Skip to content

Distance computation: Fix SSE instructions #672

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

Open
twuebker wants to merge 1 commit into
base: cpp_main
Choose a base branch
from
Open

Distance computation: Fix SSE instructions #672

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
14 changes: 7 additions & 7 deletions src/distance.cpp
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -346,10 +346,10 @@ template <typename T> float DistanceInnerProduct<T>::inner_product(const T *a, c
#else
#ifdef __SSE2__
#define SSE_DOT(addr1, addr2, dest, tmp1, tmp2) \
tmp1 = _mm128_loadu_ps(addr1); \
tmp2 = _mm128_loadu_ps(addr2); \
tmp1 = _mm128_mul_ps(tmp1, tmp2); \
dest = _mm128_add_ps(dest, tmp1);
tmp1 = _mm_loadu_ps(addr1); \
tmp2 = _mm_loadu_ps(addr2); \
tmp1 = _mm_mul_ps(tmp1, tmp2); \
dest = _mm_add_ps(dest, tmp1);
__m128 sum;
__m128 l0, l1, l2, l3;
__m128 r0, r1, r2, r3;
Expand Down Expand Up @@ -458,9 +458,9 @@ template <typename T> float DistanceFastL2<T>::norm(const T *a, uint32_t size) c
#else
#ifdef __SSE2__
#define SSE_L2NORM(addr, dest, tmp) \
tmp = _mm128_loadu_ps(addr); \
tmp = _mm128_mul_ps(tmp, tmp); \
dest = _mm128_add_ps(dest, tmp);
tmp = _mm_loadu_ps(addr); \
tmp = _mm_mul_ps(tmp, tmp); \
dest = _mm_add_ps(dest, tmp);

__m128 sum;
__m128 l0, l1, l2, l3;
Expand Down