aboutsummaryrefslogtreecommitdiff
path: root/lib/Transforms/Vectorize/LoopVectorize.cpp
AgeCommit message (Collapse)Author
2013-04-23LoopVectorizer: Fix 15830. When scalarizing and unrolling stores make sure ↵Nadav Rotem
that the order in which the elements are scalarized is the same as the original order. This fixes a miscompilation in FreeBSD's regex library. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@180121 91177308-0d34-0410-b5e6-96231b3b80d8
2013-04-23Call the potentially costly isAnnotatedParallel() only once. Pekka Jaaskelainen
Made the uniform write test's checks a bit stricter. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@180119 91177308-0d34-0410-b5e6-96231b3b80d8
2013-04-23Refuse to (even try to) vectorize loops which have uniform writes,Pekka Jaaskelainen
even if erroneously annotated with the parallel loop metadata. Fixes Bug 15794: "Loop Vectorizer: Crashes with the use of llvm.loop.parallel metadata" git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@180081 91177308-0d34-0410-b5e6-96231b3b80d8
2013-04-19LoopVectorizer: Use matcher from PatternMatch.h for the min/max patternsArnold Schwaighofer
Also make some static function class functions to avoid having to mention the class namespace for enums all the time. No functionality change intended. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@179886 91177308-0d34-0410-b5e6-96231b3b80d8
2013-04-18Fix a -Wdocumentation warningDmitri Gribenko
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@179789 91177308-0d34-0410-b5e6-96231b3b80d8
2013-04-18LoopVectorizer: Recognize min/max reductionsArnold Schwaighofer
A min/max operation is represented by a select(cmp(lt/le/gt/ge, X, Y), X, Y) sequence in LLVM. If we see such a sequence we can treat it just as any other commutative binary instruction and reduce it. This appears to help bzip2 by about 1.5% on an imac12,2. radar://12960601 git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@179773 91177308-0d34-0410-b5e6-96231b3b80d8
2013-04-18LoopVectorize: Use a set to avoid longer cycles in the reduction chain too.Benjamin Kramer
Fixes PR15748. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@179757 91177308-0d34-0410-b5e6-96231b3b80d8
2013-04-12LoopVectorizer: integer division is not a reduction operationArnold Schwaighofer
Don't classify idiv/udiv as a reduction operation. Integer division is lossy. For example : (1 / 2) * 4 != 4/2. Example: int a[] = { 2, 5, 2, 2} int x = 80; for() x /= a[i]; Scalar: x /= 2 // = 40 x /= 5 // = 8 x /= 2 // = 4 x /= 2 // = 2 Vectorized: <80, 1> / <2,5> //= <40,0> <40, 0> / <2,2> //= <20,0> 20*0 = 0 radar://13640654 git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@179381 91177308-0d34-0410-b5e6-96231b3b80d8
2013-04-04LoopVectorizer: Pass OperandValueKind information to the cost modelArnold Schwaighofer
Pass down the fact that an operand is going to be a vector of constants. This should bring the performance of MultiSource/Benchmarks/PAQ8p/paq8p on x86 back. It had degraded to scalar performance due to my pervious shift cost change that made all shifts expensive on x86. radar://13576547 git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@178809 91177308-0d34-0410-b5e6-96231b3b80d8
2013-03-14LoopVectorize: Invert case when we use a vector cmp value to query select costArnold Schwaighofer
We generate a select with a vectorized condition argument when the condition is NOT loop invariant. Not the other way around. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@177098 91177308-0d34-0410-b5e6-96231b3b80d8
2013-03-09Remove a source of nondeterminism from the LoopVectorizer.Benjamin Kramer
This made us emit runtime checks in a random order. Hopefully bootstrap miscompares will go away now. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@176775 91177308-0d34-0410-b5e6-96231b3b80d8
2013-03-09LoopVectorizer: Ignore all dbg intrinisicArnold Schwaighofer
Ignore all DbgIntriniscInfo instructions instead of just DbgValueInst. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@176769 91177308-0d34-0410-b5e6-96231b3b80d8
2013-03-09LoopVectorizer: Ignore dbg.value instructionsArnold Schwaighofer
We want vectorization to happen at -g. Ignore calls to the dbg.value intrinsic and don't transfer them to the vectorized code. radar://13378964 git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@176768 91177308-0d34-0410-b5e6-96231b3b80d8
2013-03-08Insert the reduction start value into the first bypass block to preserve ↵Benjamin Kramer
domination. Fixes PR15344. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@176701 91177308-0d34-0410-b5e6-96231b3b80d8
2013-03-02PR14448 - prevent the loop vectorizer from vectorizing the same loop twice.Nadav Rotem
The LoopVectorizer often runs multiple times on the same function due to inlining. When this happens the loop vectorizer often vectorizes the same loops multiple times, increasing code size and adding unneeded branches. With this patch, the vectorizer during vectorization puts metadata on scalar loops and marks them as 'already vectorized' so that it knows to ignore them when it sees them a second time. PR14448. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@176399 91177308-0d34-0410-b5e6-96231b3b80d8
2013-03-01LoopVectorize: Don't hang forever if a PHI only has skipped PHI uses.Benjamin Kramer
Fixes PR15384. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@176366 91177308-0d34-0410-b5e6-96231b3b80d8
2013-02-27LoopVectorize: Vectorize math builtin calls.Benjamin Kramer
This properly asks TargetLibraryInfo if a call is available and if it is, it can be translated into the corresponding LLVM builtin. We don't vectorize sqrt() yet because I'm not sure about the semantics for negative numbers. The other intrinsic should be exact equivalents to the libm functions. Differential Revision: http://llvm-reviews.chandlerc.com/D465 git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@176188 91177308-0d34-0410-b5e6-96231b3b80d8
2013-02-21Allow GlobalValues to vectorize with AliasAnalysisRenato Golin
Storing the load/store instructions with the values and inspect them using Alias Analysis to make sure they don't alias, since the GEP pointer operand doesn't take the offset into account. Trying hard to not add any extra cost to loads and stores that don't overlap on global values, AA is *only* calculated if all of the previous attempts failed. Using biggest vector register size as the stride for the vectorization access, as we're being conservative and the cost model (which calculates the real vectorization factor) is only run after the legalization phase. We might re-think this relationship in the future, but for now, I'd rather be safe than sorry. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@175818 91177308-0d34-0410-b5e6-96231b3b80d8
2013-02-13LoopVectorize: Simplify code for clarity.Benjamin Kramer
No functionality change. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@175076 91177308-0d34-0410-b5e6-96231b3b80d8
2013-02-13Metadata for annotating loops as parallel. The first consumer for this Pekka Jaaskelainen
metadata is the loop vectorizer. See the documentation update for more info. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@175060 91177308-0d34-0410-b5e6-96231b3b80d8
2013-02-08Typos.Jakob Stoklund Olesen
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@174723 91177308-0d34-0410-b5e6-96231b3b80d8
2013-02-08ARM cost model: Address computation in vector mem ops not freeArnold Schwaighofer
Adds a function to target transform info to query for the cost of address computation. The cost model analysis pass now also queries this interface. The code in LoopVectorize adds the cost of address computation as part of the memory instruction cost calculation. Only there, we know whether the instruction will be scalarized or not. Increase the penality for inserting in to D registers on swift. This becomes necessary because we now always assume that address computation has a cost and three is a closer value to the architecture. radar://13097204 git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@174713 91177308-0d34-0410-b5e6-96231b3b80d8
2013-02-08Test CommitMichael Kuperstein
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@174709 91177308-0d34-0410-b5e6-96231b3b80d8
2013-02-07fix 80-col violation and fix the docs.Nadav Rotem
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@174671 91177308-0d34-0410-b5e6-96231b3b80d8
2013-02-07Loop Vectorizer: Refactor Memory Cost ComputationArnold Schwaighofer
We don't want too many classes in a pass and the classes obscure the details. I was going a little overboard with object modeling here. Replace classes by generic code that handles both loads and stores. No functionality change intended. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@174646 91177308-0d34-0410-b5e6-96231b3b80d8
2013-02-05Loop Vectorizer: Refactor code to compute vectorized memory instruction costArnold Schwaighofer
Introduce a helper class that computes the cost of memory access instructions. No functionality change intended. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@174422 91177308-0d34-0410-b5e6-96231b3b80d8
2013-02-05Loop Vectorizer: Handle pointer stores/loads in getWidestType()Arnold Schwaighofer
In the loop vectorizer cost model, we used to ignore stores/loads of a pointer type when computing the widest type within a loop. This meant that if we had only stores/loads of pointers in a loop we would return a widest type of 8bits (instead of 32 or 64 bit) and therefore a vector factor that was too big. Now, if we see a consecutive store/load of pointers we use the size of a pointer (from data layout). This problem occured in SingleSource/Benchmarks/Shootout-C++/hash.cpp (reduced test case is the first test in vector_ptr_load_store.ll). radar://13139343 git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@174377 91177308-0d34-0410-b5e6-96231b3b80d8
2013-01-29LoopVectorize: convert TinyTripCountVectorThreshold constantPekka Jaaskelainen
to a command line switch. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@173837 91177308-0d34-0410-b5e6-96231b3b80d8
2013-01-29LoopVectorize: Clean up ValueMap a bit and avoid double lookups.Benjamin Kramer
No intended functionality change. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@173809 91177308-0d34-0410-b5e6-96231b3b80d8
2013-01-28Vectorization Factor clarificationRenato Golin
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@173691 91177308-0d34-0410-b5e6-96231b3b80d8
2013-01-25LoopVectorize: Refactor the code that vectorizes loads/stores to remove ↵Nadav Rotem
duplication. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@173500 91177308-0d34-0410-b5e6-96231b3b80d8
2013-01-25LoopVectorize: Simplify code. No functionality change.Benjamin Kramer
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@173475 91177308-0d34-0410-b5e6-96231b3b80d8
2013-01-25LoopVectorizer: Refactor more code to use the IRBuilder. Nadav Rotem
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@173471 91177308-0d34-0410-b5e6-96231b3b80d8
2013-01-25Refactor some code to use the IRBuilder.Nadav Rotem
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@173467 91177308-0d34-0410-b5e6-96231b3b80d8
2013-01-23Add support for reverse pointer induction variables. These are loops that ↵Nadav Rotem
contain pointers that count backwards. For example, this is the hot loop in BZIP: do { m = *--p; *p = ( ... ); } while (--n); git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@173219 91177308-0d34-0410-b5e6-96231b3b80d8
2013-01-21Fix a comment. Induction vars dont need to start at zero.Nadav Rotem
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@173061 91177308-0d34-0410-b5e6-96231b3b80d8
2013-01-20LoopVectorize: Fix a C++11 incompatibility.Benjamin Kramer
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@172990 91177308-0d34-0410-b5e6-96231b3b80d8
2013-01-20Fix a build error.Nadav Rotem
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@172971 91177308-0d34-0410-b5e6-96231b3b80d8
2013-01-20LoopVectorizer: Implement a new heuristics for selecting the unroll factor.Nadav Rotem
We ignore the cpu frontend and focus on pipeline utilization. We do this because we don't have a good way to estimate the loop body size at the IR level. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@172964 91177308-0d34-0410-b5e6-96231b3b80d8
2013-01-19LoopVectorizer: Emit memory checks into their own basic block.Benjamin Kramer
This separates the check for "too few elements to run the vector loop" from the "memory overlap" check, giving a lot nicer code and allowing to skip the memory checks when we're not going to execute the vector code anyways. We still leave the decision of whether to emit the memory checks as branches or setccs, but it seems to be doing a good job. If ugly code pops up we may want to emit them as separate blocks too. Small speedup on MultiSource/Benchmarks/MallocBench/espresso. Most of this is legwork to allow multiple bypass blocks while updating PHIs, dominators and loop info. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@172902 91177308-0d34-0410-b5e6-96231b3b80d8
2013-01-15LoopVectorizer cost model. Honor the user command line flag that selects the ↵Nadav Rotem
vectorization factor even if the target machine does not have any vector registers. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@172544 91177308-0d34-0410-b5e6-96231b3b80d8
2013-01-13Fix PR14547. Handle induction variables of small sizes smaller than i32 (i8 ↵Nadav Rotem
and i16). git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@172348 91177308-0d34-0410-b5e6-96231b3b80d8
2013-01-11ARM Cost Model: We need to detect the max bitwidth of types in the loop in ↵Nadav Rotem
order to select the max vectorization factor. We don't have a detailed analysis on which values are vectorized and which stay scalars in the vectorized loop so we use another method. We look at reduction variables, loads and stores, which are the only ways to get information in and out of loop iterations. If the data types are extended and truncated then the cost model will catch the cost of the vector zext/sext/trunc operations. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@172178 91177308-0d34-0410-b5e6-96231b3b80d8
2013-01-10LoopVectorizer: Fix a bug in the vectorization of BinaryOperators. The ↵Nadav Rotem
BinaryOperator can be folded to an Undef, and we don't want to set NSW flags to undef vals. PR14878 git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@172079 91177308-0d34-0410-b5e6-96231b3b80d8
2013-01-09ARM Cost model: Use the size of vector registers and widest vectorizable ↵Nadav Rotem
instruction to determine the max vectorization factor. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@172010 91177308-0d34-0410-b5e6-96231b3b80d8
2013-01-09Cost Model: Move the 'max unroll factor' variable to the TTI and add initial ↵Nadav Rotem
Cost Model support on ARM. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@171928 91177308-0d34-0410-b5e6-96231b3b80d8
2013-01-08Code cleanup: refactor the switch statements in the generation of reduction ↵Nadav Rotem
variables into an IR builder call. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@171871 91177308-0d34-0410-b5e6-96231b3b80d8
2013-01-08Rename the enum members to match the LLVM coding style.Nadav Rotem
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@171868 91177308-0d34-0410-b5e6-96231b3b80d8
2013-01-07LoopVectorizer: Add support for floating point reductionsNadav Rotem
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@171812 91177308-0d34-0410-b5e6-96231b3b80d8
2013-01-07LoopVectorizer: When we vectorizer and widen loops we process many elements ↵Nadav Rotem
at once. This is a good thing, except for small loops. On small loops post-loop that handles scalars (and runs slower) can take more time to execute than the rest of the loop. This patch disables widening of loops with a small static trip count. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@171798 91177308-0d34-0410-b5e6-96231b3b80d8
generated by cgit v1.2.3-18-g5258 (git 2.32.0) at 2025-05-10 12:18:30 +0000