quite working correctly because the StructuredBuffer<int> the int seems to be consuming uniform space.
* Work around for reflection not having all data needed for layout ordering for C++ code.
* Output constant buffers as pointers.
* Entry point parameters accessed through pointer to struct.
* WIP: Layout for CPU is reasonable for test case.
* Only output 'f' after float literal if type marks as a float.
* Cast construction works on C++.
* Made IntrinsicOp::ConvertConstruct to make intent clearer.
* C++ handling construction from scalar.
Handle access of a scalar with .x.
Check default initialization.
* Comment about need for split of kIROp_construct.
Release build works.
* Added support from constructVectorFromScalar to C/C++ target.
* Handling of in/out in C/C++.
* First pass documentation CPU support.
* Improvements to C++/C slang code generation documentation.
* Small doc change to include need for mechansim to specify cpp compiler path.
* Better handling of swizzling - allow swizzling a scalar into a vector.
|
|
* Revise new COM-lite API
This change revises the "COM-lite" API that was recently introduced to try to streamline it and introduce some missing central/base concepts.
The central new abstraction in the API is the notion of a "component type," which is a unit of shader code composition. A component type can have:
* IR code for some number of functions/types/etc.
* Zero or more global shader parameters
* Zero or more "entry point" functions at which execution can start
* Zero or more "specialization" parameters (types or values that must be filled in before kernel code can be generated)
* Zero or more "requirements" (dependencies on other component types that must be satisfied before kernel code can be generated)
Both individual compiled modules, and validated entry points are then examples of component types, and we additionally define a few services that apply to all component types:
* We can take N component types and compose them to create a new component type that combines their code, shader parameters, entry points, and specialization parameters. A composed component type may also include requirements from the sub-component types, but it is also possible that by composing thing we satisfy requirements (if `A` requires `B`, and we compose `A` and `B`, then the requirement is now satisfied, and doesn't appear on the composite).
* We can take a component type with N specialization parameters, and specialize it by giving N compatible specialization arguments. The result of specialization is a new component type with zero specialization parameters. Under the right circumstances the specialzed component type will be layout compatible with the unspecialized one.
* One more example that isn't exposed in the public API today is that we can take a component with requirements and "complete" it by automatically composing it with component types that satisfy those requirements. This can be seen as a kind of linking step that pulls together the transitive closure of dependencies.
* We can query the layout for the shader parameters and entry points of a component type, for a specific target.
* We can query compiled kernel code for an entry point in a component type (for a specific target). This only works for component types with zero specialization parameters and zero requirements.
The idea is that by giving users a fairly general algebra of operations on component types, they can compose final programs in ways that meet their requirements. For example, it becomes possible to incrementally "grow" a component type to represent the global root signature for ray tracing shaders as new entry points are added, in such a way that it always stays layout-compatible with kernels that have already been compiled.
Much of the implementation work here is in implementing the unifying component type abstraction, and in particular re-writing code that used to assume a program consisted of a flat list of modules and entry points to work with a hierarchical representation that reflects the underlying algebra (e.g., with types to represent composite and specialized component types).
There's also a hidden "legacy" case of a component type to deal with some legacy compiler behaviors that can't be directly modeled on top of the simple algebra with modules and entry points.
This API is by no means feature-complete or fully developed. It is expected that we will flesh it out more when bringing up application code (e.g., Falcor) on top of the revamped API.
One notable thing that went away in this change is explicit support for "entry point groups" and notions of local root signatures (especially the Falcor-specific handling of the `shared` keyword, which a previous change turned into an explicitly supported feature). With the new "building blocks" approach, it should be possible for a DXR application to deal with local root signatures as a matter of policy (on top of the API we provide). If/when we need to provide some kind of emulation of local root signatures for Vulkan (and/or if Vulkan is extended with an explicit notion of local root signatures), we might need to revisit this choice.
* Fix debug build
There was invalid code inside an `assert()`, so the release build didn't catch it.
* fixup: warnings
* fixup: more warnings-as-errors
* fixup: review notes
* fixup: use component type visitors in place of dynamic casting
|
|
This appears to be a regression introduced in #1001, and missed because none of our existing tests covered `static const` arrays on the GLSL/SPIR-V targets.
The basic problem is that we cannot output a `static const` definition in GLSL because `static` is a reserved word and not a keyword. Instead for GLSL we just want a `const` array. This change makes the emission of `static` for global-scope constants key on the target language for code generation, and only emit it for HLSL, C, and C++.
This change also adds a test case specifically for running Slang input that has a `static const` array on the Vulkan target.
|
|
This change adds back a little bit of explicit support for global constants in the IR, after a previous change completely removed the existing `IRGlobalConstant` node type.
The new `IRGlobalConstant` is *not* a parent instruction, and doesn't function at all like the old one.
Instead it is effectively a simple instruction that takes zero or one operands:
* The zero-operand case represents a constant with unknown value. This would usually come from another module, and thus would have an `[import(...)]` linkage decoration, so that after linking it resolves to a constant with a known value.
* In the one-operand case, the single operand represents the value of the constant, so that the operation semantically behaves like an identity function. It exists just to give decorations something to "attach" to, so that a global constant with a value can have, e.g., an `[export(...)]` decoration to establish linkage.
The IR lowering pass was updated to create the new node type to wrap any global constants. For now we do this both for global `static const` variables and function-scope `static const`, although the latter doesn't really need the extra indirection.
The IR linking logic was extended to handle linking of global constants akin to how other global instructions are handled. The new logic is mostly boilerplate, and it is likely that a refactor of the linking logic would eliminate the need for this kind of per-instruction-opcode handling of IR instructions that can have linkage.
A custom pass was added that is intended to be run right after linking (it could arguably be folded into `linkIR()`, but I thought it was safer to keep each pass as small as possible). This pass replaces any `IRGlobalConstant` that has a value (operand) with that value, so that global constants should be eliminated after the linking step. This ensures that downstream optimization/transformation passes don't have to deal with the possibility of global constants.
Almost all the existing passes would Just Work if global constants were left in the IR. The two big exceptions are:
* Anything that relies on testing `IRInst*` identity as a way to test for things having the same value would break, since a global constant is a distinct `IRInst*` from its value.
* The type legalization pass doesn't handle `IRGlobalConstant` instructions with non-simple types. This could be added if we ever wanted it, but it seemed silly to write this code now if it would always be dead (and thus untested).
I went ahead and updated the emit logic to handle an `IRGlobalConstant`s that still existing in the IR module at emit time, since the amount of code required was small so that being robust to that case seemed safest (e.g., in case we ever want to have a path that emits code directly while skipping some/all of our IR transformation passes).
There should be no visible changes to the functionality of the compiler with this change, but it should help make IR dumps from the front-end more clear/explicit (since each constant will be a distinct instruction with its own name), and paves the way for supporting proper cross-module linkage of constants.
|
|
Before this change, global and function-scope `static const` declarations were represented as instructions of type `IRGlobalConstant`, which was represented similarly to an `IRGlobalVar`: with a "body" block of instructions that compute/return the initial value.
This representation inhibited optimizations (because a reference to a global constant would not in general be replaced with a reference to its value), and also caused problems for resource type legalization because the logic for type legalization did not (and still does not) handle initializers on globals (so global *variables* that contain resource types are still unsupported).
The change here is simple at the high level: we get rid of `IRGlobalConstant` and instead handle global-scope constants as "ordinary" instructions at the global scope. E.g., if we have a declaration like:
static const int a[] = { ... }
that will be represented in the IR as a `makeArray` instruction at the global scope, referencing other global-scope instructions that represent the values in the array.
This simple choice addresses both of the main limitations. A `static const` variable of integer/float/whatever type is now represented as just a reference to the given IR value and thus enables all the same optimizations. When a `static const` variable uses a type with resources, the existing legalization logic (which can handle most of the "ordinary" instructions already) applies.
Another secondary benefit of this approach is that the hacky `IREmitMode` enumeration is no longer needed to help us special-case source code emit for `static const` variables.
Beyond just removing `IRGlobalConstant`, and updating the lowering logic to use the initializer direclty, the main change here is to the emit logic to make it properly handle "ordinary" instructions that might appear at global scope.
One open issue with this change, that could be addressed in a follow-up change, is that "extern" global constants that need to be imported from another module (but which might not have a known value when the current module is compiled) aren't supported - we don't have a way to put a linkage decoration on them. A future change might re-introduce global constants as a distinct IR instruction type that just references the value as an operand (if it is available). We would then need to replace references to an IR constant with references to its value right after linking.
|
|
* WIP: Emitting Cpp
* Added HLSLType instead of using IRInst - because they don't seem to be deduped.
* Removed need for lexer to take a String.
Added mechansim to lookup intrinsic functions on C++.
* A c/c++ cross compilation test.
* WIP Cpp output using cloning and slang types.
* More work to generate mul funcs.
* WIP: Outputting some simple C++.
* Expose findOrEmitHoistableInst to IRBuilder to aid cloning,
* Simplification for checking for BasicTypes.
Test infrastructure compiles output C++ code.
* Dot and mat/vec multiplication output.
* First pass at swizzling.
* First support for binary ops.
* Builtin binary and unary functions.
* Any and all.
* WIP adding support for other functions.
Added code to generate function signature.
* Add scalar functions to slang-cpp-prelude.h
* Support for most built in operations.
* Tested first ternary.
* Checking the emitting of corner cases functions - normalize, length, any, all, normalize, reflect.
* Check asfloat etc work.
* Fmod support.
* WIP Array handling in C++.
* First stage in being able to handl arbitrary type output for CLikeSourceEmitter
* Removed Handler/Emitter split - so can implement more easily complex type naming.
* Array passing by value first pass.
* Rename Array -> FixedArray
* Outputs structs in C++.
* Emit the thread config.
* Dimension -> TypeDimension
* SpecializedOperation -> SpecializedIntrinsic
Operation -> IntrinsicOp
Use shared impl of isNominalOp
Commented use of m_uniqueModule etc.
* Add code to test slang->cpp when compiled doesn't have errors. Does so by building shared library and exporting the entry point.
* Fix linux clang/gcc compile error about override not being specified.
* Make sure c-cross-compile is run on linux targets/smoke.
* Remove c-cross-compile.slang from smoke.
* Fix running tests/cross-compile/c-cross-compile.slang on Ubuntu 16.04
* Only add -std=c++11 for C++ source.
|
