Cell phones have to handle telecommunications and audiovisual processing while maximizing battery life.
I have few remarks and questions:
"Gather instructions fill up a vector from non-contiguous locations in memory, and scatters do the same in the opposite direction."
Generally scatter gather is at page granularity (4KB/16KB/64KB). Do you mean that they have a scatter gather at Byte granularity ?
This is what you seem to say in "1024-bit vector with 128 byte sized elements could require 128 memory accesses."
I have few remarks and questions:
"Gather instructions fill up a vector from non-contiguous locations in memory, and scatters do the same in the opposite direction."
Generally scatter gather is at page granularity (4KB/16KB/64KB). Do you mean that they have a scatter gather at Byte granularity ?
This is what you seem to say in "1024-bit vector with 128 byte sized elements could require 128 memory accesses."