Commit Graph

29 Commits

Author SHA1 Message Date
whitequark f6b391bb4e compiler: fix ARTIQ_DUMP_ELF. 2016-03-29 13:03:49 +00:00
whitequark 1d8b0d46bc compiler: mark FFI functions as ModRef=Ref using TBAA metadata.
Fascinatingly, the fact that you can mark call instructions with
!tbaa metadata is completely undocumented. Regardless, it is true:
a !tbaa metadata for an "immutable" type will cause
AliasAnalysis::getModRefBehavior to return OnlyReadsMemory for that
call site.

Don't bother marking loads with TBAA yet since we already place
!load.invariant on them (which is as good as the TBAA "immutable"
flag) and after that we're limited by lack of !nonnull anyway.

Also, add TBAA analysis passes in our pipeline to actually engage it.
2016-03-28 02:52:15 +00:00
whitequark 418f0a5234 compiler: mark loads of kernel constant attributes as load invariant.
Also, enable LICM, since it can take advantage of this.
2016-03-28 00:44:32 +00:00
whitequark f81930ffd6 compiler: run IPSCCP.
This doesn't do much, only frees some registers.
2016-03-27 02:00:05 +00:00
whitequark 7213984330 compiler: raise inliner threshold to the equivalent of -O3. 2016-03-27 01:08:13 +00:00
whitequark e75ad3d1aa compiler: extract runtime checks into separate cold functions.
This reduces register pressure as well as function size, which
favorably affects the inliner.
2016-03-27 01:02:15 +00:00
whitequark 20ad7629ac llvm_ir_generator: generate code more amenable to LLVM's GlobalOpt.
This exposes almost all embedded methods to inlining, with massive
gains.
2016-03-26 21:35:48 +00:00
whitequark 186a564ba8 compiler: make quoted functions independent of outer environment. 2016-03-26 20:46:45 +00:00
whitequark f5c720c3ee compiler: tune the LLVM optimizer pipeline (fixes #315). 2016-03-26 13:20:23 +00:00
whitequark f4ab507ab2 Bring back target print function. 2016-03-18 04:42:06 +00:00
whitequark 914bc9f360 artiq_run: allow running LLVM IR/bitcode files, even with trivial RPCs. 2016-02-24 22:50:45 +00:00
whitequark da31d29897 compiler: fix ARTIQ_DUMP_*. 2016-02-24 21:43:46 +00:00
whitequark a829b8a6fc Commit missing parts of cf41890. 2016-02-24 17:53:13 +00:00
whitequark cf41890255 Correctly display backtraces that contain inlined functions. 2016-02-24 17:44:19 +00:00
whitequark 442673076f compiler: setting ARTIQ_DUMP_* dumps to a file instead of stderr. 2016-02-24 17:44:19 +00:00
whitequark 74a75841f6 Add rtio_log() and make print() an RPC (#206). 2016-02-15 03:56:56 +00:00
whitequark ee1ef551bc Fix symbolizer invocation with no addresses. 2016-01-10 18:04:10 +00:00
whitequark a2a00e8b35 Set LLVM inlining threshold explicitly.
llvmlite's pass manager builder does not use a default when it
is not specified explicitly, unlike the C++ one.
2016-01-10 16:15:23 +00:00
whitequark 38a99fde52 Implement selective attribute writeback using shadow memory. 2016-01-02 22:51:04 +08:00
whitequark fd46690cf5 compiler: make IR dumps vastly more readable. 2015-11-17 00:23:34 +03:00
whitequark 5151adb9a8 compiler.targets: correctly pass CPU features to LLVM. 2015-08-30 16:56:58 -05:00
whitequark edf33f1643 compiler.targets: dump module signature with ARTIQ_DUMP_SIG=1. 2015-08-28 02:22:35 -05:00
whitequark d473d58b41 artiq_{compile,run}: adapt to new compiler. 2015-08-28 01:43:46 -05:00
whitequark 261515dfe5 compiler.targets.OR1KTarget: fix typo. 2015-08-10 15:47:44 +03:00
whitequark 75532d10aa Display full core device backtraces. 2015-08-10 15:12:22 +03:00
whitequark 8b7d38d203 Add ARTIQ_DUMP_ASSEMBLY. 2015-08-09 15:47:29 +03:00
whitequark 9c5ca2ae29 LLVMIRGenerator: add target data layout to LLVM modules. 2015-08-09 14:39:21 +03:00
whitequark aae2923c4c runtime: add lognonl{,_va} functions.
The kernels have print(), which prints aggregates (such as
arrays) piece-by-piece, and newlines would interfere.
2015-08-02 06:33:12 +03:00
whitequark e8943a008c Rename compiler/{targets/__init__.py → targets.py}. 2015-07-30 10:35:04 +03:00