AST/PCH files more lazy:
- Don't preload all of the file source-location entries when reading
the AST file. Instead, load them lazily, when needed.
- Only look up header-search information (whether a header was already
#import'd, how many times it's been included, etc.) when it's needed
by the preprocessor, rather than pre-populating it.
Previously, we would pre-load all of the file source-location entries,
which also populated the header-search information structure. This was
a relatively minor performance issue, since we would end up stat()'ing
all of the headers stored within a AST/PCH file when the AST/PCH file
was loaded. In the normal PCH use case, the stat()s were cached, so
the cost--of preloading ~860 source-location entries in the Cocoa.h
case---was relatively low.
However, the recent optimization that replaced stat+open with
open+fstat turned this into a major problem, since the preloading of
source-location entries would now end up opening those files. Worse,
those files wouldn't be closed until the file manager was destroyed,
so just opening a Cocoa.h PCH file would hold on to ~860 file
descriptors, and it was easy to blow through the process's limit on
the number of open file descriptors.
By eliminating the preloading of these files, we neither open nor stat
the headers stored in the PCH/AST file until they're actually needed
for something. Concretely, we went from
*** HeaderSearch Stats:
835 files tracked.
364 #import/#pragma once files.
823 included exactly once.
6 max times a file is included.
3 #include/#include_next/#import.
0 #includes skipped due to the multi-include optimization.
1 framework lookups.
0 subframework lookups.
*** Source Manager Stats:
835 files mapped, 3 mem buffers mapped.
37460 SLocEntry's allocated, 11215575B of Sloc address space used.
62 bytes of files mapped, 0 files with line #'s computed.
with a trivial program that uses a chained PCH including a Cocoa PCH
to
*** HeaderSearch Stats:
4 files tracked.
1 #import/#pragma once files.
3 included exactly once.
2 max times a file is included.
3 #include/#include_next/#import.
0 #includes skipped due to the multi-include optimization.
1 framework lookups.
0 subframework lookups.
*** Source Manager Stats:
3 files mapped, 3 mem buffers mapped.
37460 SLocEntry's allocated, 11215575B of Sloc address space used.
62 bytes of files mapped, 0 files with line #'s computed.
for the same program.
llvm-svn: 125286
gep to explicit addressing, we know that none of the intermediate
computation overflows.
This could use review: it seems that the shifts certainly wouldn't
overflow, but could the intermediate adds overflow if there is a
negative index?
Previously the testcase would instcombine to:
define i1 @test(i64 %i) {
%p1.idx.mask = and i64 %i, 4611686018427387903
%cmp = icmp eq i64 %p1.idx.mask, 1000
ret i1 %cmp
}
now we get:
define i1 @test(i64 %i) {
%cmp = icmp eq i64 %i, 1000
ret i1 %cmp
}
llvm-svn: 125271
linkage into Decl.cpp. Disable this logic for extern "C" functions, because
the operative rule there is weaker. Fixes rdar://problem/8898466
llvm-svn: 125268
exact/nsw/nuw shifts and have instcombine infer them when it can prove
that the relevant properties are true for a given shift without them.
Also, a variety of refactoring to use the new patternmatch logic thrown
in for good luck. I believe that this takes care of a bunch of related
code quality issues attached to PR8862.
llvm-svn: 125267
optimizations to be much more aggressive in the face of
exact/nsw/nuw div and shifts. For example, these (which
are the same except the first is 'exact' sdiv:
define i1 @sdiv_icmp4_exact(i64 %X) nounwind {
%A = sdiv exact i64 %X, -5 ; X/-5 == 0 --> x == 0
%B = icmp eq i64 %A, 0
ret i1 %B
}
define i1 @sdiv_icmp4(i64 %X) nounwind {
%A = sdiv i64 %X, -5 ; X/-5 == 0 --> x == 0
%B = icmp eq i64 %A, 0
ret i1 %B
}
compile down to:
define i1 @sdiv_icmp4_exact(i64 %X) nounwind {
%1 = icmp eq i64 %X, 0
ret i1 %1
}
define i1 @sdiv_icmp4(i64 %X) nounwind {
%X.off = add i64 %X, 4
%1 = icmp ult i64 %X.off, 9
ret i1 %1
}
This happens when you do something like:
(ptr1-ptr2) == 42
where the pointers are pointers to non-unit types.
llvm-svn: 125266
and generally tidying things up. Only very trivial functionality changes
like now doing (-1 - A) -> (~A) for vectors too.
InstCombineAddSub.cpp | 296 +++++++++++++++++++++-----------------------------
1 file changed, 126 insertions(+), 170 deletions(-)
llvm-svn: 125264
Natural Loop Information
Loop Pass Manager
Canonicalize natural loops
Scalar Evolution Analysis
Loop Pass Manager
Induction Variable Users
Canonicalize natural loops
Induction Variable Users
Loop Strength Reduction
into this:
Scalar Evolution Analysis
Loop Pass Manager
Canonicalize natural loops
Induction Variable Users
Loop Strength Reduction
This fixes <rdar://problem/8869639>. I also filed PR9184 on doing this sort of
thing automatically, but it seems easier to just change the ordering of the
passes if this is the only case.
llvm-svn: 125254
This is a hack because we really should only search in the 'include/clang/StaticAnalyzer' directory
if we are in 'lib/StaticAnalyzer'. My CMake knowledge is limited, so I appeal to anyone with
more expertise.
llvm-svn: 125252
When matching operands for a candidate opcode match in the auto-generated
AsmMatcher, check each operand against the expected operand match class.
Previously, operands were classified independently of the opcode being
handled, which led to difficulties when operand match classes were
more complicated than simple subclass relationships.
llvm-svn: 125245
of the CPSR during the course of executing an opcode, and modified SelectInstrSet()
to update this variable instead of the original m_inst_cpsr, which should be
the cached copy of the CPSR at the beginning of executing the opcode.
llvm-svn: 125244
Johnny Chen