RPCS3/llvm-mirror
1
0
Fork 0
mirror of https://github.com/RPCS3/llvm-mirror.git synced 2024-10-25 14:02:52 +02:00
Code Issues Projects Releases Wiki Activity
120,178 Commits 4 Branches 5 Tags 704 MiB
6cbdb95607
Commit Graph

198 Commits

Author SHA1 Message Date
Neil Booth
054a756682 Prior commit updated wrong if, apologies. 
llvm-svn: 44614
 2007-12-05 13:06:04 +00:00
Neil Booth
b5afbd8e5a Handle zero correctly. 
llvm-svn: 44613
 2007-12-05 13:01:24 +00:00
Dale Johannesen
15712dc45a Handle 0 correctly in string->APFloat conversion. 
llvm-svn: 44594
 2007-12-05 01:10:19 +00:00
Dale Johannesen
ea47b8b77d Fix denormal check in float->APInt conversion. 
PR 1804.

llvm-svn: 44201
 2007-11-17 01:02:27 +00:00
Neil Booth
cd1d243989 Remove some unnecessary C-style statics. 
Restore an assertion that arithmetic can be performed on this format.

llvm-svn: 43638
 2007-11-02 15:10:05 +00:00
Neil Booth
7eeb095d37 Add back line whose removal somehow crept into prior patch 
llvm-svn: 43627
 2007-11-01 22:51:07 +00:00
Neil Booth
d04447b7ae When converting to integer, do bit manipulations in the destination 
memory rather than in a copy of the APFloat.  This avoids problems
when the destination is wider than our significand and is cleaner.

Also provide deterministic values in all cases where conversion
fails, namely zero for NaNs and the minimal or maximal value
respectively for underflow or overflow.

llvm-svn: 43626
 2007-11-01 22:43:37 +00:00
Hartmut Kaiser
48b3098b43 Clarified operator precedence. 
Silenced VC++ warning.

llvm-svn: 43372
 2007-10-25 23:15:31 +00:00
Neil Booth
e14f19eda6 Fast-track obviously over-large and over-small exponents during decimal-> 
integer conversion.  In some such cases this makes us one or two orders
of magnitude faster than NetBSD's libc.  Glibc seems to have a similar
fast path.

Also, tighten up some upper bounds to save a bit of memory.

llvm-svn: 42984
 2007-10-15 15:00:55 +00:00
Neil Booth
1463947f03 Consolidate logic for creating NaNs. Silence compiler warning. 
llvm-svn: 42966
 2007-10-14 10:39:51 +00:00
Neil Booth
138d8ef4a7 Whether arithmetic is supported is a property of the semantics. Make it 
so, and clean up the checks by putting them in an inline function.

llvm-svn: 42965
 2007-10-14 10:29:28 +00:00
Neil Booth
a352eb14cd Separate out parsing of decimal number. Use this to only allocate 
memory for the significand once up-front.  Also ignore insignificant
trailing zeroes; this saves unnecessary multiplications later.

llvm-svn: 42964
 2007-10-14 10:16:12 +00:00
Neil Booth
f3582ac0e0 If the power of 5 is exact, and the reciprocal exact, the error is zero not one half-ulps. This prevents an infinite loop in rare cases. 
llvm-svn: 42950
 2007-10-13 03:34:08 +00:00
Neil Booth
fb373da5fd Remove duplicate comment. 
llvm-svn: 42913
 2007-10-12 16:05:57 +00:00
Neil Booth
078f455d15 Implement correctly-rounded decimal->binary conversion, i.e. conversion 
from user input strings.

Such conversions are more intricate and subtle than they may appear;
it is unlikely I have got it completely right first time.  I would
appreciate being informed of any bugs and incorrect roundings you
might discover.

llvm-svn: 42912
 2007-10-12 16:02:31 +00:00
Neil Booth
4a0b91bebe Remove a field that was never used. 
llvm-svn: 42911
 2007-10-12 15:35:10 +00:00
Neil Booth
9b83d8be07 If we're trying to be arbitrary precision, unsigned char clearly won't cut it. Needed for dec->bin conversions. 
llvm-svn: 42910
 2007-10-12 15:33:27 +00:00
Dale Johannesen
0ee2a2fb59 Next PPC long double bits. First cut at constants. 
No compile-time support for constant operations yet,
just format transformations.  Make readers and
writers work.  Split constants into 2 doubles in
Legalize.

llvm-svn: 42865
 2007-10-11 18:07:22 +00:00
Neil Booth
e46d6dbd99 Use APInt::tcExtract. It's cleaner, and works :) 
llvm-svn: 42746
 2007-10-08 14:39:42 +00:00
Neil Booth
6f9602ac12 Add back convertFromSignExtendedInteger. 
llvm-svn: 42735
 2007-10-07 12:15:41 +00:00
Neil Booth
ac998332c3 Now that convertFromUnsignedParts has a sane, constant interface, 
convertFromZeroExtendedInteger can be simplified as it doesn't need
to make a copy of the source bignum.

llvm-svn: 42734
 2007-10-07 12:10:57 +00:00
Neil Booth
db4dbedfea Reimplement convertFromUnsignedInteger so it is passed a const bignum. 
It used to modify its argument in-place.

This interface is saner and the implementation more efficient.  It will
be needed for decimal->binary conversion.

llvm-svn: 42733
 2007-10-07 12:07:53 +00:00
Neil Booth
6e01c5df37 convertFromInteger, as originally written, expected sign-extended 
input.  APInt unfortunately zero-extends signed integers, so Dale
modified the function to expect zero-extended input.  Make this
assumption explicit in the function name.

llvm-svn: 42732
 2007-10-07 11:45:55 +00:00
Neil Booth
726126f4d6 combineLostFractions does not need to be a member function 
llvm-svn: 42729
 2007-10-07 08:51:21 +00:00
Neil Booth
beba899ff0 Cleaner, more general exponent output. 
llvm-svn: 42690
 2007-10-06 07:29:25 +00:00
Chris Lattner
a1de966c80 silence warnings in no-assert build. 
llvm-svn: 42687
 2007-10-06 06:13:42 +00:00
Neil Booth
2c4419d71b Generalize tcFullMultiply so that the operands can be of differing 
part widths.  Also, return the number of parts actually required to
hold the result's value.
Remove an over-cautious condition from rounding of float->hex conversion.

llvm-svn: 42669
 2007-10-06 00:24:48 +00:00
Neil Booth
2dc065755d Add APFloat -> hexadecimal string conversion, as per %a and %A in C99. 
Useful for diagnostics and debugging.

llvm-svn: 42598
 2007-10-03 22:26:02 +00:00
Neil Booth
1e709921bc Tweak RoundAwayFromZero the bit number below which is truncated, and make 
it const.

Preparation for APFloat -> hexadecimal string conversion.

llvm-svn: 42576
 2007-10-03 15:16:41 +00:00
Dale Johannesen
4b203b8e44 Simplify and fix signed int -> FP conversions. 
llvm-svn: 42483
 2007-09-30 18:17:01 +00:00
Neil Booth
3b53530d78 Whitespace and compiler warning cleanup. 
llvm-svn: 42373
 2007-09-26 21:33:42 +00:00
Dale Johannesen
bc41438bab Make APFloat->int conversions deterministic even in 
cases with undefined behavior.

llvm-svn: 42328
 2007-09-25 23:07:07 +00:00
Dale Johannesen
ee2f3359f4 Fix long double<->shorter FP type conversions 
of zero, infinity, and NaNs.

llvm-svn: 42298
 2007-09-25 17:25:00 +00:00
Neil Booth
8cb885c389 Handle storage complications of float->float conversions. 
llvm-svn: 42220
 2007-09-22 02:56:19 +00:00
Dale Johannesen
c7279629cc Change APFloat::convertFromInteger to take the incoming 
bit width instead of number of words allocated, which
makes it actually work for int->APF conversions.
Adjust callers.  Add const to one of the APInt constructors
to prevent surprising match when called with const
argument.

llvm-svn: 42210
 2007-09-21 22:09:37 +00:00
Dale Johannesen
346a949bd8 Fix PR 1688. See comments there. 
llvm-svn: 42181
 2007-09-20 23:47:58 +00:00
Dan Gohman
53119779c2 Remove spurious consts. This fixes warnings with compilers that 
are strict about such things.

llvm-svn: 41956
 2007-09-14 20:08:19 +00:00
Dale Johannesen
4b03ae1ff5 Compensate for partCount change in Bogus definition 
(could break hash table in ConstantFP)

llvm-svn: 41874
 2007-09-12 01:22:05 +00:00
Dale Johannesen
7bc3969cea Add APInt interfaces to APFloat (allows directly 
access to bits).  Use them in place of float and
double interfaces where appropriate.
First bits of x86 long double constants handling 
(untested, probably does not work).

llvm-svn: 41858
 2007-09-11 18:32:33 +00:00
Dale Johannesen
1d92d23a89 Fix mod so it actually works. Fix conversions to 
native types to handle denormals correctly.

llvm-svn: 41726
 2007-09-05 20:39:49 +00:00
Dale Johannesen
b336f6207c Oops, should be part of 41664; won't work very well without this piece. 
llvm-svn: 41665
 2007-08-31 23:35:31 +00:00
Dale Johannesen
81d6ecb886 Enhance APFloat to retain bits of NaNs (fixes oggenc). 
Use APFloat interfaces for more references, mostly
of ConstantFPSDNode.

llvm-svn: 41632
 2007-08-31 04:03:46 +00:00
Dale Johannesen
e91a908971 Change LegalFPImmediates to use APFloat. 
Add APFloat interfaces to ConstantFP, SelectionDAG.
Fix integer bit in double->APFloat conversion.
Convert LegalizeDAG to use APFloat interface in
ConstantFPSDNode uses.

llvm-svn: 41587
 2007-08-30 00:23:21 +00:00
Dale Johannesen
9ec1e28e79 Poison APFloat::operator==. Replace existing uses with bitwiseIsEqual. 
This means backing out the preceding change to Constants.cpp, alas.

llvm-svn: 41378
 2007-08-24 22:09:56 +00:00
Dale Johannesen
0aaabb858b Revised per review feedback from previous patch. 
llvm-svn: 41353
 2007-08-24 05:08:11 +00:00
Chris Lattner
91df26be0f silence some warnings in an optimized build. 
llvm-svn: 41352
 2007-08-24 03:02:34 +00:00
Dale Johannesen
a3ab055b9d Change internal representation of ConstantFP to use APFloat. 
Interface to rest of the compiler unchanged, as yet.

llvm-svn: 41348
 2007-08-24 00:56:33 +00:00
Chris Lattner
401698e8b5 initial checkin of Neil's APFloat work. 
llvm-svn: 41203
 2007-08-20 22:49:32 +00:00