search for: fround

I am trying to upgrade to the LLVM v4.0 branch, but I am seeing failures in my TableGen descriptions for conversion from FP32 to FP16 (scalar and vector). The patterns I have are along the lines of: [(set (f16 RF16:$dst), (fround (f32 RF32:$src)))] or: [(set (v2f16 VF16:$dst), (fround (v2f32 VF32:$src)))] and these now produce the errors: error: In CONV_f32_f16: Type inference contradiction found, merging 'f32' into 'f16' or: error: In CONV_v2f32_v2f16: Type inference contradiction found, merging...

...38 and PR#1139, I pointed out that round() with digits<0 would not actually round to an integer. Brian D. Ripley (hereafter BDR) fixed it in R-1.4.0, but the fix introduced a new error, PR#1254 (caught by Ole Christensen). BDR (I'm guessing) fixed the fix; here's the relevant line from fround.c: if(dig <= 0 && fabs(res) < 1e9) res = (int)floor(res + 0.5); This solution casts the result as int if it's under a billion, but it re-introduces the original problem for large numbers: R-patched 01/13/2002> round(1e15, -2) - 1e15 [1] -0.25 So...

...6.1.3 -- Function: double fprec (double X, double DIGITS) Returns the value of X rounded to DIGITS decimal digits (after the decimal point). This is the function used by R's `round()'. ^^^^^^^^^^ -- Function: double fround (double X, double DIGITS) Returns the value of X rounded to DIGITS _significant_ decimal digits. This is the function used by R's `signif()'. ^^^^^^^^^^^ I think that is crossed vis-a-vis the the corresponding R functions,...

Hi, > 1e10+5i [1] 1e+10+0e+00i > Im(1e10+5i) [1] 5 maybe little better... --- R-3.5.1.orig/src/main/complex.c 2018-03-26 07:02:25.000000000 +0900 +++ R-3.5.1/src/main/complex.c 2018-07-10 12:50:42.523874767 +0900 @@ -381,6 +381,7 @@ r->i = fround(pow10 * x->i, digits)/pow10; } else { digits = (double)(dig); + if(digits < 1) digits=1; /* a little better */ r->r = fround(x->r, digits); r->i = fround(x->i, digits); } -- Best Regards, -- Eiji NAKAMA <nakama (a) ki.rim.or.jp> "\u4e2d\u...

...have hoped for 0. The problem seems to be specifically for negative "digits". This is in 1.3.1 on Solaris 2.6 (full description at bottom). [Apologies for making everyone read this 3 times; my bug.report() burped.] Peter Dalgaard <p.dalgaard@biostat.ku.dk> says the problem is in fround.c, which reads in part: pow10 = R_pow_di(10., dig) ... return sgn * (intx + R_rint(x * pow10) / pow10); and (says Peter), > the problem is most likely that pow10 (== 1/100) has no finite binary > representation. Patches are welcome... I'm not up to patching C code, but I will...

-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.- r-devel mailing list -- Read http://www.ci.tuwien.ac.at/~hornik/R/R-FAQ.html Send "info", "help", or "[un]subscribe" (in the "body", not the subject !) To: r-devel-request@stat.math.ethz.ch _._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._

...t;f95 -xtarget=native64" export CXX="CC -xtarget=native64" export CPPFLAGS="-I/usr/sfw/include -I/opt/csw/include" export CFLAGS="-O -xlibmieee" export F77="f95 -xtarget=native64" export CXXFLAGS=-O export FFLAGS="-O4 -xlibmopt -libmil -xvector=lib -fround=nearest" export FCFLAGS=$FFLAGS export LDFLAGS="-L/usr/sfw/lib/sparcv9 -L/opt/csw/lib/sparcv9" export SHLIB_CXXLDFLAGS="-G -lCstd" export BLAS_LIBS=-xlic_lib=sunperf export LIBS="-lmvec" export SHLIB_CFLAGS=-lmvec and: ./configure --prefix=/opt/R-2.8.1 --enable-...

Hi everyone, According to the current design an intrinsic call that depends on current floating point environment (for example, rounding mode) or change it (for example by raising FP exceptions) is represented by constrained intrinsics. The latter have attached metadata that provide info about current FP environment and have attribute `IntrInaccessibleMemOnly` that helps keeping order of

...nclude -I/usr/local/include -DHAVE_CONFIG_H -g -O2 -c wilcox.c -o wilcox.o gcc -std=gnu99 -I. -I../../src/include -I../../src/include -I/usr/local/include -DHAVE_CONFIG_H -g -O2 -c signrank.c -o signrank.o rm -rf libnmath.a ar cr libnmath.a mlutils.o d1mach.o i1mach.o fmax2.o fmin2.o fprec.o fround.o ftrunc.o sign.o fsign.o imax2.o imin2.o chebyshev.o log1p.o expm1.o lgammacor.o gammalims.o stirlerr.o bd0.o gamma.o lgamma.o gamma_cody.o beta.o lbeta.o polygamma.o bessel_i.o bessel_j.o bessel_k.o bessel_y.o choose.o snorm.o sexp.o dgamma.o pgamma.o qgamma.o rgamma.o dbeta.o pbeta.o qbeta.o r...

...u.org/onlinedocs/gcc-6.2.0/gcc/Optimize-Options.html#Optimize-Options> > > -ffp-contract=style > -ffast-math > -fno-math-errno > -funsafe-math-optimizations > -fassociative-math > -freciprocal-math > -ffinite-math-only > -fno-signed-zeros > -fno-trapping-math > -frounding-math > -fsignaling-nans > -fsingle-precision-constant > > etc, and the relevant negations of these options. We can't predict how customers will choose to chain these together, so I think the LLVM optimizer and backend designs should accommodate all possibilities derived from tho...

...ions are part of FENV_ACCESS regions and which are not. This takes into account both the semantics of the #pragma as defined by the standard, and the implementation-defined default rules that apply to code outside of any #pragma. GCC currently has the following two related command-line options: -frounding-math: Do not assume default rounding mode -ftrapping-math: Assume FP operations may trap clang accepts but (basically) ignores those options. As a first step, it might make sense to have the FENV_ACCESS default behavior triggered by these options, even while the front end does not yet support...

...ault mode assumes that the current rounding mode is the > default rounding mode. But the same holds true for many other > intrinsics and even the arithmetic IR operations like add. > > If you want to stop clang from making the default rounding mode > assumption, you need to use the -frounding-math option (or one > of its equivalents), which will cause clang to emit the corresponding > constrained intrinsics instead, for those two as well all other > affected intrinsics. > > I don't see why it would make sense to add another special case > just for those two intr...

>Date: Thu, 15 Oct 1998 15:30:22 +0200 >To: Guido Masarotto <guido at sirio.stat.unipd.it> >From: Andrea Rossetti <rossetti at stud.unipg.it> >Subject: probably "round()" bug > >Hi to all R people, >I'm using the Guido Masarotto's compiled R, version 0.62.3. >there is a probably bug in round function, in fact I've done: > >>

>Date: Thu, 15 Oct 1998 15:30:22 +0200 >To: Guido Masarotto <guido at sirio.stat.unipd.it> >From: Andrea Rossetti <rossetti at stud.unipg.it> >Subject: probably "round()" bug > >Hi to all R people, >I'm using the Guido Masarotto's compiled R, version 0.62.3. >there is a probably bug in round function, in fact I've done: > >>

...dpofa.o dposl.o dqrdc.o dqrdc2.o dqrls.o dqrsl.o dqrutl.o dsvdc.o dtrco.o dtrsl.o eigen.o lminfl.o : libappl.a making mlutils.d from mlutils.c making d1mach.d from d1mach.c making i1mach.d from i1mach.c making fmax2.d from fmax2.c making fmin2.d from fmin2.c making fprec.d from fprec.c making fround.d from fround.c making ftrunc.d from ftrunc.c making sign.d from sign.c making fsign.d from fsign.c making imax2.d from imax2.c making imin2.d from imin2.c making chebyshev.d from chebyshev.c making log1p.d from log1p.c making lgammacor.d from lgammacor.c making gammalims.d from gammalims.c making...

...S regions and which are not. This takes into account > both the semantics of the #pragma as defined by the standard, > and the implementation-defined default rules that apply to code > outside of any #pragma. GCC currently has the following two > related command-line options: > > -frounding-math: Do not assume default rounding mode > -ftrapping-math: Assume FP operations may trap > > clang accepts but (basically) ignores those options. As a first > step, it might make sense to have the FENV_ACCESS default > behavior triggered by these options, even while the front en...

...Binned Residual Plot | coefplot Generic Function for Making Coefficient Plot | contr.bayes.ordered Contrast Matrices | corrplot Correlation Plot | display Functions for Processing lm, glm, mer and polr | Output | fround Formating the Rounding of Numbers | invlogit Inverse logistic function | lalonde Lalonde Dataset | matching Single Nearest Neighborhood Matching | model.matrix.bayes Construct Design Matrices | multicomp.plot Multiple Comp...

....org/onlinedocs/gcc-6.2.0/gcc/Optimize-Options.html# > Optimize-Options > > -ffp-contract=style > -ffast-math > -fno-math-errno > -funsafe-math-optimizations > -fassociative-math > -freciprocal-math > -ffinite-math-only > -fno-signed-zeros > -fno-trapping-math > -frounding-math > -fsignaling-nans > -fsingle-precision-constant > > etc, and the relevant negations of these options. We can't predict how > customers will choose to chain these together, so I think the LLVM > optimizer and backend designs should accommodate all possibilities derived...

...are not. This takes into account > both the semantics of the #pragma as defined by the standard, > and the implementation-defined default rules that apply to code > outside of any #pragma. GCC currently has the following two > related command-line options: > > -frounding-math: Do not assume default rounding mode > -ftrapping-math: Assume FP operations may trap > > clang accepts but (basically) ignores those options. As a first > step, it might make sense to have the FENV_ACCESS default > behavior triggered by these options, even wh...

.../nmath' make[3]: Entering directory `/home/mfa5/gilp/R-versions/R-devel/src/nmath' making mlutils.d from mlutils.c making d1mach.d from d1mach.c making i1mach.d from i1mach.c making fmax2.d from fmax2.c making fmin2.d from fmin2.c making fmod.d from fmod.c making fprec.d from fprec.c making fround.d from fround.c making ftrunc.d from ftrunc.c making sign.d from sign.c making fsign.d from fsign.c making fsquare.d from fsquare.c making fcube.d from fcube.c making imax2.d from imax2.c making imin2.d from imin2.c making chebyshev.d from chebyshev.c making log1p.d from log1p.c making lgammacor.d...