techniques in low latency financial softwares

techniques in low latency financial softwares

Coming from a video game industry background, I just wondered about common
techniques those are being used in low latency ( as processing time , not
network latency or distributed time ) softwares for financial / HFT
softwares.
Yes , I know that definitely premature optimisation is devil, however in
particular , I wondered whether techniques used in AAA game engines are
being utilised , those are :
Multithreaded software architecture & multiprocessors ( for ex OpenMP or
SPUs ) & usage of SIMD for base 3d maths libraries
CPU cache friendly programmig against cache misses , load hit store etc ->
fitting data struct size to cache , organising datastructures as either
SoA or AoS , pre-fecthing data to your cache
data alignment for fast CPU fetchs
implementing a memory pool/pre allocated custom heap system to avoid OS
heap manager costs also things like memory fragmentation
minimising file system access etc, for instance if there is a need for
database using an in-memory system like SQLite etc
Are things like these being used there as well ?