Transpose of 8X8 matrix using intrinsics

[encoder]

Hello,
I am looking for an 8x8 or a 16x16 transpose of a matrix using SIMD operations through Intel intrinsics. Please let me know if such an instruction exists.

Thanks.

[SergeyKostrov]

Quoting encoder

I am looking for an 8x8 or a 16x16 transpose of a matrix using SIMD operations through Intel intrinsics...

Please take a look ata thread:

http://software.intel.com/en-us/forums/showthread.php?t=103465&o=a&s=lr

andat a thread ( more generic )about some performance issues withthe transpose operation:

http://software.intel.com/en-us/forums/showthread.php?t=104353&o=a&s=lr

[SergeyKostrov]

Quoting Sergey Kostrov

Quoting encoder

I am looking for an 8x8 or a 16x16 transpose of a matrix using SIMD operations through Intel intrinsics...

Please take a look ata thread:

http://software.intel.com/en-us/forums/showthread.php?t=103465&o=a&s=lr

andat a thread ( more generic )about some performance issues withthe transpose operation:

http://software.intel.com/en-us/forums/showthread.php?t=104353&o=a&s=lr

This is a follow up: Did it help?

Best regards,
Sergey

[SergeyKostrov]

Quoting encoder

I am looking for an 8x8 or a 16x16 transpose of a matrix...

IPP library hasa set of highly optimizedfunctionsto do a matrix transpose. Unfortunately, there is nosuch
SIMD instruction(s).