similar to: [LLVMdev] How to visualise Clang optimisation phases

Having a look at clang's source, you can find in "lib/CodeGen/BackendUtil.cpp" the functions where clang builds the passes to emit code. The optimization passes used are there and you can simulate them via the "opt" utility, by running each pass one at a time. LLVM also declares standard module/function passes on include/llvm/Support/StandardPasses.h. Have a look and see

> Having a look at clang's source, you can find in > "lib/CodeGen/BackendUtil.cpp" the functions where clang builds the > passes to emit code. The optimization passes used are there and you can > simulate them via the "opt" utility, by running each pass one at a time. > LLVM also declares standard module/function passes on >

Hi, I’ve recently encountered an issue where the `instcombine` pass replaces an `llvm.memcpy` between two distinct address spaces with an `addrspacecast` instruction. As an example, see the trivial OpenCL kernel attached. I’m compiling like this: clang -cc1 -triple spir64-unknown-unknown -x cl -O0 -emit-llvm array_init.cl -o before.ll This yields an `llvm.memcpy` to copy the array

On Wed, Dec 2, 2015 at 10:39 AM, David Chisnall <David.Chisnall at cl.cam.ac.uk > wrote: > On 2 Dec 2015, at 18:30, Phil Tomson <phil.a.tomson at gmail.com> wrote: > > > >> If you want to pass LLVM arguments to clang, then you must prefix them > with -llvm (e.g. -mllvm -loops -mllvm lcssa). > > > > > > I just tried this: > > CFLAGS +=

Hi Mehdi, Thanks for the reply - I’ve finally got round to trying to fix this based on your suggestion. I’ve got something that mostly works, but I just wanted to double-check something about the regression tests before I post a patch. > The memcpy is supposed to be equivalent to a sequence of load and store. Here we are just failing to keep the property that the load is performed through

Hi Sanjoy, Yes these are C tests (from 'gcc.c-torture/execute'), and as I indicated in my original message, the tests are not valid because the behaviour is undefined. However, it was while investigating these new failures in these tests that I realised that this optimisation existed. The optimisation itself is perfectly valid, but it does mean that integer underflow will no longer be

Hello, I would like to build samba at different optimisation levels with debug symbols, in order to test a static analysis tool I have developed. I have found the configure option "--enable-developer", which I presume enables debug symbols and sets optimisation to O0, however I would also like to be able to enable debug symbols and set the optimisation level to O2. I cannot find a way

Dear All, I would like to minimise a nonlinear function subject to linear inequality constraints as part of an R program. I have been using the constrOptim function. I have tried all of the methods that come with Optim, but nothing finds the correct solution. If I use the correct solution as the vector of starting values, though, my program does output the correct solution and optimum - the

With LLVM v5.0, I found failures in some of the 'gcc.c-torture/execute' tests due to a change in the optimisation where undefined behaviour is involved. The tests that fail are the '20040409-[123].c' tests. The underlying failure is due to the optimisation of the following: int test2(int x) { return x + INT_MIN; } from using an ADD instruction to using an OR instruction.

Hi All, I was wondering if there is an optimisation pass that moves a stack allocated array, initialised with constant values, to a global constant array. And if there is such a pass, what requirements are there for it to operate? My optimised IR is below. As you can see an array of 5 integers is created with alloca, then each element is stored to in turn. It would be nice if this array was

I am often asked what optimisations "our" compiler performs at each level. But "our" compiler is actually CLang/LLVM which we have retargeted to our proprietary target. Most of the work we do is in maintaining our target specific backend. Certainly there are optimisations that we do to take best advantage of our instruction set during lowering and instruction selection, and

Hi there, I have a piece of Matlab code I use to optimise a trding strategy. If there are any Matlab/R specialists out there, I would appreciate your help in doing the exact same optimisation in R. I suspect I would use nlm() in R but am not sure where to define my constraints. I have attached my Matlab code below for reference. Many thanks. Constraints function [c,ceq]=TriskellConstraints(X)

OK, I’ve hit one more existing regression test that I’m weary of: define void @test2_addrspacecast() { %A = alloca %T %B = alloca %T %a = addrspacecast %T* %A to i8 addrspace(1)* %b = addrspacecast %T* %B to i8 addrspace(1)* call void @llvm.memcpy.p1i8.p0i8.i64(i8 addrspace(1)* %a, i8* bitcast (%T* @G to i8*), i64 124, i32 4, i1 false) call void

I noticed that there is a '-run-pass' argument for llc. I am wondering if I can do a similar approach with machine level optimisations/passes for llc. Are those passes optional (so I can turn them off)? And how can I get MIR format as llc expects with '-run-pass'? Thanks a lot. Cheers, Yi On 22/9/17 15:10, Craig Topper wrote: > Have -O0 on your clang command line causes

I'm fairly new to LLVM, and am writing an optimisation pass to be run by 'opt'. However, i'd like this pass to be parameterisable; as such i'd like to be able to pass an option to it (preferably via the command line to 'opt'). e.g. so i can do something like this "opt -load mypass.so -MyPass=MyPassOption < inputfile.bc ..." (for example one thing I would

https://bugzilla.samba.org/show_bug.cgi?id=10380 Summary: Non-Nested Folder Optimisation Product: rsync Version: 3.1.0 Platform: All OS/Version: All Status: NEW Severity: enhancement Priority: P5 Component: core AssignedTo: wayned at samba.org ReportedBy: me at haravikk.com

Dear All I have a constrained optimisation problem, I want to maximise the following function t(weights) %*% CovarianceMatrix %*% weights for the weights, subject to constraints on each element within the weights & the weights vector summing to 1. i.e. weights = (x1, x2, x3), where x1 is within some given range (a +b, a - b). I have tried to do this using the optim function in R,

Yes, the hairy-edges of undefined behaviour - UB is UB. It does mean that given: __attribute__((noinline)) int foo(int a, int b) { return a + b; } int bar1(int x) { return foo(x, INT_MIN); } int bar2(int x) { return x + INT_MIN; } 'bar1' and 'bar2' have different outcomes. However, I think that the new optimisation is neat and valid and I am not suggesting that it should

Right but then you'll have to call each opt level. Have u looked in the pass manager? On Jun 24, 2017 3:52 PM, "Sean Silva via llvm-dev" <llvm-dev at lists.llvm.org> wrote: > The most definitive list you can probably hope to get will be obtained by > passing -mllvm -debug-pass=Structure to a clang invocation. > > -- Sean Silva > > On Jun 24, 2017 7:10 AM,

Can optimisation (simplex etc) be done through R?