Skip to content

[offload][SYCL] Add Module splitting by categories. #131347

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Open
wants to merge 8 commits into
base: main
Choose a base branch
from
Open
Show file tree
Hide file tree
Changes from all commits
Commits
File filter

Filter by extension

Filter by extension

Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
44 changes: 44 additions & 0 deletions llvm/include/llvm/Transforms/Utils/SplitModuleByCategory.h
Original file line number Diff line number Diff line change
@@ -0,0 +1,44 @@
//===-------- SplitModuleByCategory.h - module split ------------*- C++ -*-===//
//
// Part of the LLVM Project, under the Apache License v2.0 with LLVM Exceptions.
// See https://llvm.org/LICENSE.txt for license information.
// SPDX-License-Identifier: Apache-2.0 WITH LLVM-exception
//
//===----------------------------------------------------------------------===//
// Functionality to split a module by categories.
//===----------------------------------------------------------------------===//

#ifndef LLVM_TRANSFORM_UTILS_SPLIT_MODULE_BY_CATEGORY_H
#define LLVM_TRANSFORM_UTILS_SPLIT_MODULE_BY_CATEGORY_H

#include "llvm/ADT/STLFunctionalExtras.h"

#include <memory>
#include <optional>
#include <string>

namespace llvm {

class Module;
class Function;

/// Splits the given module \p M using the given \p FunctionCategorizer.
/// \p FunctionCategorizer returns integer category for an input Function.
Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

A side note, I think it'd be more helpful (at least putting on my AMD hat) to be able to determine where a global variable goes as well, if we'd like to make this pass generic to support all potential targets. The reason is, for AMDGPU, we probably need to categorize all functions that could potentially reference a global variable in the sam module, due to the lowering of LDS (shared) variables.

/// It may return std::nullopt if a function doesn't have a category.
/// Module's functions are being grouped by categories. Every such group
/// populates a call graph containing group's functions themselves and all
/// reachable functions and globals. Split outputs are populated from each call
/// graph associated with some category.
///
/// Every split output is being passed to \p Callback for further possible
/// processing.
///
/// Currently, the supported targets are SPIRV, AMDGPU and NVPTX.
Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Please update this comment. I am not sure if the targets are restrictive. I think the restriction is whether the input module has recursive calls or not.

Thanks

Copy link
Contributor

@shiltian shiltian Jun 9, 2025

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Agreed. Now we have call backs so it should just work for all.

Update:

This is probably because isKernel function.

Copy link
Contributor Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

This is probably because isKernel function.

Yes and the algorithm was implemented with assumption that the input is a heterogenous program, which usually don't have recursion.

void splitModuleByCategory(
std::unique_ptr<Module> M,
function_ref<std::optional<int>(const Function &F)> FunctionCategorizer,
function_ref<void(std::unique_ptr<Module> Part)> Callback);

} // namespace llvm

#endif // LLVM_TRANSFORM_UTILS_SPLIT_MODULE_BY_CATEGORY_H
1 change: 1 addition & 0 deletions llvm/lib/Transforms/Utils/CMakeLists.txt
Original file line number Diff line number Diff line change
Expand Up @@ -82,6 +82,7 @@ add_llvm_component_library(LLVMTransformUtils
SimplifyLibCalls.cpp
SizeOpts.cpp
SplitModule.cpp
SplitModuleByCategory.cpp
StripNonLineTableDebugInfo.cpp
SymbolRewriter.cpp
UnifyFunctionExitNodes.cpp
Expand Down
325 changes: 325 additions & 0 deletions llvm/lib/Transforms/Utils/SplitModuleByCategory.cpp
Original file line number Diff line number Diff line change
@@ -0,0 +1,325 @@
//===-------- SplitModuleByCategory.cpp - split a module by categories ----===//
//
// Part of the LLVM Project, under the Apache License v2.0 with LLVM Exceptions.
// See https://llvm.org/LICENSE.txt for license information.
// SPDX-License-Identifier: Apache-2.0 WITH LLVM-exception
//
//===----------------------------------------------------------------------===//
// See comments in the header.
//===----------------------------------------------------------------------===//

#include "llvm/Transforms/Utils/SplitModuleByCategory.h"
#include "llvm/ADT/SetVector.h"
#include "llvm/ADT/SmallPtrSet.h"
#include "llvm/ADT/StringExtras.h"
#include "llvm/IR/Constants.h"
#include "llvm/IR/Function.h"
#include "llvm/IR/InstIterator.h"
#include "llvm/IR/Instructions.h"
#include "llvm/IR/Module.h"
#include "llvm/Support/Debug.h"
#include "llvm/Transforms/Utils/Cloning.h"

#include <map>
#include <string>
#include <utility>

using namespace llvm;

#define DEBUG_TYPE "split-module-by-category"

namespace {

// A vector that contains a group of function with the same category.
using EntryPointSet = SetVector<const Function *>;

/// Represents a group of functions with one category.
struct EntryPointGroup {
int ID;
EntryPointSet Functions;

EntryPointGroup() = default;

EntryPointGroup(int ID, EntryPointSet Functions = EntryPointSet())
: ID(ID), Functions(std::move(Functions)) {}

void clear() { Functions.clear(); }

#if !defined(NDEBUG) || defined(LLVM_ENABLE_DUMP)
LLVM_DUMP_METHOD void dump() const {
constexpr size_t INDENT = 4;
dbgs().indent(INDENT) << "ENTRY POINTS"
<< " " << ID << " {\n";
for (const Function *F : Functions)
dbgs().indent(INDENT) << " " << F->getName() << "\n";

dbgs().indent(INDENT) << "}\n";
}
#endif
};

/// Annotates an llvm::Module with information necessary to perform and track
/// the result of code (llvm::Module instances) splitting:
/// - entry points group from the module.
class ModuleDesc {
std::unique_ptr<Module> M;
Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

I assume ModuleDesc "owns" a module after splitting?

Copy link
Contributor Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

ModuleDesc own an initial module and new created modules.

EntryPointGroup EntryPoints;

public:
ModuleDesc(std::unique_ptr<Module> M,
EntryPointGroup EntryPoints = EntryPointGroup())
: M(std::move(M)), EntryPoints(std::move(EntryPoints)) {
assert(this->M && "Module should be non-null");
}

Module &getModule() { return *M; }
const Module &getModule() const { return *M; }

std::unique_ptr<Module> releaseModule() {
EntryPoints.clear();
return std::move(M);
}

#if !defined(NDEBUG) || defined(LLVM_ENABLE_DUMP)
LLVM_DUMP_METHOD void dump() const {
dbgs() << "ModuleDesc[" << M->getName() << "] {\n";
EntryPoints.dump();
dbgs() << "}\n";
}
#endif
};

bool isKernel(const Function &F) {
return F.getCallingConv() == CallingConv::SPIR_KERNEL ||
F.getCallingConv() == CallingConv::AMDGPU_KERNEL ||
F.getCallingConv() == CallingConv::PTX_Kernel;
}

// Represents "dependency" or "use" graph of global objects (functions and
// global variables) in a module. It is used during device code split to
// understand which global variables and functions (other than entry points)
// should be included into a split module.
//
// Nodes of the graph represent LLVM's GlobalObjects, edges "A" -> "B" represent
// the fact that if "A" is included into a module, then "B" should be included
// as well.
//
// Examples of dependencies which are represented in this graph:
// - Function FA calls function FB
// - Function FA uses global variable GA
// - Global variable GA references (initialized with) function FB
// - Function FA stores address of a function FB somewhere
//
// The following cases are treated as dependencies between global objects:
// 1. Global object A is used within by a global object B in any way (store,
// bitcast, phi node, call, etc.): "A" -> "B" edge will be added to the
// graph;
// 2. function A performs an indirect call of a function with signature S and
// there is a function B with signature S. "A" -> "B" edge will be added to
// the graph;
class DependencyGraph {
public:
using GlobalSet = SmallPtrSet<const GlobalValue *, 16>;

DependencyGraph(const Module &M) {
// Group functions by their signature to handle case (2) described above
DenseMap<const FunctionType *, DependencyGraph::GlobalSet>
FuncTypeToFuncsMap;
for (const Function &F : M.functions()) {
// Kernels can't be called (either directly or indirectly).
if (isKernel(F))
continue;

FuncTypeToFuncsMap[F.getFunctionType()].insert(&F);
}

for (const Function &F : M.functions()) {
// case (1), see comment above the class definition
for (const Value *U : F.users())
addUserToGraphRecursively(cast<const User>(U), &F);

// case (2), see comment above the class definition
for (const Instruction &I : instructions(F)) {
const CallBase *CB = dyn_cast<CallBase>(&I);
if (!CB || !CB->isIndirectCall()) // Direct calls were handled above
continue;

const FunctionType *Signature = CB->getFunctionType();
GlobalSet &PotentialCallees = FuncTypeToFuncsMap[Signature];
Graph.emplace_or_assign(&F, std::move(PotentialCallees));
}
}

// And every global variable (but their handling is a bit simpler)
for (const GlobalVariable &GV : M.globals())
for (const Value *U : GV.users())
addUserToGraphRecursively(cast<const User>(U), &GV);
}

iterator_range<GlobalSet::const_iterator>
dependencies(const GlobalValue *Val) const {
auto It = Graph.find(Val);
return (It == Graph.end())
? make_range(EmptySet.begin(), EmptySet.end())
: make_range(It->second.begin(), It->second.end());
}

private:
void addUserToGraphRecursively(const User *Root, const GlobalValue *V) {
SmallVector<const User *, 8> WorkList;
WorkList.push_back(Root);

while (!WorkList.empty()) {
const User *U = WorkList.pop_back_val();
if (const auto *I = dyn_cast<const Instruction>(U)) {
const Function *UFunc = I->getFunction();
Graph[UFunc].insert(V);
} else if (isa<const Constant>(U)) {
if (const auto *GV = dyn_cast<const GlobalVariable>(U))
Graph[GV].insert(V);
// This could be a global variable or some constant expression (like
// bitcast or gep). We trace users of this constant further to reach
// global objects they are used by and add them to the graph.
for (const User *UU : U->users())
WorkList.push_back(UU);
} else {
llvm_unreachable("Unhandled type of function user");
}
}
}

DenseMap<const GlobalValue *, GlobalSet> Graph;
Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

nit: this const is not necessary. it can allow to use dyn_cast<xxx>, which looks more LLVM, than dyn_cast<const xxx>.

Copy link
Contributor Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

That would require me to use const_cast like the following.

void addUserToGraphRecursively(const User *Root, const GlobalValue *V) {
  SmallVector<User *, 8> WorkList;
  WorkList.push_back(const_cast<User *>(Root));
  /// ...
}

Would that be ok in LLVM style?

SmallPtrSet<const GlobalValue *, 1> EmptySet;
};

void collectFunctionsAndGlobalVariablesToExtract(
SetVector<const GlobalValue *> &GVs, const Module &M,
const EntryPointGroup &ModuleEntryPoints, const DependencyGraph &DG) {
// We start with module entry points
for (const Function *F : ModuleEntryPoints.Functions)
GVs.insert(F);

// Non-discardable global variables are also include into the initial set
for (const GlobalVariable &GV : M.globals())
if (!GV.isDiscardableIfUnused())
GVs.insert(&GV);

// GVs has SetVector type. This type inserts a value only if it is not yet
// present there. So, recursion is not expected here.
size_t Idx = 0;
while (Idx < GVs.size()) {
const GlobalValue *Obj = GVs[Idx++];

for (const GlobalValue *Dep : DG.dependencies(Obj)) {
if (const auto *Func = dyn_cast<const Function>(Dep)) {
if (!Func->isDeclaration())
GVs.insert(Func);
} else {
GVs.insert(Dep); // Global variables are added unconditionally
}
}
}
}

ModuleDesc extractSubModule(const Module &M,
const SetVector<const GlobalValue *> &GVs,
EntryPointGroup ModuleEntryPoints) {
// For each group of entry points collect all dependencies.
ValueToValueMapTy VMap;
// Clone definitions only for needed globals. Others will be added as
// declarations and removed later.
std::unique_ptr<Module> SubM = CloneModule(
M, VMap, [&](const GlobalValue *GV) { return GVs.contains(GV); });
// Replace entry points with cloned ones.
EntryPointSet NewEPs;
const EntryPointSet &EPs = ModuleEntryPoints.Functions;
llvm::for_each(
EPs, [&](const Function *F) { NewEPs.insert(cast<Function>(VMap[F])); });
ModuleEntryPoints.Functions = std::move(NewEPs);
return ModuleDesc{std::move(SubM), std::move(ModuleEntryPoints)};
}

// The function produces a copy of input LLVM IR module M with only those
// functions and globals that can be called from entry points that are specified
// in ModuleEntryPoints vector, in addition to the entry point functions.
ModuleDesc extractCallGraph(const Module &M, EntryPointGroup ModuleEntryPoints,
const DependencyGraph &DG) {
SetVector<const GlobalValue *> GVs;
collectFunctionsAndGlobalVariablesToExtract(GVs, M, ModuleEntryPoints, DG);

ModuleDesc SplitM = extractSubModule(M, GVs, std::move(ModuleEntryPoints));
LLVM_DEBUG(SplitM.dump());
return SplitM;
}

using EntryPointGroupVec = SmallVector<EntryPointGroup>;

/// Module Splitter.
/// It gets a module (in a form of module descriptor, to get additional info)
/// and a collection of entry points groups. Each group specifies subset entry
/// points from input module that should be included in a split module.
class ModuleSplitter {
private:
ModuleDesc Input;
EntryPointGroupVec Groups;
DependencyGraph DG;

private:
EntryPointGroup drawEntryPointGroup() {
assert(Groups.size() > 0 && "Reached end of entry point groups list.");
EntryPointGroup Group = std::move(Groups.back());
Groups.pop_back();
return Group;
}

public:
ModuleSplitter(ModuleDesc MD, EntryPointGroupVec GroupVec)
: Input(std::move(MD)), Groups(std::move(GroupVec)),
DG(Input.getModule()) {
assert(!Groups.empty() && "Entry points groups collection is empty!");
}

/// Gets next subsequence of entry points in an input module and provides
/// split submodule containing these entry points and their dependencies.
ModuleDesc getNextSplit() {
return extractCallGraph(Input.getModule(), drawEntryPointGroup(), DG);
}

/// Check that there are still submodules to split.
bool hasMoreSplits() const { return Groups.size() > 0; }
};

EntryPointGroupVec
selectEntryPointGroups(const Module &M,
function_ref<std::optional<int>(const Function &F)> FC) {
// std::map is used here to ensure stable ordering of entry point groups,
// which is based on their contents, this greatly helps LIT tests
std::map<int, EntryPointSet> EntryPointsMap;

for (const auto &F : M.functions())
if (std::optional<int> Category = FC(F); Category)
EntryPointsMap[*Category].insert(&F);

EntryPointGroupVec Groups;
Groups.reserve(EntryPointsMap.size());
// Start with properties of a source module
for (auto &[Key, EntryPoints] : EntryPointsMap)
Groups.emplace_back(Key, std::move(EntryPoints));

return Groups;
}

} // namespace

void llvm::splitModuleByCategory(
std::unique_ptr<Module> M,
function_ref<std::optional<int>(const Function &F)> FunctionCategorizer,
function_ref<void(std::unique_ptr<Module> Part)> Callback) {
EntryPointGroupVec Groups = selectEntryPointGroups(*M, FunctionCategorizer);
ModuleDesc MD = std::move(M);
ModuleSplitter Splitter(std::move(MD), std::move(Groups));
while (Splitter.hasMoreSplits()) {
ModuleDesc MD = Splitter.getNextSplit();
Callback(std::move(MD.releaseModule()));
}
}
17 changes: 17 additions & 0 deletions llvm/test/tools/llvm-split/SplitByCategory/amd-kernel-split.ll
Original file line number Diff line number Diff line change
@@ -0,0 +1,17 @@
; -- Per-kernel split
; RUN: llvm-split -split-by-category=kernel -S < %s -o %tC
; RUN: FileCheck %s -input-file=%tC_0.ll --check-prefixes CHECK-A0
; RUN: FileCheck %s -input-file=%tC_1.ll --check-prefixes CHECK-A1

define dso_local amdgpu_kernel void @KernelA() {
ret void
}

define dso_local amdgpu_kernel void @KernelB() {
ret void
}

; CHECK-A0: define dso_local amdgpu_kernel void @KernelB()
; CHECK-A0-NOT: define dso_local amdgpu_kernel void @KernelA()
; CHECK-A1-NOT: define dso_local amdgpu_kernel void @KernelB()
; CHECK-A1: define dso_local amdgpu_kernel void @KernelA()
Loading
Loading