Changes between Version 1 and Version 2 of SIMD/Implementation/Llvm


Ignore:
Timestamp:
Apr 16, 2012 1:25:55 PM (2 years ago)
Author:
gmainland
Comment:

--

Legend:

Unmodified
Added
Removed
Modified
  • SIMD/Implementation/Llvm

    v1 v2  
    11[[PageOutline]] 
     2 
     3'''This page is obsolete'''. Please see the [wiki:SIMD top-level SIMD project page] for further details. 
    24 
    35= Using SIMD Instructions via the LLVM Backend = 
     
    68 
    79The SIMD vector extension to GHC proposed here maps to LLVM's vector type in a straight forward manner, which in turn enables us to target a wide range of hardware capabilities. However, GHC's native code generator will simply map SIMD vector operations to ordinary scalar code (in order to avoid having to deal with the complexities of SSE, AVX, NEON, etc). 
    8  
    9 Related pages 
    10  * [wiki:SIMDPlan Implementation plan] 
    11  * There is also also the outdated [wiki:VectorComputing]. 
    1210 
    1311== Variations in the most widely used SIMD extensions == 
     
    116114 
    117115'''NB:''' The use of SIMD instructions interferes with vectorisation avoidance for scalar subcomputations. Code that avoids vectorisation also avoids the use of SIMD instructions. We would like to use SIMD instructions, but still avoid full-scale vectorisation. This should be possible, but it is not immediately clear how to realise it (elegantly). 
    118  
    119 == Implementation Details and Plan == 
    120 For implementation details and a plan, refer to the [wiki:SIMDPlan SIMD Implementation Details and Plan] Page