similar to: type.convert and doubles

Regarding this change: > CHANGES IN R 3.1.0: > NEW FEATURES: > * type.convert() (and hence by default read.table()) returns a > character vector or factor when representing a numeric input as a > double would lose accuracy. Similarly for complex inputs. > > If a file contains numeric data with unrepresentable numbers of > decimal

Hi Owen, > I have some issues with representing this as a single "fast" mode flag, it isn't a single flag, that's the whole point of using metadata. OK, right now there is only one option (the "accuracy"), true, but the intent is that others will be added, and the meaning of accuracy tightened, later. MDBuilder has a createFastFPMath method which is intended to

Duncan, I have some issues with representing this as a single "fast" mode flag, which mostly boil down to the fact that this is a very C-centric view of the world. And, since C compilers are not generally known for their awesomeness on issues of numerics, I'm not sure that's a good idea. Having something called a "fast" or "relaxed" mode implies that it is

Hi Duncan, I'm not an expert in fp accuracy question, but I had quite a few experience dealing with fp accuracy problems during compiler transformations. I think you have a step in the right direction, walking away from ULPs, which are pretty useless for the purpose of describing allowed fp optimizations IMHO. But using just "fast" keyword (or whatever else will be added in the

On 14 April 2012 20:34, Duncan Sands <baldrick at free.fr> wrote: > the verifier checks that the accuracy operand is either a floating point > number (ConstantFP) or the keyword "fast".  If "Accuracy" is zero here > then that means it wasn't ConstantFP.  Thus it must have been the keyword > "fast". I think it's assuming too much. If I write

Hi Dmitry, > I'm not an expert in fp accuracy question, but I had quite a > few experience dealing with fp accuracy problems during compiler transformations. I agree that it's a minefield which is why I intend to proceed conservatively. > I think you have a step in the right direction, walking away from ULPs, which > are pretty useless for the purpose of describing allowed

Hi Renato, > I'm not sure about this: > > + if (!Accuracy) > + // If it's not a floating point number then it must be 'fast'. > + return getFastAccuracy(); > > Since you allow accuracies bigger than 1 in setFPAccuracy(), integers > should be treated as float. Or at least assert. the verifier checks that the accuracy operand is either a floating

The attached patch is a first attempt at representing "-ffast-math" at the IR level, in fact on individual floating point instructions (fadd, fsub etc). It is done using metadata. We already have a "fpmath" metadata type which can be used to signal that reduced precision is OK for a floating point operation, eg %z = fmul float %x, %y, !fpmath !0 ... !0 = metadata

Dear all, A colleague of mine reported a problem that I fail to understand completely. He has a number of .csv files that look all very straightforward, and they all read in perfectly well using read.csv() on both his and my computer. When we try the exact same R version on the university server however, suddenly all numeric variables turn into factors. The problem is resolved by deleting the

On Sat, Apr 14, 2012 at 11:44 PM, Duncan Sands <baldrick at free.fr> wrote: > > I think you have a step in the right direction, walking away from ULPs, >> which >> are pretty useless for the purpose of describing allowed fp optimizations >> IMHO. >> But using just "fast" keyword (or whatever else will be added in the >> future) is >> not

Thanks for the updates! Minor comments: + if (!Accuracy) + // If it's not a floating point number then it must be 'fast'. + return HUGE_VALF; Can we add an assert instead of a comment? It's just as documenting and will catch any goofs. + // If it's not a floating point number then it must be 'fast'. + return !isa<ConstantFP>(MD->getOperand(0));

Here's a revised patch, plus patches showing how fpmath metadata could be turned on in clang and dragonegg (it seemed safest for the moment to condition on -ffast-math rather than on one of the flags implied by -ffast-math). Major changes: - The FPMathOperator class can no longer be used to change math settings, only to read them. Currently it can be queried for accuracy info. I split the

Hello, When evaluating different learning methods for a categorization problem with the (really useful!) caret package, I'm getting confusing results from the Kappa computation. The data is about 20,000 rows and a few dozen columns, and the categories are quite asymmetrical, 4.1% in one category and 95.9% in the other. When I train a ctree model as: model <- train(dat.dts,

Dear all, I have a set of data with 4 inputs (independent variables) and one output (dependent variable). I want to perform a regression analysis in order to fit these data to a regression model, however due to the non-linearity of the model I do not have a clue which equation to use. I am thinking of starting with a very general equation including ^3 terms and interactions between the variables

Hi, Before I try to code this myself, I'd like to know if anyone has a fixed-point routine to compute the cos and acos functions. All I need is around 3-digit accuracy. Thanks. Jean-Marc -- Jean-Marc Valin, M.Sc.A., ing. jr. LABORIUS (http://www.gel.usherb.ca/laborius) Université de Sherbrooke, Québec, Canada -------------- next part -------------- A non-text attachment was

There has been a recent thread on R-help on this, which resurrected concepts from bug reports PR#1228 and PR#6743. Since the discussion has included a lot of erroneous 'information' based on misunderstandings of floating-point computations, this is an attempt to set the record straight and explain the solutions adopted. The problem was that var(rep(0.02, 10)) was observed to be (on

Maybe a new Operator could be defined for a fast and easy double Comparison: `~~` `~~` <- function (e1, e2) all.equal(e1, e2) And document it properly.

Hi all - I am trying to tune an SVM model by optimizing the cross-validation accuracy. Maximizing this value doesn't necessarily seem to minimize the number of misclassifications. Can anyone tell me how the cross-validation accuracy is defined? In the output below, for example, cross-validation accuracy is 92.2%, while the number of correctly classified samples is (1476+170)/(1476+170+4) =

Dear all, I have a wav file that when I try to encode with the FLAC Frontend, I get "ERROR: unsupported format type 3". When I open it with an audio editor I find it is 44100 / 32 bit. I requantized it to 16 bit using the default dither and then compressed it with FLAC to get a 61 Mbyte file (the original was about 347 Mbyte). Obviously, this altered somewhat the quality, although

On Sun, Apr 15, 2012 at 1:20 PM, Renato Golin <rengolin at systemcall.org>wrote: > On 15 April 2012 09:07, Duncan Sands <baldrick at free.fr> wrote: > > Link-time optimization will sometimes result in "fast-math" functions > being > > inlined into non-fast math functions and vice-versa. This pretty much > > inevitably means that per-instruction