similar to: [LLVMdev] Generating New Functions

On 10/14/13 5:53 AM, Abhinash Jain wrote: > How can I generate a new function at compile time and insert some instruction > onto it, through llvm pass. > eg. > > 1) %y = alloca i32, align 4 > 2) %z = alloca i32, align 4 > 3) %t = alloca i32, align 4 > 4) %2 = load i32* %y, align 4 > 5) %3 = load i32* %z, align 4 > 6) %add = add nsw i32 %2, %3 > 7) %4 = load

I am primarily interested in phi nodes and their induction variables, in ValueTracking file there is an analysis of them, but if the upper bound is inf, it is not working? 2017-08-07 11:41 GMT+02:00 Anastasiya Ruzhanskaya < anastasiya.ruzhanskaya at frtk.ru>: > So, it is not supported to determine by this instruction : %cmp = icmp slt > i32 %i.03, 99, > that %i.03 = phi i32 [ 0,

Hi, I am investigating a performance degradation between llvm-3.1 and llvm-3.2 (Note: current top-of-tree shows a similar degradation) One issue I see is the following: - 'loop invariant code motion' seems to be depending on the result of the 'reassociate expression' pass: In the samples below I observer the following behavior: Both start with the same expression: %add = add

As far as I can understand of the code, the Reassociate tries to achieve this result by its "ranking" mechanism. If it dose not, it is not hard to achieve this result, just restructure the expression in a way such that the earlier definition of the sub-expression is permute earlier in the resulting expr. e.g. outer-loop1 x= outer-loop2 y =

In trying to streamline various optimization functions, I would like to have a scratch pad of working data that is shared across a number of functions. These can be called from different levels within some wrapper functions for maximum likelihood and other such computations. I'm sure there are other applications that could benefit from this. Below are two approaches. One uses the <<-

It's an interesting problem. The best stuff I've seen published is by Cooper, Eckhart, & Kennedy, in PACT '08. Cooper gives a nice intro in one of his lectures: http://www.cs.rice.edu/~keith/512/2012/Lectures/26ReassocII-1up.pdf I can't tell, quickly, what's going on in Reassociate; as usual, the documentation resolutely avoids giving any credit for the ideas. Why is that?

On Apr 25, 2013, at 10:51 AM, Preston Briggs <preston.briggs at gmail.com> wrote: > It's an interesting problem. > The best stuff I've seen published is by Cooper, Eckhart, & Kennedy, in PACT '08. > Cooper gives a nice intro in one of his lectures: http://www.cs.rice.edu/~keith/512/2012/Lectures/26ReassocII-1up.pdf > I can't tell, quickly, what's going on

Does R have a function like the S/S++ unset() function? unset(name) would remove 'name' from the current evaluation frame and return its value. It allowed you to safely avoid some memory copying when calling .C or .Call. E.g., suppose you had C code like #include <R.h> #include <Rinternals.h> SEXP add1(SEXP pX) { int nProtected = 0; int n = Rf_length(pX);

Valery, First response of several. I don't know why the demo page at UIUC is unavailable but there is an enhanced copy of it running on the mirror at http://llvm.x10sys.com/demo/ if you ever need it. Running it produced the following LLVM equivalent for the C code in your example. implementation ; Functions: int %add1(int %x) { entry: %tmp.1 = add int %x, 1 ; <int> [#uses=1]

Shouldn't the F statistic (and p value) for the x2 term in the following calls to anova() and add1() be the same? I think anova() gets it right and add1() does not. > d <- data.frame(y=1:10, x1=log(1:10), x2=replace(1/(1:10), 2:3, NA)) > anova(lm(y ~ x1 + x2, data=d)) Analysis of Variance Table Response: y Df Sum Sq Mean Sq F value Pr(>F) x1 1

Should the F statistic be the same when using add1() on models created by lm and glm(family=gaussian)? They are in the single-degree-of-freedom case but not in the multiple-degree-of-freedom case. MASS:addterm shows the same discrepancy. It looks like the deviance (==residual sum of squares) gets divided by the number of degrees of freedom for the term twice in add1.glm. Using anova() on the

Hi. I am trying to run LLVM with JIT on ARM processor (Android phone). Currently I have problems using external functions. Any call to external function crashes and gives me signal 11 (SIGSEGV) at some random address. I'm trying to run following C code: *** extern void add1(int* x); int main() { int a = 10; int b = 20; add1(&b); int c = a + b; return c; } *** It gives

On Mon, 9 Aug 2004, Reid Spencer wrote: > Attached are three files: "valery.cpp" which contains your original, "reid.cpp" > which contains corrections to most of the FIXME items and "diffs" which shows > the differences between them. The differences should be instructive on what to > do. You were really, really close .. just a few details changing. The

Valery, Attached are three files: "valery.cpp" which contains your original, "reid.cpp" which contains corrections to most of the FIXME items and "diffs" which shows the differences between them. The differences should be instructive on what to do. You were really, really close .. just a few details changing. The code in "reid.cpp" compiles but I

Hi R-devel, If you did read my survey on Rhelp about reporting, you may have seen that I am implementing a way to handle outputs for R (mainly target output destinations: xHTML and TeX). In fact: I does have something that works for basic objects, entirely done with S4 classes, with the results visible at: http://www.stat.ucl.ac.be/ROMA/sample.htm http://www.stat.ucl.ac.be/ROMA/sample.pdf To

Hello, I am trying to figure out, what vrp propagation does in llvm. I tried this program: #include <stdio.h> int main() { int s = 0; int j = 0; for (int i = 0; i < 100; i++) { j = j+i+1; s+=j; } return (s+j); } And got this under optimized version ( I don't want everything to be eliminated) define i32 @main()

Dear List Members, I have encountered two problems when using the step function to select models. To better illustrate the problems, attached is an R image which includes the objects needed to run the code attached. lm.data.frame have factor variables with 3 levels. The following run shows the first problem. AICs (* and **) are different. I noticed that the Df for rs13482096:rs13483699 is 4,

On Thu, 13 Apr 2006 18:07:42 +0200 Oscar Fuentes <oscarfv at telefonica.net> wrote: > > Simon Burton <simon at arrowtheory.com> writes: > > > I'm trying to take assembly and create machine code I can execute. > > How close am I ? > > Your test case is not complete. Besides, which version of llvm are you > using? What are the commands for compiling

On Wed, 12 Apr 2006 23:02:12 -0500 (CDT) Chris Lattner <sabre at nondot.org> wrote: > > > I'd like to directly create executable code that i can > > stick in memory somewhere and jump into (call). > > Take a look at the llvm/examples directory. There are several small > programs that create LLVM IR on the fly and JIT compile it. I'm trying to take

Dear R Help, I have noticed some inconsistent behaviour of add1 and drop1 with a weighted linear model, which affects the interpretation of the results. I have these data to fit with a linear model, I want to weight them by the relative size of the geographical areas they represent. _________________________________________________________________________________________ > example