Files
test/source/blender/blenlib/BLI_bit_ref.hh
Aras Pranckevicius acbd952abf Cleanup: fewer iostreams related includes from BLI/BKE headers
Including <iostream> or similar headers is quite expensive, since it
also pulls in things like <locale> and so on. In many BLI headers,
iostreams are only used to implement some sort of "debug print",
or an operator<< for ostream.

Change some of the commonly used places to instead include <iosfwd>,
which is the standard way of forward-declaring iostreams related
classes, and move the actual debug-print / operator<< implementations
into .cc files.

This is not done for templated classes though (it would be possible
to provide explicit operator<< instantiations somewhere in the
source file, but that would lead to hard-to-figure-out linker error
whenever someone would add a different template type). There, where
possible, I changed from full <iostream> include to only the needed
<ostream> part.

For Span<T>, I just removed print_as_lines since it's not used by
anything. It could be moved into a .cc file using a similar approach
as above if needed.

Doing full blender build changes include counts this way:
- <iostream> 1986 -> 978
- <sstream> 2880 -> 925

It does not affect the total build time much though, mostly because
towards the end of it there's just several CPU cores finishing
compiling OpenVDB related source files.

Pull Request: https://projects.blender.org/blender/blender/pulls/111046
2023-08-16 09:51:37 +02:00

246 lines
6.1 KiB
C++

/* SPDX-FileCopyrightText: 2023 Blender Authors
*
* SPDX-License-Identifier: GPL-2.0-or-later */
#pragma once
/** \file
* \ingroup bli
*
* This file provides the basis for processing "indexed bits" (i.e. every bit has an index).
* The main purpose of this file is to define how bits are indexed within a memory buffer.
* For example, one has to define whether the first bit is the least or most significant bit and
* how endianness affect the bit order.
*
* The order is defined as follows:
* - Every indexed bit is part of an #BitInt. These ints are ordered by their address as usual.
* - Within each #BitInt, the bits are ordered from least to most significant.
*/
#include "BLI_index_range.hh"
#include "BLI_utildefines.h"
#include <iosfwd>
namespace blender::bits {
/** Using a large integer type is better because then it's easier to process many bits at once. */
using BitInt = uint64_t;
/** Number of bits that fit into #BitInt. */
static constexpr int64_t BitsPerInt = int64_t(sizeof(BitInt) * 8);
/** Shift amount to get from a bit index to an int index. Equivalent to `log(BitsPerInt, 2)`. */
static constexpr int64_t BitToIntIndexShift = 3 + (sizeof(BitInt) >= 2) + (sizeof(BitInt) >= 4) +
(sizeof(BitInt) >= 8);
/** Bit mask containing a 1 for the last few bits that index a bit inside of an #BitInt. */
static constexpr BitInt BitIndexMask = (BitInt(1) << BitToIntIndexShift) - 1;
inline BitInt mask_first_n_bits(const int64_t n)
{
BLI_assert(n >= 0);
BLI_assert(n <= BitsPerInt);
if (n == BitsPerInt) {
return BitInt(-1);
}
return (BitInt(1) << n) - 1;
}
inline BitInt mask_last_n_bits(const int64_t n)
{
return ~mask_first_n_bits(BitsPerInt - n);
}
inline BitInt mask_range_bits(const int64_t start, const int64_t size)
{
BLI_assert(start >= 0);
BLI_assert(size >= 0);
const int64_t end = start + size;
BLI_assert(end <= BitsPerInt);
if (end == BitsPerInt) {
return mask_last_n_bits(size);
}
return ((BitInt(1) << end) - 1) & ~((BitInt(1) << start) - 1);
}
inline BitInt mask_single_bit(const int64_t bit_index)
{
BLI_assert(bit_index >= 0);
BLI_assert(bit_index < BitsPerInt);
return BitInt(1) << bit_index;
}
inline BitInt *int_containing_bit(BitInt *data, const int64_t bit_index)
{
return data + (bit_index >> BitToIntIndexShift);
}
inline const BitInt *int_containing_bit(const BitInt *data, const int64_t bit_index)
{
return data + (bit_index >> BitToIntIndexShift);
}
/**
* This is a read-only pointer to a specific bit. The value of the bit can be retrieved, but
* not changed.
*/
class BitRef {
private:
/** Points to the exact integer that the bit is in. */
const BitInt *int_;
/** All zeros except for a single one at the bit that is referenced. */
BitInt mask_;
friend class MutableBitRef;
public:
BitRef() = default;
/**
* Reference a specific bit in an array. Note that #data does *not* have to point to the
* exact integer the bit is in.
*/
BitRef(const BitInt *data, const int64_t bit_index)
{
int_ = int_containing_bit(data, bit_index);
mask_ = mask_single_bit(bit_index & BitIndexMask);
}
/**
* Return true when the bit is currently 1 and false otherwise.
*/
bool test() const
{
const BitInt value = *int_;
const BitInt masked_value = value & mask_;
return masked_value != 0;
}
operator bool() const
{
return this->test();
}
};
/**
* Similar to #BitRef, but also allows changing the referenced bit.
*/
class MutableBitRef {
private:
/** Points to the integer that the bit is in. */
BitInt *int_;
/** All zeros except for a single one at the bit that is referenced. */
BitInt mask_;
public:
MutableBitRef() = default;
/**
* Reference a specific bit in an array. Note that #data does *not* have to point to the
* exact int the bit is in.
*/
MutableBitRef(BitInt *data, const int64_t bit_index)
{
int_ = int_containing_bit(data, bit_index);
mask_ = mask_single_bit(bit_index & BitIndexMask);
}
/**
* Support implicitly casting to a read-only #BitRef.
*/
operator BitRef() const
{
BitRef bit_ref;
bit_ref.int_ = int_;
bit_ref.mask_ = mask_;
return bit_ref;
}
/**
* Return true when the bit is currently 1 and false otherwise.
*/
bool test() const
{
const BitInt value = *int_;
const BitInt masked_value = value & mask_;
return masked_value != 0;
}
operator bool() const
{
return this->test();
}
/**
* Change the bit to a 1.
*/
void set()
{
*int_ |= mask_;
}
/**
* Change the bit to a 0.
*/
void reset()
{
*int_ &= ~mask_;
}
/**
* Change the bit to a 1 if #value is true and 0 otherwise. If the value is highly unpredictable
* by the CPU branch predictor, it can be faster to use #set_branchless instead.
*/
void set(const bool value)
{
if (value) {
this->set();
}
else {
this->reset();
}
}
/**
* Does the same as #set, but does not use a branch. This is faster when the input value is
* unpredictable for the CPU branch predictor (best case for this function is a uniform random
* distribution with 50% probability for true and false). If the value is predictable, this is
* likely slower than #set.
*/
void set_branchless(const bool value)
{
const BitInt value_int = BitInt(value);
BLI_assert(ELEM(value_int, 0, 1));
const BitInt old = *int_;
*int_ =
/* Unset bit. */
(~mask_ & old)
/* Optionally set it again. The -1 turns a 1 into `0x00...` and a 0 into `0xff...`. */
| (mask_ & ~(value_int - 1));
}
MutableBitRef &operator|=(const bool value)
{
if (value) {
this->set();
}
return *this;
}
MutableBitRef &operator&=(const bool value)
{
if (!value) {
this->reset();
}
return *this;
}
};
std::ostream &operator<<(std::ostream &stream, const BitRef &bit);
std::ostream &operator<<(std::ostream &stream, const MutableBitRef &bit);
} // namespace blender::bits
namespace blender {
using bits::BitRef;
using bits::MutableBitRef;
} // namespace blender