summaryrefslogtreecommitdiff
path: root/compiler/cmm
Commit message (Collapse)AuthorAgeFilesLines
...
* codeGen: inline allocation optimization for clone array primopsJohan Tibell2014-03-221-2/+4
| | | | | | | | | | | | | | | | | | | | | | | | The inline allocation version is 69% faster than the out-of-line version, when cloning an array of 16 unit elements on a 64-bit machine. Comparing the new and the old primop implementations isn't straightforward. The old version had a missing heap check that I discovered during the development of the new version. Comparing the old and the new version would requiring fixing the old version, which in turn means reimplementing the equivalent of MAYBE_CG in StgCmmPrim. The inline allocation threshold is configurable via -fmax-inline-alloc-size which gives the maximum array size, in bytes, to allocate inline. The size does not include the closure header size. Allowing the same primop to be either inline or out-of-line has some implication for how we lay out heap checks. We always place a heap check around out-of-line primops, as they may allocate outside of our knowledge. However, for the inline primops we only allow allocation via the standard means (i.e. virtHp). Since the clone primops might be either inline or out-of-line the heap check layout code now consults shouldInlinePrimOp to know whether a primop will be inlined.
* Revert ad15c2, which causes Windows seg-faults (Trac #8834)Simon Peyton Jones2014-03-171-64/+21
| | | | | | | | We don't yet understand WHY commit ad15c2, which is to do with CmmSink, causes seg-faults on Windows, but it certainly seems to. So reverting it is a stop-gap, but we need to un-block the 7.8 release. Many thanks to awson for identifying the offending commit.
* codeGen: allocate small byte arrays of statically known size inlineJohan Tibell2014-03-142-3/+26
| | | | | | | This results in a 57% runtime decrease when allocating an array of 128 bytes on a 64-bit machine. Fixes #8876.
* Refactor inline array allocationSimon Marlow2014-03-111-12/+53
| | | | | | | | | | - Move array representation knowledge into SMRep - Separate out low-level heap-object allocation so that we can reuse it from doNewArrayOp - remove card-table initialisation, we can safely ignore the card table for newly allocated arrays.
* Represent offsets into heap objects with byte, not word, offsetsSimon Marlow2014-03-112-7/+18
| | | | | I'd like to be able to pack together non-pointer fields that are less than a word in size, and this is a necessary prerequisite.
* Eliminate duplicate code in Cmm pipelineJan Stolarek2014-02-031-51/+30
| | | | | | | | | End of Cmm pipeline used to be split into two alternative flows, depending on whether we did proc-point splitting or not. There was a lot of code duplication between these two branches. But it wasn't really necessary as the differences can be easily enclosed within an if-then-else. I observed no impact of this change on compilation performance.
* Document deprecations in HooplJan Stolarek2014-02-033-2/+30
|
* Remove unused importJan Stolarek2014-02-021-1/+0
|
* Remove redundant NoMonoLocalBinds pragmaJan Stolarek2014-02-021-5/+1
|
* Nuke dead codeJan Stolarek2014-02-013-680/+5
| | | | | | | | | | | | | * CmmRewriteAddignments module was replaced by CmmSink a long time ago. That module is now available at https://ghc.haskell.org/trac/ghc/wiki/Commentary/Compiler/Hoopl/Examples wiki page. * removeDeadAssignments function was not used and it was also moved to the above page. * I also nuked some commented out debugging code that was not used for 1,5 year.
* Simplify Control Flow Optimisations Cmm passJan Stolarek2014-02-011-31/+12
| | | | | | | It turns out that one of the cases in the optimization pass was a special case of another. I remove that specialization since it does not have impact on compilation time, and the resulting Cmm is identical.
* Remove unnecessary LANGUAGE pragmaJan Stolarek2014-02-011-5/+0
|
* Squash some spelling issuesGabor Greif2014-01-261-1/+1
|
* Allow the argument to 'reserve' to be a compile-time expressionSimon Marlow2014-01-163-13/+29
| | | | By using the constant-folder to reduce it to an integer.
* Add a way to reserve temporary stack space in high-level CmmSimon Marlow2014-01-162-1/+28
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | We occasionally need to reserve some temporary memory in a primop for passing to a foreign function. We've been using the stack for this, but when we moved to high-level Cmm it became quite fragile because primops are in high-level Cmm and the stack is supposed to be under the control of the Cmm pipeline. So this change puts things on a firmer footing by adding a new Cmm construct 'reserve'. e.g. in decodeFloat_Int#: reserve 2 = tmp { mp_tmp1 = tmp + WDS(1); mp_tmp_w = tmp; /* Perform the operation */ ccall __decodeFloat_Int(mp_tmp1 "ptr", mp_tmp_w "ptr", arg); r1 = W_[mp_tmp1]; r2 = W_[mp_tmp_w]; } reserve is described in CmmParse.y. Unfortunately the argument to reserve must be a compile-time constant. We might have to extend the parser to allow expressions with arithmetic operators if this is too restrictive. Note also that the return instruction for the procedure must be outside the scope of the reserved stack area, so we have to extract the values from the reserved area before we close the scope. This means some more local variables (r1, r2 in the example above). The generated code is more or less identical to what we had before though.
* Typo in commentGabor Greif2014-01-161-1/+1
|
* Documentation on the stack layout algorithmSimon Marlow2014-01-161-5/+94
|
* Document the fact that Areas overlap, and why.Simon Marlow2014-01-101-0/+69
|
* -ddump-cmm: don't dump the proc point stage if we didn't do anythingSimon Marlow2013-11-281-3/+6
|
* Fix up shortcut for slow callsPatrick Palka2013-11-281-4/+4
|
* Implement shortcuts for slow calls (#6084)Simon Marlow2013-11-282-0/+26
|
* When removing unreachable code, remove unreachable info tables tooSimon Peyton Jones2013-11-221-9/+22
| | | | | | | | | | This bug only shows up when you are using proc-point splitting. What was happening was: * We generate a proc-point for the stack check * And an info table * We eliminate the stack check because it's redundant * And the dangling info table caused a panic in CmmBuildInfoTables.bundle
* Improve panic printoutSimon Peyton Jones2013-11-221-1/+2
|
* Add debug dump of the list of Cmm proc pointsSimon Peyton Jones2013-11-221-0/+1
|
* Silence two AMP warningsHerbert Valerio Riedel2013-11-031-1/+5
| | | | Signed-off-by: Herbert Valerio Riedel <hvr@gnu.org>
* Revert "Implement shortcuts for slow calls that would require PAPs (#6084)"Austin Seipp2013-10-262-26/+0
| | | | This reverts commit 2f5db98e90cf0cff1a11971c85f108a7480528ed.
* Discard dead assignments in tryToInlineSimon Marlow2013-10-251-4/+26
| | | | | | | Inlining global registers and constants made code slightly larger in some cases. I finally got around to looking into why, and discovered one reason: we weren't discarding dead code in some cases. This patch fixes it.
* Implement shortcuts for slow calls that would require PAPs (#6084)Simon Marlow2013-10-252-0/+26
|
* Remove unnecessary check in CmmContFlowOptJan Stolarek2013-10-241-8/+6
| | | | Fixes #8456
* More comments about stack layoutSimon Peyton Jones2013-10-181-5/+16
|
* Improve control flow optimisation algorithmJan Stolarek2013-10-181-164/+271
| | | | | | | Fixes #8456. Previous version of control flow optimisations did not update the list of block predecessors, leading to unnecessary duplication of blocks in some cases. See Trac and comments in the code for more details.
* Clarify comments and liberalise stack-check optimisation slightlySimon Peyton Jones2013-10-182-10/+15
| | | | | | The only substantive change here is to change "==" into ">=" in the Note [Always false stack check] code. This is semantically correct, but won't have any practical impact.
* Comments about control-flow optimisationSimon Peyton Jones2013-10-181-8/+9
|
* Optimise stack checks that are always falseJan Stolarek2013-10-171-12/+12
| | | | | Fix a bug introduced in 94125c97e49987e91fa54da6c86bc6d17417f5cf. See Note [Always false stack check]
* Remove unused codeJan Stolarek2013-10-161-48/+0
| | | | | | | I am removing old loopification code that has been commented out for long long time. We now have loopification implemented in the code generator (see Note [Self-recursive tail calls]) so we won't need to resurect this old code.
* Trailing whitespacesJan Stolarek2013-10-161-5/+4
|
* Generate (old + 0) instead of Sp in stack checksJan Stolarek2013-10-161-10/+2
| | | | | | | | | | | | | | | | | | | | When compiling a function we can determine how much stack space it will use. We therefore need to perform only a single stack check at the beginning of a function to see if we have enough stack space. Instead of referring directly to Sp - as we used to do in the past - the code generator uses (old + 0) in the stack check. Stack layout phase turns (old + 0) into Sp. The idea here is that, while we need to perform only one stack check for each function, we could in theory place more stack checks later in the function. They would be redundant, but not incorrect (in a sense that they should not change program behaviour). We need to make sure however that a stack check inserted after incrementing the stack pointer checks for a respectively smaller stack space. This would not be the case if the code generator produced direct references to Sp. By referencing (old + 0) we make sure that we always check for a correct amount of stack: when converting (old + 0) to Sp the stack layout phase takes into account changes already made to stack pointer. The idea for this change came from observations made while debugging #8275.
* TyposKrzysztof Gogolewski2013-10-121-1/+1
|
* Make MO_Prefetch_Data a no-op in the C backend (#8437).Reid Barton2013-10-121-0/+1
| | | | Signed-off-by: Erik de Castro Lopo <erikd@mega-nerd.com>
* Future-proof code for upcoming `array-0.5.0.0`Herbert Valerio Riedel2013-10-111-4/+4
| | | | | | This way CPP conditionals can be avoided for the transition period. Signed-off-by: Herbert Valerio Riedel <hvr@gnu.org>
* Add support for prefetch with locality levels.Austin Seipp2013-10-013-8/+24
| | | | | | | | | | | | | | | | | This patch adds support for several new primitive operations which support using processor-specific instructions to help guide data and cache locality decisions. We have levels ranging from [0..3] For LLVM, we generate llvm.prefetch intrinsics at the proper locality level (similar to GCC.) For x86 we generate prefetch{NTA, t2, t1, t0} instructions. On SPARC and PowerPC, the locality levels are ignored. This closes #8256. Authored-by: Carter Tazio Schonwald <carter.schonwald@gmail.com> Signed-off-by: Austin Seipp <austin@well-typed.com>
* Typos in commentsGabor Greif2013-10-011-2/+2
|
* Globally replace "hackage.haskell.org" with "ghc.haskell.org"Simon Marlow2013-10-014-5/+5
|
* Expand the commentary in Note [syntax of cmm files]Simon Marlow2013-10-011-4/+29
|
* Fix BCO bitmap generation on 32-bit platforms (#8377)Takano Akio2013-09-291-40/+0
| | | | | | | On 32-bit platforms, the bitmap should be an array of 32-bit words, not Word64s. Signed-off-by: Austin Seipp <austin@well-typed.com>
* By default, only pass 128-bit SIMD vectors in registers on X86-64.Geoffrey Mainland2013-09-221-3/+20
| | | | | | | LLVM's GHC calling convention only allows 128-bit SIMD vectors to be passed in machine registers on X86-64. This may change in LLVM 3.4; the hidden flag -fllvm-pass-vectors-in-regs causes all SIMD vector widths to be passed in registers on both X86-64 and on X86-32.
* Pass 512-bit-wide vectors in registers.Geoffrey Mainland2013-09-223-0/+11
|
* Add Cmm support for 512-bit-wide values.Geoffrey Mainland2013-09-223-2/+14
|
* Pass 256-bit-wide vectors in registers.Geoffrey Mainland2013-09-223-0/+11
|
* Add Cmm support for 256-bit-wide values.Geoffrey Mainland2013-09-223-2/+14
|
View Lorry Controller Status