similar to: [LLVMdev] Intrinsics and dead instruction/code elimination

On May 19, 2010, at 7:07 AM, o.j.sivart at gmail.com wrote: > Hi all, > > I'm interested in the impact of representing code via intrinsic functions, in contrast to via an instruction, when it comes to performing dead instruction/code elimination. As a concrete example, lets consider the simple case of the llvm.*.with.overflow.* intrinsics. > > If I have some sequence (> 1)

On 20/05/2010, at 3:01 AM, Chris Lattner wrote: > > On May 19, 2010, at 7:07 AM, o.j.sivart at gmail.com wrote: > >> Hi all, >> >> I'm interested in the impact of representing code via intrinsic functions, in contrast to via an instruction, when it comes to performing dead instruction/code elimination. As a concrete example, lets consider the simple case of the

On May 19, 2010, at 3:13 PM, o.j.sivart at gmail.com wrote: >> >> Intrinsics should be optimized as well as instructions. In this specific case, these intrinsics should be marked readnone, which means that load/store optimization will ignore them. Dead code elimination will delete the intrinsic if it is dead etc. > > I understand that dead code elimination is able to delete

On 20/05/2010, at 8:16 AM, Chris Lattner wrote: > > On May 19, 2010, at 3:13 PM, o.j.sivart at gmail.com wrote: > >>> >>> Intrinsics should be optimized as well as instructions. In this specific case, these intrinsics should be marked readnone, which means that load/store optimization will ignore them. Dead code elimination will delete the intrinsic if it is dead

Hi All, I've been successfully using the with function for analyses and the transform function for multiple transformations. Then I thought, why not use "with" for both? I ran into problems & couldn't figure them out from help files or books. So I created a simplified version of what I'm doing: rm( list=ls() ) x1<-c(1,3,3) x2<-c(3,2,1) x3<-c(2,5,2)

I am having the following error in my function function(theta,reqdIRR) { theta1<-theta[1] theta2<-theta[2] n<-length(reqdIRR) constant<- n*(theta1+theta2) sum1<-lapply(reqdIRR*exp(theta1),FUN = sum) sum2<-lapply(exp(theta2 - reqdIRR*exp(theta1)),FUN = sum) sum = sum1 + sum2 log.fcn = constant - as.numeric(sum) result = - log.fcn return(result) } *error :

Greetings. For R gurus this may be a no brainer, but I could not find pointers to efficient computation of this beast in past help files. Background - I wish to implement a Cramer-von Mises type test statistic which involves double sums of max(X_i,Y_j) where X and Y are vectors of differing length. I am currently using ifelse pointwise in a vector, but have a nagging suspicion that there is a

The attached patch eliminates some warnings while compiling for ARM4 targets. It also simplifies the asm constraints a bit. Now we can use the ARM4 optimisations when compiling for PortalPlayer targets in Rockbox. Cheers, Dan -------------- next part -------------- A non-text attachment was scrubbed... Name: speex_arm4.patch Type: text/x-diff Size: 1550 bytes Desc: not available Url :

The following series adds initial vector support for PowerPC64. On POWER9, flac --best is about 3.3x faster. Amitay Isaacs (2): Add m4 macro to check for C __attribute__ features Check if compiler supports target attribute on ppc64 Anton Blanchard (5): configure.ac: Remove SPE detection code configure.ac: Add VSX enable/disable configure.ac: Fix FLAC__CPU_PPC on little endian, and add

Hello! # I have 3 vectors of values: values1<-rnorm(10) values2<-rnorm(10) values3<-rnorm(10) # In real life, all 3 vectors have a length of 25 # I create all possible combinations of 4 based on 10 elements: mycombos<-expand.grid(1:10,1:10,1:10,1:10) dim(mycombos) # Removing rows that contain pairs of identical values in any 2 of these columns: mycombos<-mycombos[!(mycombos$Var1

Jean-Marc, At least VS2005 (what I'm using) won't compile resample_sse.h with _USE_SSE2 defined because it refuses to cast __m128 to __m128d and vice versa. While there are intrinsics to do the casts, I thought it would be simpler to just use an intrinsic that accomplishes the same thing without all the casting. Thanks, --John @@ -91,7 +91,7 @@ static inline double

Maybe security related mails should be sent elsewhere? I didn't notice any so here it goes: sender.c:receive_sums() s->count = read_int(f); .. s->sums = (struct sum_buf *)malloc(sizeof(s->sums[0])*s->count); if (!s->sums) out_of_memory("receive_sums"); for (i=0; i < (int) s->count;i++) { s->sums[i].sum1 = read_int(f);

Hello :) The attached patch (which is not in any way finished) optimizes the resampler. (For those following the discussions on IRC; this version includes optimizations for both direct and interpolate cases). Using GCC 4.3, x86_64, Valgrind to measure instruction counts, resampling 10 frames of 320 floats at quality 3. Direct was measured with a 16=>48 resampling, and interpolate with a

Mark II of the patch set. The first patch (dynsumlen2.patch) increments the protocol version to support per-file dynamic block checksum sizes. It is a prerequisite for varsumlen2.patch. varsumlen2.patch implements per-file dynamic block and checksum sizes. The current block size calculation only applies to files between 7MB and 160MB setting the block size to 1/10,0000 of the file length for a

Hi, I run this code to get the power of the test for modified bartlett's test..but I'm not really sure that my coding is right.. #normal distribution unequal variance asim<-5000 pv<-rep(NA,asim) for(i in 1:asim) {print(i) set.seed(i) n1<-20 n2<-20 n3<-20 mu<-0 sd1<-sqrt(25) sd2<-sqrt(50) sd3<-sqrt(100) g1<-rnorm(n1,mu,sd1) g2<-rnorm(n2,mu,sd2)

The attached patch cleans up Opus's x86 intrinsics configury. It: * Makes ?enable-intrinsics work with clang and other non-GCC compilers * Enables RTCD for the floating-point-mode SSE code in Celt. * Disables use of RTCD in cases where the compiler targets an instruction set by default. * Enables the SSE4.1 Silk optimizations that apply to the common parts of Silk when Opus is built in

From: Jonathan Lennox <jonathan at vidyo.com> * Makes ?enable-intrinsics work with clang and other non-GCC compilers * Enables RTCD for the floating-point-mode SSE code in Celt. * Disables use of RTCD in cases where the compiler targets an instruction set by default. * Enables the SSE4.1 Silk optimizations that apply to the common parts of Silk when Opus is built in floating-point mode, not

From: Jonathan Lennox <jonathan at vidyo.com> * Makes ?enable-intrinsics work with clang and other non-GCC compilers * Enables RTCD for the floating-point-mode SSE code in Celt. * Disables use of RTCD in cases where the compiler targets an instruction set by default. * Enables the SSE4.1 Silk optimizations that apply to the common parts of Silk when Opus is built in floating-point mode, not

From: Thorvald Natvig <slicer at users.sourceforge.net> Some optimizing compilers miscompile the current SSE optimizations when full optimizations are enabled. By using output value pointer instead of a return value, we can bypass this misbehaviour. --- libspeex/resample.c | 8 ++++---- libspeex/resample_sse.h | 24 ++++++++---------------- 2 files changed, 12 insertions(+), 20

Hi, Is there a sure way to test 2 cd's if they are true copies in rsync. I have tried this... rsync -avv /mnt/cdrom/ /mnt/cdrom2 ** result....quite shotened... misc/rpm2header is uptodate pkg-9.2-FiveStar-download-i586.idx is uptodate total: matches=0 tag_hits=0 false_alarms=0 data=0 wrote 65301 bytes read 20 bytes 1789.62 bytes/sec total size is 679261427 speedup is 10398.82 ** Is