This repository was archived by the owner on Mar 25, 2025. It is now read-only.

Merge `llvm` branch to `master`#932

Draft

iomaganaris wants to merge 105 commits intomasterfrom

magkanar/llvm-rebase

Commits on Sep 15, 2022

Disable python bindings for faster build

pramodk
authored and
iomaganaris
committed
Integrate LLVM into CMake build system

pramodk
authored and
iomaganaris
committed
Code infrastructure for LLVM code generation backend

pramodk
authored and
iomaganaris
committed
Azure CI fixes for LLVM build and README update

pramodk
authored and
iomaganaris
committed
Print build status after cmake configure stage

pramodk
authored and
iomaganaris
committed
Adding test template for LLVM codegen

pramodk
authored and
iomaganaris
committed
Initial LLVM codegen vistor routines (#457)

authored and
iomaganaris
committed
FunctionBlock code generation and terminator checks (#470)

georgemitenkov
authored and
iomaganaris
committed
Add option to run LLVM optimisation passes (#471)

pramodk
authored and
iomaganaris
committed
Add function call LLVM code generation (#477)

georgemitenkov
authored and
iomaganaris
committed
Support for IndexedName codegen (#478)

georgemitenkov
authored and
iomaganaris
committed
Improvements for code generation specific transformations (#483)

pramodk
authored and
iomaganaris
committed
nrn_state function generation in NMODL AST to help LLVM codegen (#484)

pramodk
and
iomaganaris
committed
Running functions from MOD files via LLVM JIT (#482)

authored and
iomaganaris
committed
Extended support for binary ops and refactoring (#489)

georgemitenkov
authored and
iomaganaris
committed
Avoid converting LOCAL statement in all StatementBlocks (#492)

pramodk
authored and
iomaganaris
committed
Handle CodegenVarType type in JSON printer (#494)

pramodk
authored and
iomaganaris
committed
Integrating LLVM helper into LLVM visitor (#497)

georgemitenkov
authored and
iomaganaris
committed
LLVM code generation for if/else statements (#499)

georgemitenkov
authored and
iomaganaris
committed
Added error handling for values not in scope (#502)

georgemitenkov
authored and
iomaganaris
committed
Added support for WHILE statement (#501)

georgemitenkov
authored and
iomaganaris
committed
Create mechanism instance struct in LLVM IR (#507)
iomaganaris
committed
Printf support in LLVM IR codegen (#510)

georgemitenkov
authored and
iomaganaris
committed
Fix issue error: ‘runtime_error’ is not a member of ‘std’ (#512)
iomaganaris
committed
Move code gen specific InstanceStruct node to codegen.yaml (#526)

pramodk
authored and
iomaganaris
committed
* Improvements to codegen helper (Part I)

pramodk
authored and
iomaganaris
committed
Addressing TODOs for Instance struct (#533) Part II

georgemitenkov
authored and
iomaganaris
committed
Unit test for scalar state kernel generation in LLVM (#547)

georgemitenkov
authored and
iomaganaris
committed
Indexed name codegen improvements (#550)

georgemitenkov
authored and
iomaganaris
committed
Add InstanceStruct test data generation helper and unit test (#546)

iomaganaris
and
pramodk
committed
Add the remainder loop for vectorization of DERIVATIVE block (#534)

Nicolas Cornu
authored and
iomaganaris
committed
Always initialize return variable in function block (#554)

Nicolas Cornu
authored and
iomaganaris
committed
Running a kernel with NMODL-LLVM JIT (#549)

georgemitenkov
authored and
iomaganaris
committed
Loop epilogue fix for LLVM visitor helper (#567)

georgemitenkov
authored and
iomaganaris
committed
Gather support and vectorisation fixes for LLVM code generation (#568)

georgemitenkov
authored and
iomaganaris
committed
Verification and file utilities for LLVM IR codegen (#582)

georgemitenkov
authored and
iomaganaris
committed
Add gather execution test (#591)

georgemitenkov
authored and
iomaganaris
committed
Fixed loop allocations (#590)

georgemitenkov
authored and
iomaganaris
committed
Benchmarking LLVM code generation (#583)

authored and
iomaganaris
committed
Minor benchmarking improvement (#593)

pramodk
authored and
iomaganaris
committed
Bug fix in codegen helper: delete LOCAL statement (#595)

pramodk
authored and
iomaganaris
committed
LLVM 13 compatibility and fixing void* type (#603)

georgemitenkov
authored and
iomaganaris
committed
Allow LOCAL variable inside StatementBlock for LLVM IR generation (#599)

pramodk
authored and
iomaganaris
committed
Update CI with LLVM v13 (trunk) (#605)

pramodk
authored and
iomaganaris
committed
Integrating vector maths library into LLVM codegen (#604)

authored and
iomaganaris
committed
Using shared libraries in LLVM JIT (#609)

georgemitenkov
authored and
iomaganaris
committed
Avoid local std::ofstream object causing segfault (#614)

pramodk
authored and
iomaganaris
committed
Refactoring of runners' infrastructure and dumping object files (#620)

authored and
iomaganaris
committed
Optimisation levels for benchmarking (#623)

georgemitenkov
authored and
iomaganaris
committed
Adding function debug information (#628)

authored and
iomaganaris
committed
Fixed using benchmarking_info in TestRunner (#631)

georgemitenkov
authored and
iomaganaris
committed
Fixed addition of SOLVE block to kernel's FOR loop (#636)

georgemitenkov
authored and
iomaganaris
committed
IR builder redesign for LLVM IR code generation pipeline (#634)

georgemitenkov
authored and
iomaganaris
committed
Fixed initialisation of `CodegenAtomicStatement` (#642)

georgemitenkov
authored and
iomaganaris
committed
Fix instance struct data generation for testing/benchmarking (#641)

pramodk
authored and
iomaganaris
committed
Basic scatter support (#643)

georgemitenkov
authored and
iomaganaris
committed
Benchmarking code re-organisation and minor improvements (#647)

pramodk
authored and
iomaganaris
committed
Added attributes and metadata to LLVM IR compute kernels (#648)

georgemitenkov
authored and
iomaganaris
committed
Added loaded value to the stack (#655)

georgemitenkov
authored and
iomaganaris
committed
Basic predication support for LLVM backend (#652)

georgemitenkov
authored and
iomaganaris
committed
Improvements for LLVM code generation and benchmarking (#661)

georgemitenkov
authored and
iomaganaris
committed
Fixed `alloca`s insertion point for LLVM backend (#663)

georgemitenkov
authored and
iomaganaris
committed
Fast math flags for LLVM backend (#662)

georgemitenkov
authored and
iomaganaris
committed
Avoid generating LLVM IR for Functions and Procedures if inlined (#664)
iomaganaris
committed
Fixed typo in benchmarking metrics (#665)

georgemitenkov
authored and
iomaganaris
committed
Remove only inlined blocks from AST based on symtab properties (#668)
iomaganaris
committed
Use VarName on the RHS of assignment expression (#669)

pramodk
authored and
iomaganaris
committed
[LLVM] SLEEF and libsystem_m vector libraries support (#674)

authored and
iomaganaris
committed
[LLVM] Enhancements for optimization pipeline (#683)

authored and
iomaganaris
committed
[LLVM] Added saving to file utility (#685)

authored and
iomaganaris
committed
[LLVM] Aliasing and `cpu` options for LLVM visitor and the benchmark (#686)

georgemitenkov
and
iomaganaris
committed
Fix azure yaml pipeline from merge (#687)

pramodk
authored and
iomaganaris
committed
[LLVM] Support for newer versions of LLVM APIs

georgemitenkov
authored and
iomaganaris
committed
Fix build issues for the rebased branch

pramodk
authored and
iomaganaris
committed
[LLVM] Allocate InstanceStruct on the GPU using cudaMallocManaged (#815)
iomaganaris
committed
[LLVM][GPU] Separated CPU and GPU CLI options (#817)

georgemitenkov
authored and
iomaganaris
committed
[LLVM][refactoring] Added platform abstraction (#818)

georgemitenkov
authored and
iomaganaris
committed
[LLVM][GPU] Added GPU-specific AST transformations (#819)

committed
[LLVM][GPU] Basic code generation for NVPTX backend (#820)

authored and
iomaganaris
committed
Print kernel wrappers and nrn_init based on Instance Struct (#551)

iomaganaris
and
pramodk
committed
[LLVM][GPU] NVPTX specific passes for code generation (#833)

georgemitenkov
and
iomaganaris
committed
[LLVM] Code formatting changes (#838)

iomaganaris
and
Nicolas Cornu
committed
[LLVM][GPU][+refactoring] Replacement of math intrinsics with library calls (#835)

georgemitenkov
and
iomaganaris
committed
JIT invocation from python for benchmarks (#832)

Omar Awile
authored and
iomaganaris
committed
Fixes issue with debug printing of visitors (#854)
iomaganaris
committed
Support for Breakpoint block (nrn_cur) for code generation (#645)

authored and
iomaganaris
committed
[LLVM][GPU] Added CUDADriver to execute benchmark on GPU (#829)
iomaganaris
committed
[LLVM][GPU] Atomic updates support (#853)

georgemitenkov
and
iomaganaris
committed
Replaced fmt literals with fmt::format
iomaganaris
committed
[LLVM][FIX] Float generation fix in LLVM helper visitor (#865)

georgemitenkov
authored and
iomaganaris
committed
[CP-859] Code generation changes for "inline" scopmath solvers. (#859) (#868)

iomaganaris
and
olupton
committed
[LLVM] Fixes compilation with LLVM codegen disabled (#867)
iomaganaris
committed
[CP-870] Cherry-pick sympy fix from master (#872)

committed
[LLVM][SIMD] Atomic updates support (#864)

georgemitenkov
and
iomaganaris
committed
Install explicitly LLVM 13.0.1 in MacOS builds in Azure (#898)
iomaganaris
committed
Fix nrn_cur kernel code generation unit test (#892)
iomaganaris
committed
[LLVM][refactoring] Annotations pass and no more wrappers (#893)

georgemitenkov
authored and
iomaganaris
committed
[LLVM] Instantiate CodegenLLVMVisitor of PyJIT with fast math and debug symbols options (#912)
iomaganaris
committed
Fix issues after rebase
iomaganaris
committed
Update desired cmake version in github actions
iomaganaris
committed
Make Clang and CMake format happy
iomaganaris
committed
Disable llvm backend by default
iomaganaris
committed

Commits

Commits on Sep 15, 2022

Commits on Sep 16, 2022