[clang] Introduce paged vector (PR #66430)

Jakub Kuderski via cfe-commits cfe-commits at lists.llvm.org
Mon Sep 18 09:39:02 PDT 2023


================
@@ -0,0 +1,301 @@
+//===- llvm/ADT/PagedVector.h - 'Lazyly allocated' vectors --------*- 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
+//
+//===----------------------------------------------------------------------===//
+//
+// This file defines the PagedVector class.
+//
+//===----------------------------------------------------------------------===//
+#ifndef LLVM_ADT_PAGEDVECTOR_H
+#define LLVM_ADT_PAGEDVECTOR_H
+
+#include "llvm/Support/Allocator.h"
+#include <cassert>
+#include <iostream>
+#include <vector>
+
+namespace llvm {
+// A vector that allocates memory in pages.
+// Order is kept, but memory is allocated only when one element of the page is
+// accessed. This introduces a level of indirection, but it is useful when you
+// have a sparsely initialised vector where the full size is allocated upfront
+// with the default constructor and elements are initialised later, on first
+// access.
+//
+// Notice that this does not have iterators, because if you
+// have iterators it probably means you are going to touch
+// all the memory in any case, so better use a std::vector in
+// the first place.
+//
+// Pages are allocated in SLAB_SIZE chunks, using the BumpPtrAllocator.
+template <typename T, std::size_t PAGE_SIZE = 1024 / sizeof(T)>
+class PagedVector {
+  static_assert(PAGE_SIZE > 0, "PAGE_SIZE must be greater than 0. Most likely "
+                               "you want it to be greater than 16.");
+  // The actual number of element in the vector which can be accessed.
+  std::size_t Size = 0;
+
+  // The position of the initial element of the page in the Data vector.
+  // Pages are allocated contiguously in the Data vector.
+  mutable std::vector<T *> PageToDataIdx;
+  // Actual page data. All the page elements are added to this vector on the
+  // first access of any of the elements of the page. Elements default
+  // constructed and elements of the page are stored contiguously. The order of
+  // the elements however depends on the order of access of the pages.
+  uintptr_t Allocator = 0;
+
+  constexpr static T *invalidPage() { return reinterpret_cast<T *>(SIZE_MAX); }
+
+public:
+  // Default constructor. We build our own allocator.
+  PagedVector()
+      : Allocator(reinterpret_cast<uintptr_t>(new BumpPtrAllocator) | 0x1) {}
+  PagedVector(BumpPtrAllocator *A)
+      : Allocator(reinterpret_cast<uintptr_t>(A)) {}
+
+  ~PagedVector() {
+    // If we own the allocator, delete it.
+    if (Allocator & 0x1) {
+      delete getAllocator();
+    }
+  }
+
+  // Get the allocator.
+  BumpPtrAllocator *getAllocator() const {
+    return reinterpret_cast<BumpPtrAllocator *>(Allocator & ~0x1);
+  }
+  // Lookup an element at position Index.
+  T &operator[](std::size_t Index) const { return at(Index); }
+
+  // Lookup an element at position i.
+  // If the associated page is not filled, it will be filled with default
+  // constructed elements. If the associated page is filled, return the element.
+  T &at(std::size_t Index) const {
+    assert(Index < Size);
+    assert(Index / PAGE_SIZE < PageToDataIdx.size());
+    auto *&PagePtr = PageToDataIdx[Index / PAGE_SIZE];
+    // If the page was not yet allocated, allocate it.
+    if (PagePtr == invalidPage()) {
+      PagePtr = getAllocator()->template Allocate<T>(PAGE_SIZE);
+      // We need to invoke the default constructor on all the elements of the
+      // page.
+      for (std::size_t I = 0; I < PAGE_SIZE; ++I) {
+        new (PagePtr + I) T();
+      }
+    }
+    // Dereference the element in the page.
+    return *((Index % PAGE_SIZE) + PagePtr);
+  }
+
+  // Return the capacity of the vector. I.e. the maximum size it can be expanded
+  // to with the expand method without allocating more pages.
+  std::size_t capacity() const { return PageToDataIdx.size() * PAGE_SIZE; }
+
+  // Return the size of the vector. I.e. the maximum index that can be
+  // accessed, i.e. the maximum value which was used as argument of the
+  // expand method.
+  std::size_t size() const { return Size; }
+
+  // Expands the vector to the given NewSize number of elements.
+  // If the vector was smaller, allocates new pages as needed.
+  // It should be called only with NewSize >= Size.
+  void expand(std::size_t NewSize) {
+    // You cannot shrink the vector, otherwise
+    // one would have to invalidate contents which is expensive and
+    // while giving the false hope that the resize is cheap.
+    if (NewSize <= Size) {
+      return;
+    }
+    // If the capacity is enough, just update the size and continue
+    // with the currently allocated pages.
+    if (NewSize <= capacity()) {
----------------
kuhar wrote:

Could we set `Size = NewSize` before this `if` to handle both outcomes?

https://github.com/llvm/llvm-project/pull/66430


More information about the cfe-commits mailing list