Debian Codegraph @

gmp4/mpn/x86_64/fastsse/README

raw
1 This directory contains code for x86-64 processors with fast
2 implementations of SSE operations, hence the name "fastsse".
3
4 Current processors that might benefit from this code are:
5
6 AMD K10
7 AMD Bulldozer
8 Intel Nocona
9 Intel Nehalem/Westmere
10 Intel Sandybridge/Ivybridge
11 VIA Nano
12
13 Current processors that do not benefit from this code are:
14
15 AMD K8
16 AMD Bobcat
17 Intel Atom
18
19 Intel Conroe/Penryn is a border case; its handling of non-aligned
20 128-bit memory operands is poor. VIA Nano also have poor handling of
21 non-aligned operands.