similar to: Bug or inaccuracy in cumsum( )

Full_Name: Jeremiah Cohen Version: 2.9.0 OS: Windows XP Submission from: (NULL) (129.59.230.235) I believe there is a bug in the seq() function for certain values of the "from" argument. Here are examples: > seq(-.2, .1, .1) [1] -0.2 -0.1 0.0 0.1 > seq(-.3, .1, .1) [1] -3.000000e-01 -2.000000e-01 -1.000000e-01 5.551115e-17 1.000000e-01 > seq(-.4, .1, .1) [1] -0.4 -0.3

Hi everyone, I've tried the below on R 1.9.1 and the 2004-08-30 builds of R 1.9.1 Patched and R 2.0.0 on Windows 2000, and the results are consistent. > seq(0.5, 0, by = -0.1) [1] 0.5 0.4 0.3 0.2 0.1 0.0 > seq(0.7, 0, by = -0.1) [1] 7.000000e-01 6.000000e-01 5.000000e-01 4.000000e-01 3.000000e-01 2.000000e-01 1.000000e-01 -1.110223e-16 Is this really the intended behaviour?

Hello everyone, I am using the following program to get the p-value of some numbers (column 'LR' of the data.dat file). I want to write the 1st and 2nd column of the output file (data.out) as an integer while the program change them. Could anybody please tell me how I can write the code which writes the values of the first two columns as integer? For your convenience, I have attached the

Hi, I'm having trouble with the "which" or the "seq" function, I'm not sure. Here's an example : > lat=seq(1,2,by=0.1) > lat [1] 1.0 1.1 1.2 1.3 1.4 1.5 1.6 1.7 1.8 1.9 2.0 > which(lat==1) [1] 1 > which(lat==1.1) [1] 2 > which(lat==1.2) [1] 3 > which(lat==1.3) [1] 4 > which(lat==1.4) [1] 5 > which(lat==1.5) [1] 6 >

Hi, I have a following vector: > smallch [1] 0.0652840 0.1181300 0.0319370 0.0155700 0.0464110 0.0107850 [7] 0.0158970 0.0375900 0.0603090 0.0310300 0.0105920 0.0540580 [13] -0.0177740 0.0039393 Pretty (R 1.5.1) has problems with zero: > pretty(smallch) [1] -2.000000e-02 -3.469447e-18 2.000000e-02 4.000000e-02 6.000000e-02 [6] 8.000000e-02 1.000000e-01 1.200000e-01

Hi, I have a following vector: > smallch [1] 0.0652840 0.1181300 0.0319370 0.0155700 0.0464110 0.0107850 [7] 0.0158970 0.0375900 0.0603090 0.0310300 0.0105920 0.0540580 [13] -0.0177740 0.0039393 Pretty (R 1.5.1) has problems with zero: > pretty(smallch) [1] -2.000000e-02 -3.469447e-18 2.000000e-02 4.000000e-02 6.000000e-02 [6] 8.000000e-02 1.000000e-01 1.200000e-01

I have the following output from pretty(): > pretty(c(-.1, 1)) # note 2nd element [1] -2.000000e-01 -2.775558e-17 2.000000e-01 4.000000e-01 6.000000e-01 [6] 8.000000e-01 1.000000e+00 > as.character(pretty(c(-.1, 1))) [1] "-0.2" "-2.77555756156289e-17" "0.2" [4] "0.4" "0.6"

Hi Roel, the code you list below is precisely what I expect to get (of course the stores must happen but the constant folding should happen as well). It all looks very strange. LLVM is behaving as if the __restrict__ keyword was not used at all. Even more strange is the fact that for this function: double f(double *__restrict__ x, double *__restrict__ y, double *__restrict__ z) { *x = 1.0;

Hi Brent, I think this is a problem in the easy-cse transform. In this transform load operations can be replaced by their subexpression, in this case the propagated constant, based on the value of the 'CurrentGeneration' of memory writes. This implies that any store operation invalidates the knowledge about previously stored subexpressions. In general, this is a safe assumption but

Hi Brent, Looking at your code I can see at least one reason why some of the store operations remain in the output since you are (through x, y, and z) writing in memory which exists outside of your function (p). Constant propagation also seems to work in the first few lines, *y = *x +1 (%3) is stored directly. The strange thing to me is that the same doesn't happen for *z = *x + 2. Here

I have no clue, I didn't have time to look into that example yet. How does the IR (before optimization) differ from the other version? Roel On 01/24/2012 04:45 PM, Brent Walker wrote: > Can you explain please why it works for this version of the function: > > double f(double *__restrict__ x, double *__restrict__ y, double > *__restrict__ z); > > What is different here?

I think the problem here is that the IR doesn't have any way to attach restrict information to loads/stores/pointers. It works on arguments because they can be given the 'noalias' attribute, and then the alias analyzer must understand what that means. Pete On Jan 24, 2012, at 7:47 AM, Roel Jordans wrote: > I have no clue, I didn't have time to look into that example yet.

Can you explain please why it works for this version of the function: double f(double *__restrict__ x, double *__restrict__ y, double *__restrict__ z); What is different here? There are stores here as well. Brent On Wed, Jan 25, 2012 at 12:34 AM, Roel Jordans <r.jordans at tue.nl> wrote: > Hi Brent, > > I think this is a problem in the easy-cse transform. In this transform

Peter Cooper wrote: > I think the problem here is that the IR doesn't have any way to attach restrict information to loads/stores/pointers. I think we do now, actually. Now that the loads and stores have TBAA metadata, I think the restrict attribute can go there. It needs to be attached to every use of a restrict pointer, but that's similar to how TBAA already works. > It works

Hi LLVMers, I would like to ask a question regarding aliasing. Suppose I have the following program: double f(double** p ) { double a,b,c; double * x = &a; double * y = &b; double * z = &c; *x = 1; *y = *x + 2; *z = *x + 3; return *x+*y+*z; } LLVM can tell that the three pointers do not alias each other so can perform the constant folding at compile time.

Hi all, R> seq(-.4, .8, length=7) [1] -4.000000e-01 -2.000000e-01 5.551115e-17 2.000000e-01 4.000000e-01 [6] 6.000000e-01 8.000000e-01 Well that may not be a "bug", but it has this unfortunate consequence: R> pretty(c(-.4,.8)) [1] -4.000000e-01 -2.000000e-01 5.551115e-17 2.000000e-01 4.000000e-01 [6] 6.000000e-01 8.000000e-01 And thus my plot axes look funny

I am using 2.2.0 If the QR decomposition of an N*M matrix is such that the pivoting order is not 1:M, Q%*%R does not result in the original matrix but in a matrix with the columns permuted. This is clearly intentional, and probably to be expected if pivoting is used --- chol() behaves in the same manner (it would perhaps be nice if the qr help page made that clear in the same way that the chol()

Full_Name: Version: 2.7.1 (2008-06-23) OS: windows vista Submission from: (NULL) (216.82.144.137) Hello, I have been working with various probability distributions in R, and it seems the gamma distribution is inaccurate for some inputs. For example, qgamma(1e-100, 5e-101, lower.tail=FALSE) gives: 1.0. However, it seems this is incorrect; I think the correct answer should be

Hello, I have been working with various probability distributions in R, and it seems the gamma distribution is inaccurate for some inputs. For example, qgamma(1e-100, 5e-101, lower.tail=FALSE) gives: 1.0. However, it seems this is incorrect; I think the correct answer should be 0.082372029620717283. When I check these numbers using pgamma, I get: pgamma(1,5e-101, lower.tail=FALSE) =

This has been driving me crazy. In order to JIT stuff in a simple REPL I'm trying to wrap them in anonymous functions and then evaluate them (as per recommendation of the Kaleidoscope guide). With simple expressions it works fine. Like if I try and add 4 and 5, I get this from the bare llvm::Value dump: double 9.000000e+00 and this after it's wrapped in a function: define double @0() {