similar to: [LLVMdev] Folding vector instructions

Alex wrote: > Hello. > > Sorry I am not sure this question should go to llvm or mesa3d-dev mailing > list, so I post it to both. > > I am writing a llvm backend for a modern graphics processor which has a ISA > very similar to that of Direct 3D. > > I am reading the code in Gallium-3D driver in a mesa3d branch, which > converts the shader programs (TGSI tokens) to

In the output of the code below, I want to do the following: - get hats over some of the betas - get the polygons stippled, not coloured grey - remove the tick marks at the ends of the axes. If I put tick = false, the whole axis disappears. betahat <- c(0.04*0:150) betahatdens <- dnorm(betahat, 3, 1) plot(betahat, betahatdens, xlim = c(-0.2, 6.2),

I am running this code in JIT on x86 (32 bit). It crashes when 'alloca %object' instruction is within the body of the cycle, and it finishes successfully when this instruction is in the beginning of main (outside the cycle). Crash occurs in the middle of the cycle after few hundreds of thousands iterations. What is wrong? Does alloca inside the cycle forces it to allocate it on stack

When I compile my C fibonacci example fib.c with 'clang -O3 -c -emit-llvm -o fib-clang.bc fib.c&& llvm-dis fib-clang.bc' I get fib-clang.ll that has some degree of inlining in it. But when I get an equivalent to fib.c file fib.ll and run it through opt with the command 'llvm-as fib.ll&& opt -O3 fib.bc -o fib-opt.bc&& llvm-dis fib-opt.bc' resulting

The BB vectorizer has an option 'bb-vectorizer-search-limit'. Is there a similar option for the SLP vectorizer? Maybe an analysis pass' scope that can be widen? I have large basic blocks with instructions that should be merged into packed versions. However, the blocks are optimized independently from each other. Now, if the instructions to be merged aren't too far apart the

Hi, It looks like SimplifyLibCalls has a tendency to emit calls to libm functions without checking with TLI whether these calls are available. For example, PowOpt has this code: struct PowOpt : public UnsafeFPLibCallOptimization { PowOpt(bool UnsafeFPShrink) : UnsafeFPLibCallOptimization(UnsafeFPShrink) {} virtual Value *callOptimizer(Function *Callee, CallInst *CI, IRBuilder<> &B)

On Tue, Aug 13, 2013 at 5:58 AM, Kuperstein, Michael M < michael.m.kuperstein at intel.com> wrote: > Hi,**** > > ** ** > > It looks like SimplifyLibCalls has a tendency to emit calls to libm > functions without checking with TLI whether these calls are available.**** > > For example, PowOpt has this code:**** > > ** ** > > struct PowOpt : public

Hey, as some of you may know we're in process of experimenting with LLVM in Gallium3D (Mesa's new driver model), where LLVM would be used both in the software only (by just JIT executing shaders) and hardware (drivers will implement LLVM code-generators) cases. While the software only case is pretty straight forward I just realized I missed something in my initial evaluation. That

> > which goes through memory. Is there some idiom I'm missing so that it would use > > for instance movmsk for SSE or vcmpgt & cr6 for altivec? > > I don't think you are missing anything: LLVM IR has no support for horizontal > operations like or'ing the elements of a vector of boolean together. The code > generators do try to recognize a few idioms and

I agree with John; also, if you decide to go this route, you can reuse the code from CodeGenPrepare::optimizeSelectInst: https://github.com/llvm/llvm-project/blob/master/llvm/lib/CodeGen/CodeGenPrepare.cpp#L6065 Alexey On Thu, Jan 30, 2020 at 9:00 PM John Regehr via llvm-dev < llvm-dev at lists.llvm.org> wrote: > Several different passes introduce select instructions, such as >

Hi, I would like to know if there's a way to avoid select instructions during the IR generation. What are the optimization passes that can result in a select instruction? i.e. I want to preserve branches in my code without disabling any other optimizations applicable. For example, void foo(int* x, int* y){ if(*x > 0){ *y = *x + 10; } else{ *y = *x + 20; } }

I used &Idx[0]. In future, please avoid tabs in your patch. I applied your patch. http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20071217/056403.html - Devang On Dec 17, 2007, at 2:57 AM, Richard Pennington wrote: > Devang Patel wrote: >> On Dec 15, 2007, at 12:15 PM, Richard Pennington wrote: >>> I got the current version of LLVM via svn yesterday and

Devang Patel wrote: > On Dec 15, 2007, at 12:15 PM, Richard Pennington wrote: > >> I got the current version of LLVM via svn yesterday and modified my >> code to >> use the LLVMFoldingBuilder. Very nice! >> >> My question is this: I noticed that the folding builder doesn't fold >> some >> operations, e.g. casts. Is there some reason why? If

I made a list of /usr/bin scripts which allows /tmp races. Following ones creates /tmp/something.$$, then, with no permission/ownership checking, /tmp/something.$$.x (x may vary ;), or even performs suitable checks, but gives enough time to alter /tmp contents: glibcbug, bashbug, znew, mailstat, autoupdate, x11perfcomp, gccmakedep, pnmindex, xcopy, autoheader, cvsbug, rcs2log, updatedb, igawk,

Hi Stephen, > Hi all, llvm newbie here. welcome! > I'm trying to branch based on a vector compare. I've found a slow way (below) > which goes through memory. Is there some idiom I'm missing so that it would use > for instance movmsk for SSE or vcmpgt & cr6 for altivec? I don't think you are missing anything: LLVM IR has no support for horizontal operations like

Hi all, llvm newbie here. I'm trying to branch based on a vector compare. I've found a slow way (below) which goes through memory. Is there some idiom I'm missing so that it would use for instance movmsk for SSE or vcmpgt & cr6 for altivec? Or do I need to resort to calling the intrinsic directly? Thanks, Stephen. %16 = fcmp ogt <4 x float> %15, %cr %17 =

Am 04.09.2012 00:08, schrieb Stephen: >>> which goes through memory. Is there some idiom I'm missing so that it would > use >>> for instance movmsk for SSE or vcmpgt & cr6 for altivec? >> >> I don't think you are missing anything: LLVM IR has no support for horizontal >> operations like or'ing the elements of a vector of boolean together.

Hello all, I tried unsuccessfully to run libguestfs on debian squeeze (with kernel from testing) using the packages you provide here: http://libguestfs.org/download/binaries/debian-packages/. The output I get from libguestfs-test tool is this: ===== Test starts here ===== LIBGUESTFS_DEBUG=1 TMPDIR=(not set) libguestfs: new guestfs handle 0x17f32c0 library version: 1.14.0 guestfs_get_append:

Hello, I had been using guestmount to mount a raw image and it has been working great until I updated the kernel on 01/10/2018. I am running ubuntu 16.04 LTS, and have changed the read permissions to the vmlinuz as I have done with previous kernels. The only change to my system between working and not working is the kernel update from this morning. I am concerned this might be related to the

Hi， Rich， follow your advice， I have succeeded in compiling the “libguestfs-1.26.5” my ”libguestfs-test-tool is passed “ “ libguestfs: command: run: rm libguestfs: command: run: \ -rf /tmp/libguestfs7jxKy0 ===== TEST FINISHED OK =====” But， When I run “virt-df `virsh domuuid 147`” the error show：（debug mode) “supermin helper [00000ms] host_cpu = x86_64 supermin helper [00000ms]