Skip to content

Preserve imaginary sign in complex tan overflow handling #1434

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

Closed
agourakis82 wants to merge 1 commit into from
Closed

Preserve imaginary sign in complex tan overflow handling #1434

Show file tree
Hide file tree
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
2 changes: 1 addition & 1 deletion include/xsimd/arch/common/xsimd_common_trigo.hpp
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -870,7 +870,7 @@ namespace xsimd
return batch_type(wreal, real_batch(1.)), batch_type(wreal, wimag / d);
#else
batch_type winf(constants::infinity<real_batch>(), constants::infinity<real_batch>());
batch_type wres = select(isinf(wimag), batch_type(wreal, real_batch(1.)), batch_type(wreal, wimag / d));
batch_type wres = select(isinf(wimag), batch_type(wreal, copysign(real_batch(1.), wimag)), batch_type(wreal, wimag / d));
return select(d == real_batch(0.), winf, wres);
#endif
}
Expand Down
34 changes: 34 additions & 0 deletions test/test_complex_trigonometric.cpp
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -108,6 +108,40 @@ struct complex_trigonometric_test

void test_tan()
{
#ifndef __FAST_MATH__
// Fast-math drops overflow recovery and does not match std::tan.
using R = real_value_type;
const R zero(0);
const R huge = std::is_same_v<R, float> ? R(45) : R(356);
const R re[] = { -R(2), -R(1), -R(0.5), -zero, zero, R(0.5), R(1), R(2) };
const R im[] = { -huge, -R(20), -R(1), -zero, zero, R(1), R(20), huge };
vector_type inputs;
inputs.reserve(64 + size);
for (R real : re)
{
for (R imag : im)
{
inputs.push_back(value_type(real, imag));
}
}
while (inputs.size() % size != 0)
{
inputs.push_back(inputs.front());
}
for (size_t i = 0; i < inputs.size(); i += size)
{
batch_type in, out, ref;
detail::load_batch(in, inputs, i);
out = tan(in);
vector_type scalar(size);
for (size_t lane = 0; lane < size; ++lane)
{
scalar[lane] = std::tan(inputs[i + lane]);
}
detail::load_batch(ref, scalar, 0);
CHECK_BATCH_EQ(ref, out);
}
#endif
test_conditional_tan<real_value_type>();
}

Expand Down