| Age | Commit message (Collapse) | Author |
|---|
|
The main change I was working on here was to start having more of the builtin functions (in this case, `cos`, `sin`, and `saturate`) just lower to the IR as calls to builtin functions (with declarations but no definition), rather than expect/require them to map to individual IR opcodes in every case.
The main change there was the removal of some `intrinsic_op` modifiers in the stdlib. This then requires the `isTargetInstrinsic` logic in IR-based code emit to avoid emitting declarations for these intrinsics.
The corresponding logic for emitting *calls* to these intrinsics is currently being skipped.
Along the way, a variety of fixups were added:
- In order to support lowering to GLSL, we need to handle cases where a variable/function name uses a GLSL reserved word. The right long-term fix there is to always use generated or mangled names, but for now I'm hacking it by adding a `_s` prefix to all names during IR-based emit.
- This needs a flag to disable it, since some of our tests currently rely on checking binding information from generated HLSL/SPIR-V that will include these mangled/modified names.
- Emit matrix layout modifiers appropriately for GLSL
- Specialize IR parameter-block emission between GLSL and HLSL
- Fix up argument count/index logic for a couple of opcodes that weren't fixed when removing the types from the explicit operand list
- Fix up IR generation for calls to declarations with generic arguments. We were briefly adding the generic args to the ordinary argument list, which added complexity in several places. We now rely on the declaration-reference nodes in the IR to carry that extra info.
- TODO: We actually need to make sure that this is the case, since we don't currently correctly generated specialized decl-refs when building IR for function calls
The main test that would have been affected by this is `cross-compile-entry-point`, but I was not able to get that working fully with the IR. The main problem in this case was that when emitting GLSL we will need to perform certain required transformations on the IR to get legal code for GLSL. Notably:
- We need to hoist entry-point parameters away from being function parameters, and make them be global variables. This is currently being hand-waved during the emit logic, but it seems way better to have it all get cleaned up in the IR first.
- We need to scalarize entry-point parameters, because structure input/output is not supported as vertex input or fragment output (and it may be best to always scalarize anyway, to match HLSL semantics). (Note: "scalarize" here means to bust up structures, but not matrices/vectors)
|
|
* First attempt at a Linux build
- Fix up places where C++ idioms were written assuming lenient behavior of Microsoft's compiler
- Add a few more alternatives for platform-specific behavior where Windows was the only platform accounted for.
- Add a basic Makefile that can at least invoke our build, even if it isn't going good dependency tracking, etc.
- Build `libslang.so` and `slangc` that depends on it, using a relative `RPATH` to make the binary portable (I hope)
- Add an initial `.travis.yml` to see if we can trigger their build process.
* Fixup: const bug in `List::Sort`
I'm not clear why this gets picked up by the gcc *and* clang that Travis uses, but not the (newer) gcc I'm using on Ubuntu here, but I'm hoping it is just some missing `const` qualifiers.
* Fixup: reorder specialization of "class info"
Clang complains about things being specialized after being instantiated (implicilty), and I hope it is just the fact that I generate the class info for the roots of the hierarchy after the other cases. We'll see.
* Fixup: add `platform.cpp` to unified/lumped build
* Fixup: Windows uses `FreeLibrary`
and not `UnloadLibrary`
* Fixup: fix Windows project file to include new source file
This obviously points to the fact that we are going to need to be generating these files sooner or later.
|
|
None of these changes are made "live" at the moment. I'm just trying to get them checked in to avoid divering too far from `master` at any point during development.
- Add basic emit logic to produce GLSL from the IR in a few cases (the existing IR emit logic was ad hoc and HLSL-specific)
- When lowering a function declaration, walk up its chain of parent declarations to collect additional parameters as needed
- When lowering a call, make sure to add generic arguments that come from the declaration reference being called
- Attach a "mangled name" to symbols when lowering, so that we can eventually use that name to resolve things for linkage.
- After the above work, I had to apply some fixups to make sure that generic arguments *don't* get added when the user is calling an `__intrinsic_op` function, since those should map 1-to-1 down to instructions with just their ordinary parameter list.
A big open question right now is whether I should continue to represent the generic arguments as just part of the ordinary argument list for a function, or split them out into separate `applyGeneric` and `apply` steps.
A strongly related question is whether a declaration with generic parameters should lower into a single declaration, or one declaration nested inside an outer generic declaration.
A good future step at this point would be to eliminate a lot of the `__intrinsic_op` stuff in favor of having the builtin functions include their own definitions, which might be in terms of a new expression-level construct for writing inline IR operations. This can't be done until the existing AST-to-AST path is no longer needed for cross-compilation purposes.
More immediate next steps here:
- We need a way to round-trip calls to external declaration that get handled by this mangled-name logic. Basically, if we are asked to output HLSL and we see a call to `_S...GetDimensions...(float4, t, a, ...)` we need to be able to walk the mangled name and get back to `t.getDimensions(a, ...)` without a whole lot of manual definitions to make things round-trip.
- In the other case, where a declaration isn't built-in for the chosen target, we need to be able to load a module of target-specific definitions (which will somehow map back to symbols with certain mangled names) and then look these up (by mangled name) and then load/link/inline them into the user's IR to satisfy requirements in their code.
|
|
At a high level, this commit adds two things:
1. A "bytecode" format for serializing Slang IR instructions and related structure (functions, "registers")
2. A virtual machine that can load and then execute code in that bytecode format.
The reason for kicking off this work right now is that we *need* a way to run tests on Slang code generation that doesn't rely on having a GPU present (given that our CI runs on VM instances without GPUs), nor on textual comparison to the output of other compilers. With these features I've implemented a slapdash `slang-eval-test` test fixture that can run a (trivial) compute shader to very our compilation flow through to bytecode.
Some key design constraints/challenges:
- The bytecode format should be "position independent" so that a user can just load a blob of data and then inspect it without having to deserialize into another format, allocate memory, etc. Eventually the bytecode format might be a replacement for out current reflection API (we used to base reflection off a similar format, but the cost/benefit wasn't there at the time and we switched to just using the AST).
- The VM should be able to execute bytecode functions without doing any per-operation translation, JIT, etc. (translation of more coarse-grained symbols is okay). For now the VM is just being used to run tests, but eventually I'd like it to be viable for:
- Running Slang-based code in the context of the compiler itself. This starts with stuff like constant-folding in the front-end, but could expand to more general metaprogramming features.
- Running Slang-based ocde within a runtime application (e.g., a game engine) that wants to be able to run things like "parameter shader" code, or even just evaluate compute-like code on CPU (e.g., when supporting particles on both CPU and GPU).
- Finally, the bytecode format should ideally be able to round-trip back to the IR without unacceptable loss of information. This requirement and the previous one play off of each other, because things like a traditional SSA phi operation is ugly when you have to actually *execute* it. This doesn't matter right now when we don't have SSA yet, but it might be part of the decision-making here.
The actual implementation is centralized in `bytecode.{h,cpp}` and `vm.{h.cpp}`.
Big picture notes:
- The space of opcodes is shared between IR and bytecode (BC), with the hope that this makes translation of operations between the two easy.
- The actual bytecode instruction stream relies on a variable-length encoding for integer values, including opcodes and operand numbers, so that the common case is single-byte encoding.
- In the long term I intend to have a rule that if you use a single-byte encoding for an opcode, then all operands are required to use single-byte encodings too. Operations that need multi-byte operands would then be forced to use a multi-byte encoding of the op, and would be sent down a slower path in the interpeter.
- The "bytecode"'s outer structure is based on ordinary data structures linked with pointers, but they are "relative pointers" so the actual structure is position-independent.
- There are two main kinds of operands: registers and "constants." An operand is a signed integer where non-negatie values indicate registers (with `index == operandVal`) and negative values indicate constants (with `index == ~operandVal`).
- Registers are stored in the "stack frame" for a VM function call, and each has a fixed offset based on the size of the type and those that come before it. Conceptually, registers are allowed to overlap if they aren't live at the same time, and we manage this with a simple stack model: every register is supposed to identify the register that comes directly before it (this isn't implemented yet).
- "Constants" are more realistically a representation of "captured" values, but they are currently also how constants come in. Basically we can use a compact range of indices in the bytecode for a function, and each of these indices indirectly refers to some value in the next outer scope.
- The actual encoding of bytecode instructions right now is largely ad-hoc and very wasteful (we encode the type on everything, and we also encode everything as if it had varargs).
- In some cases, an instruction needs to know the types of the values involved (e.g., because it needs to load an array element, which means copying a number of bytes based on the size). The way the VM works we have types attached to our registers, so we currently get sneaky and look at those types in some ops. Longer term is makes sense to encode the required type info directly in the BC.
- There's a whole lot of hand-waving going on with how the actual top-level bytecode module gets loaded, because of the way we currently treat the top-level module as an instruction stream in the IR. This means that we try to represent the loaded module as a "stack frame" for a call to the module as a function, but that approach as serious problems, and isn't realistically what we want to do.
|
|
* IR: handle control flow constructs
This change includes a bunch of fixes and additions to the IR path:
- `slang-ir-assembly` is now a valid output target (so we can use it for testing)
- This uses what used to be the IR "dumping" logic, revamped to support much prettier output.
- A future change will need to add back support for less prettified output to use when actually debugging
- IR generation for `for` loops and `if` statements is supported
- HLSL output from the above control flow constructs is implemented
- Revamped the handling of l-values, and in particular work on compound ops like `+=`
- Add basic IR support for `groupshared` variables
- Add basic IR support for storing compute thread-group size
- Output semantics on entry point parameters
- This uses the AST structures to find semantics, so its still needs work
- Pass through loop unroll flags
- This is required to match `fxc` output, at least until we implement
unrolling ourselves.
* Fixup: 64-bit build issues.
* fixup for merge
|
|
This gets us far enough that we can convert a single test case to use the IR, under the new `-use-ir` flag.
Getting this merged into mainline will at least ensure that we keep the IR path working in a minimal fashion, even when we have to add functionality the existing AST-based path
There is definitely some clutter here from keeping both IR-based and AST-based translation around, but I don't want to have a long-lived branch for the IR that gets further and further away from the `master` branch that is actually getting used and tested.
Summary of changes:
- Add pointer types and basic `load` operation to be able to handle variable declarations
- Add basic `call` instruction type
- Add simple address math for field reference in l-value
- Always add IR for referenced decls to global scope
- Add notion of "intrinsic" type modifier, which maps a type declaration directly to an IR opcode (plus optional literal operands to handle things like texture/sampler flavor)
- Improve printing of IR instructions, types, operands
- Add constant-buffer type to IR
- Allow any instruction to be detected as "should be folded into use sites" and use this to tag things of constant-buffer type
- Also add logic for implicit base on member expressions, to handle references to `cbuffer` members
- Add connection back to original decl to IR variables (including global shader parameters...)
- Use reflection name instead of true name when emitting HLSL from IR (so that we can match HLSL output)
- Make IR include decorations for type layout
- Re-use existing emit logic for HLSL semantics to output `register` semantics for IR-based code
- Make IR-based codegen be an option we can enable from the command line
- It still isn't on by default (it can barely manage a trivial shader), but it seems better to enable it always instead of putting it under an `#ifdef`
- Fix up how we check for intrinsic operations suring AST-based cross compilation so that adding new intrinsic ops for the IR won't break codegen.
|
|
The `-split-mixed-types` flag can be provided to command-line `slangc`, and the `SLANG_COMPILE_FLAG_SPLIT_MIXED_TYPE` flag can be passed to `spSetCompileFlags`.
Either of these turns on a mode where Slang will split types that included both resource and non-resource fields.
The declaration of such a type will just drop the resource fields, while a variable declare using such a type turns into multiple declararations: one for the non-resource fields, and then one for each resource field (recursively).
This behavior was already implemented for GLSL support, and this change just adds a flag so that the user can turn it on unconditionally.
Caveats:
- This does not apply in "full rewriter" mode, which is what happens if the user doesn't use any `import`s. I could try to fix that, but it seems like in that mode people are asking to bypass as much of the compiler as possible.
- When it *does* apply, it applies to user code as well as library/Slang code. So this will potentially rewrite the user's own HLSL in ways they wouldn't expect. I don't see a great way around it, though.
|
|
With this change, basic generation of IR works for a trivial shader, and there is some basic support for dumping the generated IR in an assembly-like format.
As with the other IR change, the use of the IR is statically disabled for now, so that existing users won't be affected.
|
|
Fixes #23
Up to this point, the compiler has used the ordinary `String` type to represent declaration names, which means a bunch of lookup structures throughout the compiler were string-to-whatever maps, which can reduce efficiency.
It also means that things like the `Token` type end up carying a `String` by value and paying for things like reference-counting.
This change adds a `Name` type that is used to represent names of variables, types, macros, etc.
Names are cached and unique'd globally for a session, and the string-to-name mapping gets done during lexing.
From that point on, most mapping is from pointers, which should make all the various table lookups faster.
More importantly (possibly), this brings us one step closer to being able to pool-allocate the AST nodes.
|
|
Fixes #24
So far the code has used a representation for source locations that is heavy-weight, but typical of research or hobby compilers: a `struct` type containing a line number and a (heap-allocated) string.
This is actually very convenient for debugging, but it means that any data structure that might contain a source location needs careful memory management (because of those strings) and has a tendency to bloat.
The new represnetation is that a source location is just a pointer-sized integer.
In the simplest mental model, you can think of this as just counting every byte of source text that is passed in, and using those to name locations.
Finding the path and line number that corresponds to a location involves a lookup step, but we can arrange to store all the files in an array sorted by their start locations, and do a binary search.
Finding line numbers inside a file is similarly fast (one you pay a one-time cost to build an array of starting offsets for lines).
More advanced compilers like clang actually go further and create a unique range of source locations to represent a file each time it gets included, so that they can track the include stack and reproduce it in diagnostic messages.
I'm not doing anything that clever here.
|
|
- API users can use this to get "clean" output to aid with debugging Slang issues
- Also changes the prefix on intermediate files that Slang dumps, to make them easier to ignore with a regexp
|
|
- Change the `slang` project from a static library to a dynamic one
- Add some details around `slang.h` to make sure DLL export stuff is working
- Make the `slangc` executable use the dynamic library
- Rename the `glslang` sub-project to `slang-glslang` and move it into the main source hierarchy
- This reflects the fact that it isn't a stand-alone tool, and isn't in any way a standard binary of glslang, but rather just an artifact of how Slang uses glslang
|
|
Fixes #12
- This was a latent issue, but the previous commit brought it to the front.
- As indicated in #12, I don't allocate a descriptor-table slot to the block
- Instead I allocate a `PushConstantBuffer`
- Unlike what #12 asks for, I don't use a different resource type for the contents of the block
- Pretty much all the logic is easiest if these continue to be just plain `Uniform` data
|
|
Calling:
spSetDumpIntermedites(compileRequest, true);
will set up a mode where Slang tries to dump every intermediate HLSL, GLSL, DXBC, SPIR-V, etc. file it generates. If SPIR-V or DXBC is requested then we also dump assembly of those.
Right now the files are all named as `slang-<counter>.<ext>`, and get dropped in whatever the working directory is, but I'm open to ideas on how to improve that.
Note: this change introduces a new binary interface to `glslang`, so pulling it requires an updated `glslang.dll`.
|
|
Fixes #83
- The basic idea is that I added a bunch of more specific profile names line `glsl_vertex_430` which indicate the desired GLSL version the user wants.
- An explicit `#version` line in the code always overrides one specified by profile, though
|
|
- This really just checks two basic things:
1. Was there any global variable declared with `in` and `sample`?
2. Did any code encountered during lowering referenece `gl_SampleIndex`?
- This doesn't cover what HLSL could need, nor what we would need for cross-compilation. Consider it GLSL-specific for now.
- In order to generate the information with even a reasonable chance of being accurate (not giving a ton of false positives) I tried to integrate the checks into the lowering process (so they only see code that is referenced, one hopes).
- For this to work with my testing setup, I needed to make sure that lowering is always performed, prior to emitting reflection info
- This change broke several reflection tests, because they had been using code that wouldn't actually pass the downstream compiler. I checked in fixes for those.
|
|
- This also adds reflection API for querying:
- Entry point name
- Entry point parameter list
|
|
EntryPointResult/TranslationUnitResult, added helper functionality; Ensure null termination when printing raw data
|
|
|
|
- Allow a code-generation target of `NONE` in order to suppress ordinary output in test cases where we don't care about the actual output (just pass/fail result)
- Add explicit `location` layout qualifiers to intermediate vertex-to-fragment variables in GLSL test cases for rendering, to work around apparent Intel driver bugs.
|
|
These are mostly copy-pasted from the existing `cbuffer` support, so there might be details I'm missing.
|
|
`SPECIALIZTION` -> `SPECIALIZATION`
|
|
- The big change here is the introduction of a "lowering" pass that takes an input AST from the semantic checker, and produces an output AST suitable for emitting. The intention is that he lowering pass is responsible for:
- Stripping out unused code (when we have enough information to do so), by only outputting declarations that are transitively references from an entry point
- When cross-compiling to GLSL, generating a suitable `void main()` entry point to wrap the user-written entry-point function
- (Eventually) legalizing types in the program, by scalarizing aggregate types that mix uniform and resource types
- (Eventually) instantiating generic declarations so that the resulting code only deals with fully specialized declarations
- (Eventually) de-sugaring OOP constructs into basic "structs and functions" form
- (Eventually) instantiating code that depends on interface types at the concrete types chosen
- It is clear that there is still a lot of work to be done there, to this change is really about getting infrastructure in place without breaking the existing test cases.
- One cleanup here is that we get rid of the idea of whole-translation-unit output, since that was specific to HLSL output, and there is really no strong reason for keeping it. Users should now just ask for the output for each entry point that they wanted to generate.
- The biggest source of complexity for the lowering process is that it needs to produce the same AST structure as the input, to deal with the complexity of the rewriter case. That is, we need the output to be able to reproduce the input exactly in the case where we are rewriting and nothing needs to change, so the output format needs at least the degrees of freedom of the input.
- As a result, we end up having to distinguish "rewriter" and "full" modes in both lowering and code-emit steps, so that we can react appropriately.
- Generating a GLSL `main()` also adds a lot of complexity. Right now I'm using the simplest approach, where we always output the Slang/HLSL entry point as an ordinary function (as written) and then emit a simple GLSL `main()` to call it. I generate globals for all the shader inputs/outputs (these need to be scalarized and have explicit `location`s attached), and then collect these into the `struct` types of the original parameters as needed.
- This approach will start to have some major down-sides once we have to deal with "arrayed" input/output
- A long-term question here is how to replace entry-point parameter types with scalarized and/or "transposed" versions, while still letting the original code work as written (including copying those inputs to temporary arrays)
- Split `BlockStatementSyntaxNode` into:
- `BlockStmt` which just provides a scope around a `body` statement
- `SeqStmt` which just allows multiple statements to be treated as one
- Change how we emit `for` loops, to deal with the case where the initialization part might expand into multiple statements
- Basically `for(A;B;C) {D}` becomes `{A; for(;B;C) {D}}`, so we can handle arbitrary statements for `A`
- As an additional wrinkle, when we are rewriting HLSL, we just generate `A; for(;B;C) {D}` to deal with the broken scoping there
- This change is needed because the lowering pass was sometimes expanding the original initialization statement `A` into a block `{A}`. Certainly if it declared multiple variables we'd need to handle it, and this seemed the easiest way
- A more significant challenge for lowering would come if/when we ever wanted to support true short-circuiting behavior for `&&` and `||`
- For right now I'm not changing the behavior of the "rewriter" mode, so we still have `UnparsedStmt` instances being generated, but it is clear that eventually we need to parse *all* input, even if we can't type-check 100% of it. This is required so that we can rewrite user code that might refer to a shader input with interface type.
|
|
Right now `#import` only differs from `#include` in that it takes a string literal for a file name instead of a raw identifier (to which `.slang` gets appended).
The next step is to make `#import` respect preprocessor state, while `__import` doesn't.
|
|
The main user-visible change here is that instead of `spAddTranslationUnitEntryPoint` we have `spAddEntryPoint`, to reflect that the list of entry points is "global" to a compile request.
As a result, `spGetEntryPointSource` now only needs the entry point index, and not the translation unit index.
There are a bunch more behind-the-scenes changes, though, reflecting a streamlining of the concepts related to compilation into a smaller number of classes.
Now there is:
- `Session` (unchanged) to manage the lifetimes of shared stuff like the stdlib
- `CompileRequest` (merges in `CompileOptions`) to handle all the lifetime related to a single invocation of the compiler
- `TranslationUnitRequest` (merges `TranslationUnitOptions`, `CompileUnit`) to represent a single translation unit ("module") that the user is trying to compile. This is a single file for HLSL/GLSL, but can be multiple files for Slang.
- `EntryPointRequest` (merges `EntryPointOption` and a bit of `EntryPointResult`) to track a single entry point that the user is asking to compile (that entry point always comes from a single translation unit)
A lot of functions used to take some combination of these and end up with really long signatures.
I've given most of the objects "parent" pointers so that they can get back to all the context they need, so most functions don't need as many parameters.
It may eventually be important to tease these apart again, in particular:
- The code-generation side of things (the `*Result` types) might need to be pulled out in case we want to codegen multiple times from the same AST
- Similarly, the layout stuff may also need to be pulled out, in case we want to lay things out multiple times with different rules.
|
|
The basic idea of this change is that user code can just write:
#include "foo.h"
and then if `foo.h` gets found in a list of registered directories for "auto-import," then it actually gets interpreted as if the user had writte, more or less:
__import foo;
That is, the code in `foo.h` will be treated as Slang, and will be fully parsed and checked (no matter what the source language had been), and the scoping rules will be those of `__import` instead of `#include`.
This is a really big hammer, and I could imagine it smashing fingers if used poorly.
I'm not sure this feature will pan out, but we need to try things to know.
One big piece of that that I'll likely keep in either case is an overhaul of command-line options parsing for `slangc`. In particular, this logic has been moved into the core `slang` library (so that users can just pass options in via the API), and it is all done on UTF-8 strings rather than wide strings (which was always going to be Windows-specific).
|
|
It is always easier to add back code when you need it, than it is to maintain code you aren't using.
|
|
This is a large change that contains many pieces:
- Update the `cross-compile0` test to actually make use of cross compilation.
Now the `cross-compile0.hlsl` file contains both HLSL and GLSL source code, and then imports code from `cross-compile0.slang`, which provides a "library" (one function) that can be shared between both the HLSL and GLSL version of things.
- Fixed a bug in the support for backslash-escaped newlines.
- Added a new `__import` declaration type (replaces the `using` directive that was still around in a vestigial form)
An `__import` causes the compiler to look for a Slang source file (currently using the ordinary `#include` lookup logic), and then parse/check the found file as an additional module ("translation unit"), before making its declarations visible in the current scope.
- Refactored the main compilation flow to be simpler. There were the `ShaderCompiler` and `ShaderCompilerImpl` classes that weren't relaly doing anything, but added complexity to the whole workflow.
- The `render-test` application has been heavily modified to better support testing cross-compilation workflows. At the most basic level we are starting to distinguish pass-through vs. rewriter workflows, and are passing various `#define`s down to the compiler(s) to let the source code be customized as needed for each case.
Several annoying corner cases are caused here by having to support the GLSL compilation model, which really wants each entry point in its own specific translation unit, whereas we really want to keep things nicely contained in single files.
- Added support for `__intrinsic` operations to have target-specific behavior.
This allows a function to be given a different name for some specific target (so a call gets emitted as a call to that other operation).
More generally, the library writer can put together an arbitrary format string that will be used in place of expressions that call the given function, e.g.:
__intrinsic(hlsl, "$1 - $0") __intrinsic int foo(int a, int b);
Given this declaration, a call like `foo(x,y)` will code generate as `x - y` for HLSL, and as `foo(x,y)` for all other targets.
Annoying things still to be dealt with:
- The way that I'm filtering the user-provided options when passing things down to the compilation of dynamically loaded modules is a bit ad hoc. It would be good to have a systematic notion of which options will be inherited and which won't. There is also more code duplication than I'd like, so we risk having the compiler behave differently when compiling a file at the top level, vs. because of `__import`.
- Adding target-specific behavior to intrinsics is all well and good, but the current approach means we can only add this to the original declaration, which limits the ability to easily extend the set of targets.
A better approach long-term would be to add a more robust notion of target-based overload resolution (which would happen after semantic checking). Then one mechanism would be used to find the right target-specific overload to use for an operation, and then each (target-specific) definition could use a simpler attribute to intercept code-generation behavior.
Note that we might eventually need a similar notion to deal with stage- or profile-specific functions and the overloading behavior around them, so using this for intrinsics doesn't seem like a bad idea.
|
|
|
