Currently sle and ule have to call slt/ult and eq to get the proper answer. This results in extra code for both calls and additional scans of multiword APInts.
This patch replaces slt/ult with a compareSigned/compare that can return -1, 0, or 1 so we can cover all the comparison functions with a single call.
While I was there I removed the activeBits calls and other checks at the start of the slow part of ult. Both of the activeBits calls potentially scan through each of the APInts separately. I can't imagine that's any better than just scanning them in parallel and doing the compares. Now we just share the code with tcCompare.
These changes seem to be good for about a 7-8k reduction on the size of the opt binary on my local x86-64 build.
Differential Revision: https://reviews.llvm.org/D32339
llvm-svn: 300995
This question comes up in many places in SimplifyDemandedBits. This makes it easy to ask without allocating additional temporary APInts.
The BitVector class provides a similar functionality through its (IMHO badly named) test(const BitVector&) method. Though its output polarity is reversed.
I've provided one example use case in this patch. I plan to do more as a follow up.
Differential Revision: https://reviews.llvm.org/D32258
llvm-svn: 300851
Summary: This is a simple question we should be able to answer without creating a temporary to hold the AND result. We can also get an early out as soon as we find a word that intersects.
Reviewers: RKSimon, hans, spatel, davide
Reviewed By: hans, davide
Subscribers: llvm-commits
Differential Revision: https://reviews.llvm.org/D32253
llvm-svn: 300812
This patch uses lshrInPlace to replace code where the object that lshr is called on is being overwritten with the result.
This adds an lshrInPlace(const APInt &) version as well.
Differential Revision: https://reviews.llvm.org/D32155
llvm-svn: 300566
This merges the two different multiword shift right implementations into a single version located in tcShiftRight. lshrInPlace now calls tcShiftRight for the multiword case.
I retained the memmove fast path from lshrInPlace and used a memset for the zeroing. The for loop is basically tcShiftRight's implementation with the zeroing and the intra-shift of 0 removed.
Differential Revision: https://reviews.llvm.org/D32114
llvm-svn: 300503
This was throwing an assert because we determined the intra-word shift amount by subtracting the size of the full word shift from the total shift amount. But we failed to account for the fact that we clipped the full word shifts by total words first. To fix this just calculate the intra-word shift as the remainder of dividing by bits per word.
llvm-svn: 300405
Switch from Euclid's algorithm to Stein's algorithm for computing GCD. This
avoids the (expensive) APInt division operation in favour of bit operations.
Remove all memory allocation from within the GCD loop by tweaking our `lshr`
implementation so it can operate in-place.
Differential Revision: https://reviews.llvm.org/D31968
llvm-svn: 300252
Summary:
APInt is currently implemented with an unsigned BitWidth field first and then a uint_64/pointer union. Due to the 64-bit size of the union there is a hole after the bitwidth.
Putting the union first allows the class to be packed. Making it 12 bytes instead of 16 bytes. An APSInt goes from 20 bytes to 16 bytes.
This shows a 4k reduction on the size of the opt binary on my local x86-64 build. So this enables some other improvement to the code as well.
Reviewers: dblaikie, RKSimon, hans, davide
Reviewed By: davide
Subscribers: davide, llvm-commits
Differential Revision: https://reviews.llvm.org/D32001
llvm-svn: 300171
This patch is one step to attempt to unify the main APInt interface and the tc functions used by APFloat.
This patch adds a WordType to APInt and uses that in all the tc functions. I've added temporary typedefs to APFloat to alias it to integerPart to keep the patch size down. I'll work on removing that in a future patch.
In future patches I hope to reuse the tc functions to implement some of the main APInt functionality.
I may remove APINT_ from BITS_PER_WORD and WORD_SIZE constants so that we don't have the repetitive APInt::APINT_ externally.
Differential Revision: https://reviews.llvm.org/D31523
llvm-svn: 299341
Summary:
GreatestComonDivisor currently makes a copy of both its inputs. Then in the loop we do one move and two copies, plus any allocation the urem call does.
This patch changes it to take its inputs by value so that we can do a move of any rvalue inputs instead of copying. Then in the loop we do 3 move assignments and no copies. This way the only possible allocations we have in the loop is from the urem call.
Reviewers: dblaikie, RKSimon, hans
Reviewed By: dblaikie
Subscribers: llvm-commits
Differential Revision: https://reviews.llvm.org/D31572
llvm-svn: 299314
This method is pretty new and probably isn't use much in the code base so this should have a negligible size impact. The OR and XOR operators are already inline.
llvm-svn: 298870
I'm not sure if zeroing VAL before writing pVal is really necessary, but at least one other place did it in code.
But by taking the store out of line, this reduces the opt binary by about 20k on my local x86-64 build.
llvm-svn: 298233
We currently have to insert bits via a temporary variable of the same size as the target with various shift/mask stages, resulting in further temporary variables, all of which require the allocation of memory for large APInts (MaskSizeInBits > 64).
This is another of the compile time issues identified in PR32037 (see also D30265).
This patch adds the APInt::insertBits() helper method which avoids the temporary memory allocation and masks/inserts the raw bits directly into the target.
Differential Revision: https://reviews.llvm.org/D30780
llvm-svn: 297458
This extends an earlier change that did similar for add and sub operations.
With this first patch we lose the fastpath for the single word case as operator&= and friends don't support it. This can be added there if we think that's important.
I had to change some functions in the APInt class since the operator overloads were moved out of the class and can't be used inside the class now. The getBitsSet change collides with another outstanding patch to implement it with setBits. But I didn't want to make this patch dependent on that series.
I've also removed the Or, And, Xor functions which were rarely or never used. I already commited two changes to remove the only uses of Or that existed.
Differential Revision: https://reviews.llvm.org/D30612
llvm-svn: 297121
Summary:
There are quite a few places in the code base that do something like the following to set the high or low bits in an APInt.
KnownZero |= APInt::getHighBitsSet(BitWidth, BitWidth - 1);
For BitWidths larger than 64 this creates a short lived APInt with malloced storage. I think it might even call malloc twice. Its better to just provide methods that can set the necessary bits without the temporary APInt.
I'll update usages that benefit in a separate patch.
Reviewers: majnemer, MatzeB, davide, RKSimon, hans
Reviewed By: hans
Subscribers: llvm-commits
Differential Revision: https://reviews.llvm.org/D30525
llvm-svn: 297111
Summary:
This patch moves the clearUnusedBits calls into the two different initialization paths for APInt from a uint64_t. This allows the compiler to better optimize the clearing of the unused bits for the single word case. And it puts the clearing for the multi word case into the initSlowCase function to save code. In the common case of initializing with 0 this allows the clearing to be completely optimized out for the single word case.
On my local x86 build this is showing a ~45kb reduction in the size of the opt binary.
Reviewers: RKSimon, hans, majnemer, davide, MatzeB
Reviewed By: hans
Subscribers: llvm-commits
Differential Revision: https://reviews.llvm.org/D30486
llvm-svn: 296677
The current pattern for extract bits in range is typically:
Mask.lshr(BitOffset).trunc(SubSizeInBits);
Which can be particularly slow for large APInts (MaskSizeInBits > 64) as they require the allocation of memory for the temporary variable.
This is another of the compile time issues identified in PR32037 (see also D30265).
This patch adds the APInt::extractBits() helper method which avoids the temporary memory allocation.
Differential Revision: https://reviews.llvm.org/D30336
llvm-svn: 296272
The current pattern for extract bits in range is typically:
Mask.lshr(BitOffset).trunc(SubSizeInBits);
Which can be particularly slow for large APInts (MaskSizeInBits > 64) as they require the allocation of memory for the temporary variable.
This is another of the compile time issues identified in PR32037 (see also D30265).
This patch adds the APInt::extractBits() helper method which avoids the temporary memory allocation.
Differential Revision: https://reviews.llvm.org/D30336
llvm-svn: 296147
The current pattern for extract bits in range is typically:
Mask.lshr(BitOffset).trunc(SubSizeInBits);
Which can be particularly slow for large APInts (MaskSizeInBits > 64) as they require the allocation of memory for the temporary variable.
This is another of the compile time issues identified in PR32037 (see also D30265).
This patch adds the APInt::extractBits() helper method which avoids the temporary memory allocation.
Differential Revision: https://reviews.llvm.org/D30336
llvm-svn: 296141
The current pattern for setting bits in range is typically:
Mask |= APInt::getBitsSet(MaskSizeInBits, LoPos, HiPos);
Which can be particularly slow for large APInts (MaskSizeInBits > 64) as they require the allocation memory for the temporary variable.
This is one of the key compile time issues identified in PR32037.
This patch adds the APInt::setBits() helper method which avoids the temporary memory allocation completely, this first implementation uses setBit() internally instead but already significantly reduces the regression in PR32037 (~10% drop). Additional optimization may be possible.
I investigated whether there is need for APInt::clearBits() and APInt::flipBits() equivalents but haven't seen these patterns to be particularly common, but reusing the code would be trivial.
Differential Revision: https://reviews.llvm.org/D30265
llvm-svn: 296102
Summary: As per title. I ran into that limitation of the API doing some other work, so I though that'd be a nice addition.
Reviewers: jroelofs, compnerd, majnemer
Subscribers: llvm-commits
Differential Revision: https://reviews.llvm.org/D29503
llvm-svn: 294063
We had various variants of defining dump() functions in LLVM. Normalize
them (this should just consistently implement the things discussed in
http://lists.llvm.org/pipermail/cfe-dev/2014-January/034323.html
For reference:
- Public headers should just declare the dump() method but not use
LLVM_DUMP_METHOD or #if !defined(NDEBUG) || defined(LLVM_ENABLE_DUMP)
- The definition of a dump method should look like this:
#if !defined(NDEBUG) || defined(LLVM_ENABLE_DUMP)
LLVM_DUMP_METHOD void MyClass::dump() {
// print stuff to dbgs()...
}
#endif
llvm-svn: 293359
Summary:
There's a comment in XorSlowCase that says "0^0==1" which isn't true. 0 xored with 0 is still 0. So I don't think we need to clear any unused bits here.
Now there is no difference between XorSlowCase and AndSlowCase/OrSlowCase other than the operation being performed
Reviewers: majnemer, MatzeB, chandlerc, bkramer
Reviewed By: MatzeB
Subscribers: chfast, llvm-commits
Differential Revision: https://reviews.llvm.org/D28986
llvm-svn: 292873
This patch changes LLVM_CONSTEXPR variable declarations to const
variable declarations, since LLVM_CONSTEXPR expands to nothing if the
current compiler doesn't support constexpr. In all of the changed
cases, it looks like the code intended the variable to be const instead
of sometimes-constexpr sometimes-not.
llvm-svn: 279696
This adds versions of operator + and - which are optimized for the LHS/RHS of the
operator being RValue's. When an RValue is available, we can use its storage space
instead of allocating new space.
On code such as ConstantRange which makes heavy use of APInt's over 64-bits in size,
this results in significant numbers of saved allocations.
Thanks to David Blaikie for all the review and most of the code here.
llvm-svn: 276470
APInt::operator+(uint64_t) just forwarded to operator+(const APInt&).
Constructing the APInt for the RHS takes an allocation which isn't
required. Also, for APInt's in the slow path, operator+ would
call add() internally which iterates over both arrays of values. Instead
we can use add_1 and sub_1 which only iterate while there is something to do.
Using the memory for 'opt -O2 verify-uselistorder.lto.opt.bc -o opt.bc'
(see r236629 for details), this reduces the number of allocations from
23.9M to 22.7M.
llvm-svn: 270959
APInt::slt was copying the LHS and RHS in to temporaries then making
them unsigned so that it could use an unsigned comparision. It did
this even on the paths which were trivial to give results for, such
as the sign bit of the LHS being set while RHS was not set.
This changes the logic to return out immediately in the trivial cases,
and use an unsigned comparison in the remaining cases. But this time,
just use the unsigned comparison directly without creating any temporaries.
This works because, for example:
true = (-2 slt -1) = (0xFE ult 0xFF)
Also added some tests explicitly for slt with APInt's larger than 64-bits
so that this new code is tested.
Using the memory for 'opt -O2 verify-uselistorder.lto.opt.bc -o opt.bc'
(see r236629 for details), this reduces the number of allocations from
26.8M to 23.9M.
llvm-svn: 270881
We always clear the unused bits in the most signifant word so there is
no need to mask them out in countLeadingZerosSlowCase().
Differential Revision: http://reviews.llvm.org/D16621
llvm-svn: 260911
Previously the code used getActiveBits() to determine the highest set
bit of each APInt first. However doing so requires the same amount of
memory accesses as simply comparing both numbers right away.
Removing all the active bit checks leads to simpler code and is faster
in my benchmark.
Differential Revision: http://reviews.llvm.org/D16620
llvm-svn: 260447
The patch is generated using this command:
tools/clang/tools/extra/clang-tidy/tool/run-clang-tidy.py -fix \
-checks=-*,llvm-namespace-comment -header-filter='llvm/.*|clang/.*' \
llvm/lib/
Thanks to Eugene Kosov for the original patch!
llvm-svn: 240137
Summary: This patch fixes step D4 of Knuth's division algorithm implementation. Negative sign of the step result was not always detected due to incorrect "borrow" handling.
Test Plan: Unit test that reveals the bug included.
Reviewers: chandlerc, yaron.keren
Reviewed By: yaron.keren
Subscribers: llvm-commits
Differential Revision: http://reviews.llvm.org/D9196
llvm-svn: 235699
APInt uses Knuth's D algorithm for long division. In rare cases the
implementation applied a transformation that was not needed.
Added unit tests for long division. KnuthDiv() procedure is fully covered.
There is a case in APInt::divide() that I believe is never used (marked with
a comment) as all users of divide() handle trivial cases earlier.
Patch by Pawel Bylica!
http://reviews.llvm.org/D8448
llvm-svn: 233312
Now that SmallString is a first-class citizen, most SmallString::str()
calls are not required. This patch removes a whole bunch of them, yet
there are lots more.
There are two use cases where str() is really needed:
1) To use one of StringRef member functions which is not available in
SmallString.
2) To convert to std::string, as StringRef implicitly converts while
SmallString do not. We may wish to change this, but it may introduce
ambiguity.
llvm-svn: 232622
This mirrors the behavior of APInt::udiv and APInt::urem. Some
architectures, like X86, have a single instruction which can compute
both division and remainder.
llvm-svn: 224217
We have a transform that changes:
(x lshr C1) udiv C2
into:
x udiv (C2 << C1)
However, it is unsafe to do so if C2 << C1 discards any of C2's bits.
This fixes PR21255.
llvm-svn: 219634
Sooooo many of these had incorrect or strange main module includes.
I have manually inspected all of these, and fixed the main module
include to be the nearest plausible thing I could find. If you own or
care about any of these source files, I encourage you to take some time
and check that these edits were sensible. I can't have broken anything
(I strictly added headers, and reordered them, never removed), but they
may not be the headers you'd really like to identify as containing the
API being implemented.
Many forward declarations and missing includes were added to a header
files to allow them to parse cleanly when included first. The main
module rule does in fact have its merits. =]
llvm-svn: 169131
new hash_value infrastructure, and replace their implementations using
hash_combine. This removes a complete copy of Jenkin's lookup3 hash
function (which is both significantly slower and lower quality than the
one implemented in hash_combine) along with a somewhat scary xor-only
hash function.
Now that APInt and APFloat can be passed directly to hash_combine,
simplify the rest of the LLVMContextImpl hashing to use the new
infrastructure.
llvm-svn: 152004
toString() now takes an optional bool argument that,
depending on the radix, adds the appropriate prefix
to the integer's string representation that makes it into a
meaningful C literal, e.g.:
hexademical: '-f' becomes '-0xf'
octal: '77' becomes '077'
binary: '110' becomes '0b110'
Patch by nobled@dreamwidth.org!
llvm-svn: 133032
of a constant had a minor typo introduced when copying it from the book, which
caused it to favor negative approximations over positive approximations in many
cases. Positive approximations require fewer operations beyond the multiplication.
In the case of division by 3, we still generate code that is a single instruction
larger than GCC's code.
llvm-svn: 126097
zextOrTrunc(), and APSInt methods extend(), extOrTrunc() and new method
trunc(), to be const and to return a new value instead of modifying the
object in place.
llvm-svn: 121120
logic to use the new APInt methods. Among other things this
implements rdar://8501501 - llvm.smul.with.overflow.i32 should constant fold
which comes from "clang -ftrapv", originally brought to my attention from PR8221.
llvm-svn: 116457
payloads. APFloat's internal folding routines always make QNaNs now,
instead of sometimes making QNaNs and sometimes SNaNs depending on the
type.
llvm-svn: 97364
Craig Topper