Is there any reference for the latency of AVX2 instructions, such as latency for vgather, vpshufb, etc.? I got some related information from APPENDIX C of the Intel® 64 and IA-32 Architectures Optimization Reference Manual, but looks not all the AVX2 instructions are fully contained in that manual.
AVX2 is not available in any processors. Maybe Intel is still implementing some features and therefore does not exactly know latency. Or it still takes some time to update information. For example, information for all AVX instructions for Ivy Bridge in intrisics guide was updated a few weeks ago.
I looked in several manuals but could not find information about the instructions you mentioned - at least not about latency. Maybe some Intel member can provide additional information.