SSE4

SSE4 (Streaming SIMD Extensions 4) is a set of instructions that will be used in AMD since K10 and Intel since the Penryn version of the Core 2 processors. The second part, called SSE4.2, was introduced with the Intel Nehalem microarchitecture.

Intel SSE4 consists of 54 instructions. The first part of 47 commands appeared under the names of SSE4.1. In addition, once published seven commands as SSE4.2 from Core i7 Nehalem from the variant.

AMD supported until the release of FX processors in October 2011, no commands from SSE4, but added instead added four of his own commands and published this instruction set under the name SSE4a. The processors of the Bulldozer microarchitecture support in addition to SSE4a also SSE4.1 and 4.2 completely.

Instructions

The following is a partial list of the newly introduced commands and their applications.

SSE 4.1

  • Scalar determine - DPPS, DPPD
  • Conditional Crossfade - BLENDPS / D, BLENDVPS / D, PBLENDVB, PBLENDDW
  • Minima or maxima determine - PMINSB, PMAXSB, PMINUW, PMAXUW, PMINUD, PMAXUD, PMINSD, PMAXSD
  • Integer conversion - PMOVSXBW / -D / -Q, PMOVZXBW / -D / -Q, PMOVSXWD / Q, PMOVZXWD / Q, PMOVSXDQ, PMOVZXDQ

SSE 4.2

  • Cyclic redundancy check - CRC32
  • Advanced string operations - PCMPESTRI, PCMPESTRM, PCMPISTRI, PCMPISTRM
  • Compare two strings
  • Find characters from a specified amount
  • Find characters from specified intervals
  • Search if a string is contained in another
AMD 10h Bulldozer (microarchitecture) Advanced Vector Extensions Multiply–accumulate operation#Fused multiply.E2.80.93add
743338
de