similar to: [LLVMdev] Vectorization: Next Steps

On Fri, 2012-02-03 at 20:59 -0800, Preston Briggs wrote: > Hal wrote: > > 3. Loop vectorization - It would be nice to have, in addition to > > basic-block vectorization, a more-traditional loop vectorization > pass. I > > think that we'll need a better loop analysis pass in order for this > to > > happen. Some of this was started in LoopDependenceAnalysis, but

On Sat, Feb 4, 2012 at 2:27 PM, Hal Finkel <hfinkel at anl.gov> wrote: > On Fri, 2012-02-03 at 20:59 -0800, Preston Briggs wrote: >> so are building a dependence graph for a complete function.  Of >> course, such a thing is useful for vectorization and all sorts of >> other dependence-based loop transforms. >> >> I'm looking at the problem in two parts:

On Mon, 2012-02-06 at 14:13 -0600, Sebastian Pop wrote: > On Sat, Feb 4, 2012 at 2:27 PM, Hal Finkel <hfinkel at anl.gov> wrote: > > On Fri, 2012-02-03 at 20:59 -0800, Preston Briggs wrote: > >> so are building a dependence graph for a complete function. Of > >> course, such a thing is useful for vectorization and all sorts of > >> other dependence-based

Preston Briggs <preston.briggs at gmail.com> writes: > On Mon, Feb 6, 2012 at 12:13 PM, Sebastian Pop <spop at codeaurora.org> wrote: >> [many things, but I'm only going to focus on one of them] >> Would you consider using Polly http://polly.grosser.es to avoid >> writing this code? > > My impression is that Polly (and polyhedral analysis generally) >

On Mon, Feb 6, 2012 at 12:13 PM, Sebastian Pop <spop at codeaurora.org> wrote: > [many things, but I'm only going to focus on one of them] > Would you consider using Polly http://polly.grosser.es to avoid > writing this code? My impression is that Polly (and polyhedral analysis generally) doesn't do want I want. But I'm happy to talk about it 'cause I might be

Shanmuhka wrote: > I looked at the sanjoys patch for SIV Test. And i figured out that this is exactly what i need. > as the comments said, check if subscript A can possibly have the same value as B in analyseSIV(A,B) > but i didn't get How to use this information ? > > lets just say in the above program > When i use depends function it shows the dependency from load of x to

Gents, I spent some time reading over Sanjoy's patch for LoopDependenceAnalysis. Unfortunately, an early version of these notes escaped; this is the complete review. First off, I agree with his choice to implement the SIV tests. For scientific Fortran, the SIV (and the simpler ZIV) tests cover about 85% of the cases in practice. For C and C++, I expect the percentage will be much higher.

Shanmukha Rao wrote: > I am using LLVM for implementing LoopFission pass. > I am using LoopPass. > I know that for checking circular dependency in loop I have to use LoopDependenceAnalysis > > This is what i want to do. >         for(int i = 0; i< n ; i++){ > s1 : a[i] = a[i] + x[i]; > s2 : x[i] = x[i+1] + i*2 ; > } > >

On Thu, 15 Mar 2012 09:57:00 -0700 Preston Briggs <preston.briggs at gmail.com> wrote: > Shanmukha Rao wrote: > > I am using LLVM for implementing LoopFission pass. > > I am using LoopPass. > > I know that for checking circular dependency in loop I have to use > > LoopDependenceAnalysis > > > > This is what i want to do. > >         for(int i =

Hi Sanjoy, Reading through LoopDependenceAnalysis::analyseStrongSIV(), I noticed one problem and one confusion. My confusion related to your naming of the two instructions as A and B. It's consistent all through LoopDependenceAnalysis. I'd prefer something like source and destination, so I can keep track of which is which. It didn't matter so much when you were simply proving or

Hi, The DependenceAnalysis pass will find loop-carried dependences. However, it is a conservative analysis and will sometimes suggest there may be more dependences than actually exist. In your example, I expect the analysis is confused for some reason and is returning the default confused response. You could test it using the isConfused() method. Note that the DVEntry::ALL direction is always

On 02/07/2012 05:12 PM, David A. Greene wrote: > > Preston Briggs<preston.briggs at gmail.com> writes: > >> On Mon, Feb 6, 2012 at 12:13 PM, Sebastian Pop<spop at codeaurora.org> wrote: >>> [many things, but I'm only going to focus on one of them] >>> Would you consider using Polly http://polly.grosser.es to avoid >>> writing this code?

Hi Hal, Preston! Sorry for the delay! Got busy with some offline work. I've worked on my previous code to calculate direction and distance vectors whenever possible (strong SIV, basically). I think the current code is much clearer and would like your opinions on it. I have attached the patch and also pushed to the github repo I mentioned [1]. Thanks! [1]

Nadav Rotem <nrotem at apple.com> wrote: > I sent a patch to llvm-commit with a new loop vectorizer. > This is a very simple loop vectorizer, but we have to start somewhere. > With this new loop vectorizer we can already vectorize a good number of loops. > I know that we can improve the new loop vectorizer in a number of ways. > We can implement a precise dependence test, >

Hi, Here is a preliminary (monolithic) version you can comment on. This is still buggy, however, and I'll be testing for and fixing bugs over the next few days. I've used your version of the strong siv test. Thanks! -- Sanjoy Das. http://playingwithpointers.com -------------- next part -------------- A non-text attachment was scrubbed... Name: patch.diff Type: application/octet-stream

On Feb 6, 2012, at 1:56 PM, Hal Finkel wrote: >> If you do not want to use polly, you could use ISL >> http://freecode.com/projects/isl to set up the dependence problem and >> use ISL's ILP to solve it. > > isl is an LGPL project. It is not clear to me what the general consensus > would be on having a core analysis pass carry an LGPL dependency. This is fine for

Hi Nadav, Do you have any small write-up of current design of loop vectorizer?. If so, can you please send it across?. I want to see if there are dependencies such as unrolling for the vectorization. In the design we may also have to consider BB vectorizer and loop vectorizer working well together with no ambiguous requirements/dependencies. Regards, Shivaram -----Original Message-----

On 02/07/2012 07:04 PM, Preston Briggs wrote: >>> The 1st loop is trivially parallel. I think Polly would recognize >>> this and do good things. >> >> This case is trivial. >> >> But keep in mind that unsigned loop ivs and integers can cause modulo >> wrapping which is not trivial to handle - Both Polly, but also any other >> tool, need to

On 02/07/2012 01:17 AM, Hal Finkel wrote: > On Tue, 2012-02-07 at 00:47 +0100, Tobias Grosser wrote: >> On 02/06/2012 11:28 PM, Chris Lattner wrote: >>> On Feb 6, 2012, at 1:56 PM, Hal Finkel wrote: >>>>> If you do not want to use polly, you could use ISL >>>>> http://freecode.com/projects/isl to set up the dependence problem and >>>>>

On Tue, 2012-02-07 at 00:47 +0100, Tobias Grosser wrote: > On 02/06/2012 11:28 PM, Chris Lattner wrote: > > On Feb 6, 2012, at 1:56 PM, Hal Finkel wrote: > >>> If you do not want to use polly, you could use ISL > >>> http://freecode.com/projects/isl to set up the dependence problem and > >>> use ISL's ILP to solve it. > >> > >> isl