Skip to content

limits: Construct floating point infinity value from bit pattern #463

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.

Sign up for GitHub

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

Merged
stotko merged 1 commit into from
Oct 14, 2025
Merged

limits: Construct floating point infinity value from bit pattern #463

Changes from all commits
Commits
Show all changes
1 commit
Select commit
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: 42 additions & 2 deletions src/stdgpu/impl/limits_detail.h
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -19,13 +19,40 @@
#include <cfloat>
#include <climits>
#include <cmath>
#include <cstddef>
#include <cstdint>

#include <stdgpu/compiler.h>

namespace stdgpu
{

namespace detail
{

// Almost identical to variant in bit.h
template <typename To, typename From>
constexpr STDGPU_HOST_DEVICE To
simple_bit_cast(const From& object) noexcept
{
To result;

// NOLINTNEXTLINE(cppcoreguidelines-pro-type-reinterpret-cast)
auto* result_bytes = reinterpret_cast<std::byte*>(&result);

// NOLINTNEXTLINE(cppcoreguidelines-pro-type-reinterpret-cast)
const auto* object_bytes = reinterpret_cast<const std::byte*>(&object);

for (std::size_t i = 0; i < sizeof(To); ++i)
{
result_bytes[i] = object_bytes[i];
}

return result;
}

} // namespace detail

template <typename T>
constexpr STDGPU_HOST_DEVICE T
numeric_limits<T>::min() noexcept
Expand Down Expand Up @@ -641,7 +668,13 @@ numeric_limits<float>::round_error() noexcept
constexpr STDGPU_HOST_DEVICE float
numeric_limits<float>::infinity() noexcept
{
// Definition of HUGE_VALF has changed in recent versions and may cause warnings
#if STDGPU_HOST_COMPILER == STDGPU_HOST_COMPILER_MSVC
constexpr uint32_t infinity_bits = 0b01111111100000000000000000000000;
return detail::simple_bit_cast<float>(infinity_bits);
#else
return HUGE_VALF;
#endif
}

constexpr STDGPU_HOST_DEVICE double
Expand Down Expand Up @@ -677,7 +710,13 @@ numeric_limits<double>::round_error() noexcept
constexpr STDGPU_HOST_DEVICE double
numeric_limits<double>::infinity() noexcept
{
// Definition of HUGE_VAL has changed in recent versions and may cause warnings
#if STDGPU_HOST_COMPILER == STDGPU_HOST_COMPILER_MSVC
constexpr uint64_t infinity_bits = 0b0111111111110000000000000000000000000000000000000000000000000000;
return detail::simple_bit_cast<double>(infinity_bits);
#else
return HUGE_VAL;
#endif
}

constexpr STDGPU_HOST_DEVICE long double
Expand Down Expand Up @@ -713,9 +752,10 @@ numeric_limits<long double>::round_error() noexcept
constexpr STDGPU_HOST_DEVICE long double
numeric_limits<long double>::infinity() noexcept
{
// Suppress long double is treated as double in device code warning for MSVC on CUDA
// Definition of HUGE_VALL has changed in recent versions and may cause warnings
#if STDGPU_HOST_COMPILER == STDGPU_HOST_COMPILER_MSVC
return HUGE_VAL;
constexpr uint64_t infinity_bits = 0b0111111111110000000000000000000000000000000000000000000000000000;
return detail::simple_bit_cast<double>(infinity_bits);
#else
return HUGE_VALL;
#endif
Expand Down