SSE2

[Smart_Lubobya]

how can i transpose a matrix using SSE INSTRUCTIONS?

[TimP]

Quoting - Smart Lubobya

how can i transpose a matrix using SSE INSTRUCTIONS?

If you mean by in-lined intrinsics, this is worth while only for certain cases (4x4 for float data type), which you could find documented. More generally, you would look at MKL library which comes with Intel C++.

[JenniferJ]

You can use /QxSSE2 or /arch:SSE2 to compile, use /Qvec-report3 to report if the loop is vectorized with SSE intruction.

Jennifer

[Smart_Lubobya]

Quoting - tim18

If you mean by in-lined intrinsics, this is worth while only for certain cases (4x4 for float data type), which you could find documented. More generally, you would look at MKL library which comes with Intel C++.

"certain cases (4x4 for float data type), which you could find documented".where are these documented?

[joex26]

Quoting - Smart Lubobya

"certain cases (4x4 for float data type), which you could find documented".where are these documented?

Just guessing:

Intel optimization manual?

[TimP]

Quoting - Smart Lubobya

"certain cases (4x4 for float data type), which you could find documented".where are these documented?

Didn't you get enough hits when you typed "SSE transpose" in the search box?