Changes between Version 1 and Version 2 of Commentary/Compiler/Backends/LLVM/Installing


Ignore:
Timestamp:
Feb 25, 2010 3:07:29 AM (4 years ago)
Author:
dterei
Comment:

Move some content to other pages

Legend:

Unmodified
Added
Removed
Modified
  • Commentary/Compiler/Backends/LLVM/Installing

    v1 v2  
    6868A nice demonstration of the improvements the LLVM back-end can bring to some code though can be see at http://donsbot.wordpress.com/2010/02/21/smoking-fast-haskell-code-using-ghcs-new-llvm-codegen/ 
    6969 
    70 = LLVM Back-end Design = 
    71  
    72 The initial design tries to fit into GHC's current pipeline stages as seamlessly as possible. This allows for quicker development and focus on the core task of LLVM code generation. 
    73  
    74 The LLVM pipeline works as follows: 
    75   * New path for LLVM generation, separate from C and NCG. (path forks at compiler/main/CodeOutput.lhs, same place where C and NCG fork). 
    76   * LLVM code generation will output LLVM assembly code. 
    77   * The LLVM assembly code is translated to an object file as follows 
    78      * First, there is an '!LlvmAs' phase which generates LLVM bitcode from LLVM assembly code (using the {{{llvm-as}}} tool).  
    79      * The LLVM optimizer is run which is a series of bitcode to bitcode optimization passes (using the {{{llc}}} tool). 
    80      * Finally an object file is created from the LLVM bitcode (using the {{{llc}}} tool) 
    81   * This brings the LLVM path back to the other back-ends. 
    82   * The final state is the Link stage, which uses the system linker as with the other back-ends. 
    83  
    84 Here is a diagram of the pipeline: 
    85  
    86 {{{ 
    87   Cmm -> (codeOutput) --->(ncg) Assembler                -->(mangler, splitter) --> ('As' phase) -----> Object Code --> (link) --> executable 
    88                           \---> LLVM Assembler           --> LLVM Optimizer     --> ('llc' phase) -----/ 
    89 }}} 
    90  
    91 This approach was the easiest and thus quickest way to initially implement the LLVM back-end. Now that it is working, there is some room for additional optimisations. A potential optimisation would be to add a new linker phase for LLVM. Instead of each module just being compiled to native object code ASAP, it would be better to keep them in the LLVM bitcode format and link all the modules together using the LLVM linker. This enable all of LLVM's link time optimisations. All the user program LLVM bitcode will then be compiled to a native object file and linked with the runtime using the native system linker. 
    92  
    93  
    94 = Implementation Issues = 
    95  
    96 == LLVM Changes == 
    97  
    98 The biggest problem is that LLVM doesn't provide all the features we need. The two issues below, 'Register Pinning' and 'TNTC' are the primary examples of this. While there is a patch for LLVM to partially correct fix this, this is a problem in itself as we now must include in GHC our own version of LLVM. Eventually we need to either get the changes we need included in LLVM or improve LLVM so that the features we require could be included dynamically. 
    99  
    100 == Register Pinning == 
    101  
    102 The new back-end supports a custom calling convention to place the STG virtual registers into specific hardware registers. The current approach taken by the C back-end and NCG of having a fixed assignment of STG virtual registers to hardware registers for performance gains is not implemented in the LLVM back-end. Instead, it uses a custom calling convention to support something semantically equivalent to register pinning. The custom calling convention passes the first N variables in specific hardware registers, thus guaranteeing on all function entries that the STG virtual registers can be found in the expected hardware registers. This approach is believed to provide better performance than the register pinning used by NCG/C back-ends as it keeps the STG virtual registers mostly in hardware registers but allows the register allocator more flexibility and access to all machine registers. 
    103  
    104 == TABLES_NEXT_TO_CODE == 
    105  
    106 GHC for heap objects places the info table (meta data) and the code adjacent to each other. That is, in memory, the object firstly has a head structure, which consists of a pointer to an info table and a payload structure. The pointer points to the bottom of the info table and the closures code is placed to be straight after the info table, so to jump to the code we can just jump one past the info table pointer. The other way to do this would be to have the info table contain a pointer to the closure code. However this would then require two jumps to get to the code instead of just one jump in the optimised layout. Achieving this layout can create some difficulty, the current back-ends handle it as follows: 
    107  
    108   * The NCG can create this layout itself 
    109   * The C code generator can't. So the [wiki:Commentary/EvilMangler Evil Mangler] rearranges the GCC assembly code to achieve the layout.  
    110  
    111 There is a build option in GHC to use the unoptimised layout and instead use a pointer to the code in the info table. This layout can be enabled/disabled by using the compiler {{{#def TABLES_NEXT_TO_CODE}}}. As LLVM has no means to achieve the optimised layout and we don't wish to write an LLVM sister for the Evil Mangler, the LLVM back-end currently uses the unoptimised layout. This apparently incurs a performance penalty of 5% (source, Making a ''Fast Curry: Push/Enter vs. Eval/Apply for Higher-order Languages'', Simon Marlow and Simon Peyton Jones, 2004). 
    112  
    113 == Shared Code with NCG == 
    114  
    115 It is probable that some of the code needed by the LLVM back-end is already implemented for the NCG back-end. Some examples of this code would be the following two functions in ''compiler/main/AsmCodeGen.lhs'': 
    116  
    117   ''fixAssignsTop'':: 
    118     Changes assignments to global registers to instead assign to the !RegTable, used for non-pinned virtual registers. 
    119  
    120   ''cmmToCmm'':: 
    121     Optimises the cmm code, in particular it changes loads from global registers to instead load from the !RegTable. 
    122  
    123 == LLVM IR Representation == 
    124  
    125 The LLVM IR is modeled in GHC using an algebraic data type to represent the first order abstract syntax of the LLVM assembly code. The LLVM representation lives in the 'Llvm' subdirectory and also contains code for pretty printing. This is the same approach taken by [http://www.cs.uu.nl/wiki/Ehc/WebHome EHC]'s LLVM Back-end, and we adapted the [https://subversion.cs.uu.nl/repos/project.UHC.pub/trunk/EHC/src/ehc/LLVM.cag module] developed by them for this purpose. 
    126  
    127 It is an open question as to if this binding should be split out into its own cabal package. Please contact the GHC mailing list if you think you might be a user of such a package. 
    128  
    129  
    13070= Validate = 
    13171