similar to: postindexed load/store

Hi, I would like to transform a LLVM function containing a load and an add of the base address inside a loop to a post-incremented load. In DAGCombiner.cpp::CombineToPostIndexedLoadStore(), it says it cannot fold the add for instance if it is a predecessor/successor of the load. I find this odd, as this is exactly what I would like to handle: a simple loop with an address that is inremented in

On Jan 27, 2011, at 11:13 PM, Jonas Paulsson wrote: > Hi, > > I would like to transform a LLVM function containing a load and an add of the base address inside a loop to a post-incremented load. In DAGCombiner.cpp::CombineToPostIndexedLoadStore(), it says it cannot fold the add for instance if it is a predecessor/successor of the load. I find this odd, as this > is exactly what I

Hi, I'm trying to enable the pre-decrementing and post-incrementing based addressing modes for my target. So far without any success :( I enabled these modes in my TargetLowering class using setIndexedLoadAction and setIndexedStoreAction calls. I also defined getPreIndexedAddressParts and getPostIndexedAddressParts. And I can see that DAGCombiner::CombineToPostIndexedLoadStore is invoked.

Hi Hal, On Tuesday 01 October 2013 06:22 PM, Hal Finkel wrote: > ----- Original Message ----- >> Hi Hal, >> >> >> Our Architecture has indirect move instruction which increments >> the >> pointer implicitly and the target only has i8 type values. So the >> load >> of i16 will be converted to two i8 target loads wherein after the

Hal, (and everyone who might care about post increment generation)... I have an interesting question/observation. Consider this vector loop. void vec_add_const(unsigned N, short __attribute__ ((aligned (16))) *A, short __attribute__ ((aligned (16))) val) { unsigned i,j; for (i=0; i<N; i++) { for (j=0; j<N; j++) { A[i*N+j] += val; } } } The

----- Original Message ----- > From: "Sergei Larin" <slarin at codeaurora.org> > To: "Hal Finkel" <hfinkel at anl.gov> > Cc: llvmdev at cs.uiuc.edu > Sent: Friday, March 1, 2013 10:24:39 AM > Subject: Interesting post increment situation in DAG combiner > > Hal, (and everyone who might care about post increment generation)... Sergei, Perhaps

----- Original Message ----- > Hi Hal, > > > Our Architecture has indirect move instruction which increments > the > pointer implicitly and the target only has i8 type values. So the > load > of i16 will be converted to two i8 target loads wherein after the > first > load instruction, the pointer to the first i8 address will > automatically >

Hal, Here is my patch for the post inc case. I think it is symmetrically applicable to the pre-inc, but I have not tested it for that. I think you can clearly see my intent here - I simply select the "latest" candidate when multiple are available. Who else might be interested in this? Sergei --- Qualcomm Innovation Center, Inc. is a member of Code Aurora Forum, hosted by The

Hi all, My old and stable backup script don't work anymore. It's based on smbclient's tar command using crX parameters for creating archive, excluding files based on basic regular expressions. The box is running Fedora core 2, fully updated: kernel-2.6.8-1.521 samba-common-3.0.7-2.FC2 samba-client-3.0.7-2.FC2 The error is easily reproducible: smbclient //ntpdc/share -A credfile \ -c

Dear Emily, Using a look-behind solves the split problem in this case. (Note: Using Regex is in most/many cases the simplest solution.) str = c("leucocyten + gramnegatieve staven +++ grampositieve staven ++", "leucocyten ? grampositieve coccen +") tokens = strsplit(str, "(?<=[-+])\\s++", perl=TRUE) PROBLEM The current expression does NOT work for a different

Full_Name: Stefan Raberger Version: 2.8.1 OS: Windows XP Submission from: (NULL) (213.185.163.242) Hi there, I recently noticed some strange behaviour of the command "type.convert", depending on the startup mode used. But there also seems to be different behaviour on different PCs (all running the same OS and the same version of R). On PC1: When I start R in SDI mode (RGui --no-save

Hi all, I have a property, defined in an external stylesheet: .my_class { background-color: #FF0000; } I want to retrieve the value of this background-color of the class my_class. How can I do that with prototype/scriptaculous ? Thanks in advance, Nicolas Terray

Hi Chris, --- Chris Lattner <sabre at nondot.org> wrote: > On Wed, 29 Nov 2006, Roman Levenstein wrote: > > Thanks, this is a good idea. > > > > But I cannot figure out how to make a machine function pass run > > _BEFORE_ the RA. I guess I'm missing something very obvious. > > In your target's TargetMachine::addInstSelector method, add it to the

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);

Hi, attached is a patch by Chris Faylor <cgf at cygnus.com> relative to 2.2.0p1. Description: OpenSSH does not allow port gatewaying by default. This means that only the local host can access forwarded ports. Adding "GatewayPorts yes" to .ssh/config usually does this job. Unfortunately, OpenSSH does not recognize the same hosts.allow/ hosts.deny options as ssh.com's sshd

Windows usernames are case-insensitive and almost any Unicode character is allowed in a username. The user should be able to login with her username given in any case and not be refused. However, this opens up a security problem in terms of the sshd_config Match rules. The match rules for user and group names have to operate case-insensitive as well, otherwise the user can override her settings

Hi, I was wondering if there is an optimization pass in LLVM or Polly that can optimize this given loop: int main(int argc, char* argv[]) { char s = 0; // Loop is setting s = argv[1][0]; for (int i=0; i<argv[1][0]; i++) { s++; } printf("s = %c\n", s); return 0; }

I am trying to port the splancs library for S++ version 4.0 to R. I have modified the original FORTRAN to use dllwrap/dlltool (e.g. removed common blocks). I have been able to generate dll's (splancs.dll) using dllwrap/dlltool and Ming32 g77 with a -mdll switch. As of yet I have not been able to generate a dll that can be loaded by R. Does anyone out there have any suggestions. An example

Hi Hal, Our Architecture has indirect move instruction which increments the pointer implicitly and the target only has i8 type values. So the load of i16 will be converted to two i8 target loads wherein after the first load instruction, the pointer to the first i8 address will automatically increment to point to the next i8 value. So the post increment nature is in the Target. A normal

Hi all, these are the remaining unapplied patches of the ARM HDLCD patch series. Changes in v7: - rebased on b61ed421d2c85b5b106c63f2c14f8aa162b282f0; - turn more printk and panic into early_printk and early_panic. Changes in v6: - rebased on 77d3a1db3196b1b5864469f8d3f41d496800c795; - remove useless initializations to NULL in lfb_init; - more compact checks in lfb_init. Changes in v5: - move