similar to: [LLVMdev] A faster instruction selector?

Hi Nicolas and Dan, Thanks for your replies. I've been playing around with various setting, as you suggested. > What version of LLVM are you using here? I'm using 2.4 My original time ratios of reg-alloc to instruction selection (1:12) referred to the local register allocator and the standard instruction selector (all passes), not a sensible comparison, I realise. > I did

Hi everyone, llvm is great! But there is one exception ;) llvm components are generally fast, but instruction selection is slooow. Let me explain. I am developing a toolkit for building virtual machines which can automatically generate a JIT compiler using the interpreter specification. llvm does the hard work of machine code generation. (Thanks to you all) I discovered that JIT compilation is

As you may have noticed, I have been working on a fast register allocator in the last week. The new allocator is intended to replace the local allocator for debug builds. Both allocators use a local allocation strategy. Each basic block is scanned from top to bottom, and virtual registers are assigned to physical registers as they appear. There are no live registers between blocks. Everything is

Welcome to the world of pluggable machine passes. This work was done to lighten the load of some llvm tools, ie., only link in the register allocators and instruction schedulers that are actually needed in a particular circumstance. I guess I will have to write this up, but generally it works like this. In your register allocator .cpp file add the following include; #include

Hello , I have written a very simple MachineFunction Pass that currently does nothing. It compiles fine but when I try to load it with llc it give me following error: llc -optimize-regalloc=0 -load lib/GCRA.dylib -regalloc=gc test/fibo.bc Pass 'Bundle Machine CFG Edges' is not initialized. Verify if there is a pass dependency cycle. What is going wrong here ? Here is my very simple

summary: I can display a lon-lat map on a lattice::layerplot, and I can display a Lambert conformal conic (LCC) map on a spam::image, but I can't display an LCC map on a lattice::layerplot. Example follows. What am I doing wrong? details: I've been using `lattice` (via `rasterVis`) successfully to display global atmospheric data, which works well enough (though I am definitely intrigued

I am running 1.3.1 on a Windows (NT 4.0) machine. I've fit a nonlinear model intended to predict crop yield from nutrient information, and want to use the profile function. If I type say, profile(simparj.fm) I get the following error message: "Error in prof$getProfile(): number of iterations exceeded maximum of 5.25515e-308" I used the profiler function to profile simparj,fm step

I'm trying to get the Syncrosoft License Control Center to run under wine. for some reason it can only load MFC42.DLL and MSVCRT.DLL if I put 'sudo' on the command line: [paul@localhost LCC]$ wine LCC.exe err:module:import_dll Library MFC42.DLL (which is needed by L"C:\\Program Files\\Syncrosoft\\LCC\\LCC.exe") not found err:module:import_dll Library MSVCRT.dll (which is

I've read something about this problem, but I don't know how can i avoid this problem. Why the order of the factors give different results? I suppose it's because the order of the factors, i've just changed "lcc" from the first position to the last in the model, and the significance change completely >

I've run an ancova, edadysexo is a factor with 3 levels,and log(lcc) is the covariate (continous variable) I get this results > ancova<-aov(log(peso)~edadysexo*log(lcc)) > summary(ancova) Df Sum Sq Mean Sq F value Pr(>F) edadysexo 2 31.859 15.9294 803.9843 <2e-16 *** log(lcc) 1 11.389 11.3887 574.8081 <2e-16 ***

In the following sequence: $llc -march=c -o diag01_c.c diag01.ll $vpolcc.ss -S -k diag01_c.c <<== **FAILS** $vpolcc.ss -o diag01 diag01_c.s I am using llc (from llvm) to convert an .ll file into a .c file. I am then running the .c file through vpolcc.ss (from zephyr) to further process it for mips architecture. The vpolcc.ss script fails at lcc (not to be confused with llc), with

Dear Anton, you can add your register allocator strait iin the "lib/CodeGen/Passes.cpp", and then 're-make' it: "makellvm llc", on the top of lib/CodeGen. It is faster than running make from LLVMOBJDIR. The problem is that it only add to llc the changes on the lib/CodeGen directory. If you change other parts, a make from LLVMOBJDIR will synchronize it. Try adding

Hello, I'm new to OGG and I must say than it's impressive. I plan to use it in my project (BASIC programming langage) but I use LccWin32 as main C compiler (free and very powerful) and OGG has some difficulties to compile. I would like to know if it was already reported. If not, here it is: - Inline ASM functions for vorbis_itoa() isn't ok for LCC (not the same syntax than VC). I

For those interested in experimenting with (compiling / developing) a version of R for the Common Language Runtime (CLR) environment (Microsoft .Net, Novell Ximan Mono & DotGNU) -- a few links to some "free" compliers: C DotGNU (the official GNU project) http://dotgnu.org/ http://www.southern-storm.com.au/pnet_faq.html#q1_7 FAQ "1.7. What is pnetC? Since version 0.4.4 of

On Tuesday 04 May 2010 05:45:36 Josef Eisl wrote: > >> - As far as I understand it, register allocators are implemented as > >> MachineFunctionPasses. Does a MachineFunction object contain all > >> information needed for a (classic) allocator? > > > > It has the instructions, operands and dependencies among them. There's > > a

Hi all, Just a notice for out-of-tree maintainer, I removed RegisterScheduler::setDefault in r243388. It was used to override the default command line option that selects the DAG scheduler to use. Instead of hacking into the command line plumbing, a target that wants more control over the selection of the scheduler than it already has with setSchedulingPreference has a SubTargetHook

Changes since last time: - split auto-generated code in to c and h files - un-break the build system - fix ocaml binding due to libxl API change - lot''s of tidy-ups too numerous to mention Please consider and apply :) -----8<--------------------------------------------------------------- Introduce python binding for libxl. The binding is not yet complete but list_domains,

I managed to link my RegAlloc.a library to llc tool but can;t make the same with lli tool. Should I use USEDLIBS variable in lli makefile (like llc) or there is another way? On 8/20/06, Fernando Magno Quintao Pereira <fernando at cs.ucla.edu> wrote: > > > Dear Anton, > > you can add your register allocator strait iin the > "lib/CodeGen/Passes.cpp", and then

On Wednesday 11 July 2007 15:07, Christopher Lamb wrote: > Could it be possible for there to be a harness type interface that > would allow coalescers that support both modes to be hooked into the > pass registration, and those that depend on the allocator not be > registered as passes? I have a patch for this kind of thing attached. Please take a look and let me know if it looks

SUMMARY: Specific problem: I'm regridding biomass-burning emissions from a global/unprojected inventory to a regional projection (LCC over North America). I need to have boundaries for Canada, Mexico, and US (including US states), but also Caribbean and Atlantic nations (notably the Bahamas). I would also like to add Canadian provinces and Mexican states. How to put these together? General