[pull] master from vosen:master#8
Open
pull[bot] wants to merge 398 commits into
Open
Conversation
Replaces traditional LALRPOP-based parser with winnow-based parser to handle out-of-order instruction modifer. Generate instruction type and instruction visitor from a macro instead of writing by hand. Add separate compilation path using the new parser that only works in tests for now
This is very incomplete. Just enough code to emit LLVM bitcode and continue further development
The overarching goal is to refactor all passes so they are module-scoped and not function-scoped. Additionally, make improvements to the most egregiously buggy/unfit passes (so the code is ready for the next major features: linking, ftz handling) and continue adding more code to the LLVM backend
Additionally remove unused compilation paths
It's weird that it fails without `-mno-link-builtin-bitcode-postopt`. I've tested it only on ROCm 6.2, might be broken on older or newer ROCm
* Refactor emit_intrinsic to allow struct return type
* Implement extended precision integer addition
Uses `llvm.*add.with.overflow.*`. That intrinsic does not take a carry argument, so handling carry in requires multiple additions and combining the carry out, but the AMDGPU target is able to translate that pattern into a single instruction.
These four PTX instructions:
```
add.cc.u32 r0, a0, b0;
addc.cc.u32 r1, a1, b1;
addc.cc.u32 r2, a2, b2;
addc.u32 r3, a3, b3;
```
are translated into four RDNA3 instructions:
```
v_add_co_u32 v0, vcc_lo, v0, v4
v_add_co_ci_u32_e32 v1, vcc_lo, v1, v5, vcc_lo
v_add_co_ci_u32_e32 v2, vcc_lo, v2, v6, vcc_lo
v_add_co_ci_u32_e32 v3, vcc_lo, v7, v3, vcc_lo
```
* cargo fmt
* Rename to match convention
* cargo fmt
…ws library loading (#612)
Adds `sub.cc`, `subc`, `subc.cc`, `mad.cc`, `madc`, and `madc.cc` on top of the work done in #607
Support cuMemAllocPitch_v2 and cuMemcpy2D_v2 and some other memory-related functions
Same cleanup as #610 did for emit_clz: replaces the hand-rolled LLVMGetNamedFunction/LLVMAddFunction/LLVMBuildCall2 sequence with a call to the existing emit_intrinsic helper. Output IR is unchanged, the brev_llvm test fixture already expects the standard @llvm.bitreverse.iN intrinsic call form that emit_intrinsic produces.
Rename the parameter to match the function name and align with the sibling functions emit_vector_write and emit_vector_repack which name their parameters after their argument types/intent. No behavior change.
The `?` operator already converts the error type via the `From` trait, and `CompilerError` has `#[from]` impls (or explicit From impls) for `io::Error`, `Utf8Error`, `Vec<PtxError<'_>>`, `TranslateError`, and trivially for `CompilerError` itself. The explicit map_err calls are no-ops. No behavior change.
This does not add full 32 bit support, just support in the compiler for a special 32bit on 64 bit mode. I expected the changes to be much smaller, but since they are so wide-reaching I prefer to get this piece out as soon as possible
Add cuCtxCreate_v4 and exit and a broken printf implementation
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
12 participants
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
See Commits and Changes for more details.
Created by
pull[bot] (v2.0.0-alpha.4)
Can you help keep this open source service alive? 💖 Please sponsor : )