similar to: How can I tell llvm, that a branch is preferred ?

Hi, Recently 10% performance regression on an important benchmark showed up after we integrated https://reviews.llvm.org/rL318299. The analysis showed that rL318299 triggered loop rotation on an multi exits loop, and the loop rotation introduced code layout issue. The performance regression is a side-effect of rL318299. I got two testcases a.ll and b.ll attached to illustrate the problem. a.ll

On Mon, Dec 18, 2017 at 5:46 PM Xinliang David Li <davidxl at google.com> wrote: > The introduction of cleanup.cond block in b.ll without loop-rotation > already makes the layout worse than a.ll. > > > Without introducing cleanup.cond block, the layout out is > > entry->while.cond -> while.body->ret > > All the arrows are hot fall through edges which is

The Problem? Nowaday, SCEV called "Scalar Evolution" does only evolate instructions that has predictable operand, Constant-Based operand. such as that can evolute as a constant. otherwise we couldn't evolate it as SCEV node, evolated as SCEVUnknown. important thing that we remember is, we do not use SCEV only for Loop Deletion, which that doesn't really needed on nature loops

Currently the AtomicExpandPass will lower the following IR: define i1 @foo(i32* %obj, i32 %old, i32 %new) { entry: %v0 = cmpxchg weak volatile i32* %obj, i32 %old, i32 %new _*release acquire*_ %v1 = extractvalue { i32, i1 } %v0, 1 ret i1 %v1 } to the equivalent of the following on AArch64: _*ldxr w8, [x0]*_ cmp w8, w1 b.ne .LBB0_3 // BB#1:

Hi @ll, while clang/LLVM recognizes common bit-twiddling idioms/expressions like unsigned int rotate(unsigned int x, unsigned int n) { return (x << n) | (x >> (32 - n)); } and typically generates "rotate" machine instructions for this expression, it fails to recognize other also common bit-twiddling idioms/expressions. The standard IEEE CRC-32 for "big

"Sanjay Patel" <spatel at rotateright.com> wrote: > IIUC, you want to use x86-specific bit-hacks (sbb masking) in cases like > this: > unsigned int foo(unsigned int crc) { > if (crc & 0x80000000) > crc <<= 1, crc ^= 0xEDB88320; > else > crc <<= 1; > return crc; > } To document this for x86 too: rewrite the function

Hello. I am writing a back end in which I combined the existing BPF LLVM back end with the Mips MSA vector extensions (from the Mips back end) I have encountered an error when compiling with llc: the instruction selector uses a vector register instead of a scalar register with type i64 . I have the following part of LLVM IR program: vector.body.preheader:

Hi, I'm using LLVM to convert expressions to native assembly, the problem is when LLVM compiles this code: define void @fn_0000000000000000(i8*, i8*, i8*) { bb: %res = alloca i32 %3 = load i32* %res %4 = bitcast i8* %0 to i32* %5 = load i32* %4 %6 = bitcast i8* %0 to i32* %7 = load i32* %6 %8 = xor i32 %5, %7 store volatile i32 %8, i32* %res %9 = load i32* %res %10 = icmp

On Sun, May 11, 2014 at 8:19 AM, Stephan Tolksdorf <st at quanttec.com> wrote: > Hi, > > When clang/LLVM can't prove that a noexcept function only contains > non-throwing code, it seems to insert an explicit exception handler that > calls std::terminate. Why doesn't clang leave it to the eh personality > function to call std::terminate when an exception is thrown

Hi I have some inline function C code, that llvm could be optimizing better. Since I am new to this, I wonder if someone could give me a few pointers, how to approach this in LLVM. Should I try to change the IR code -somehow- to get the code generator to generate better code, or should I rather go to the code generator and try to add an optimization pass ? Thanks for any feedback. Ciao

Hello all, I've been working for the last month or so on a comprehensive mitigation approach to variant #1 of Spectre. There are a bunch of reasons why this is desirable: - Critical software that is unlikely to be easily hand-mitigated (or where the performance tradeoff isn't worth it) will have a compelling option. - It gives us a baseline on performance for hand-mitigation. - Combined

I suspect this is a straight forward problem so I thought I'd ask. I'm developing a new backend. I recently updated from the LLVM repository and now my output assembly is branching to labels/blocks that have been removed. It had been working fine two weeks ago. What looks suspicious is the following message: TryTailMergeBlocks: BB#1, BB#3, BB#4 Looking for common tails of

I am looking for guidance on how to: 1.

the clang 3.5 loop optimizer seems to jump in unintentional for simple loops the very simple example ---- const int SIZE = 3; int the_func(int* p_array) { int dummy = 0; #if defined(ITER) for(int* p = &p_array[0]; p < &p_array[SIZE]; ++p) dummy += *p; #else for(int i = 0; i < SIZE; ++i) dummy += p_array[i]; #endif return dummy; } int main(int argc, char** argv) {

Hi all, I am slowly working on a SwitchInst optimizer (http://llvm.org/PR8125) and now I am running into a deficiency of the x86 peephole optimizer (or jump-threader?). Here is what I get: andl $3, %edi je .LBB0_4 # BB#2: # %nz # in Loop: Header=BB0_1 Depth=1 cmpl $2, %edi

Hi Michael, Sometime back I did some experiments with interleave vectorizer and did not found any degrade, probably my tests/benchmarks are not extensive enough to cover much. Elina is the right person to comment on it as she already experienced cases where it hinders performance. For interleave vectorizer on X86 we do not have any specific costing, it goes to BasicTTI where the costing is not

Am 03.03.2015 um 19:49 schrieb Philip Reames <listmail at philipreames.com>: Hi Philip first thanks for your response, > You'll need to prove a bit more information to get any useful response. Questions: > 1) What's you're use case? Are you using clang to compile C code? Are you manually generating LLVM IR? yes the "inline function C code" will be compiled

It's not a solution to the actual bug (which is, as the thread you linked discusses, a problem with the assumption on LLVM's part that the __chkstk function lies within 2GB of the emitted code's address space) but there is a simple workaround: hoist all allocas to the first basic block of your function. This allows the JIT to perform all stack allocations in a single adjustment of the

Hi, We recently found a testcase showing that simplifications in instcombine sometimes change the instruction without reducing the instruction cost, but causing problems in TwoAddressInstruction pass. And it looks like the problem is generic and other simplification may have the same issue. I want to get some ideas about what is the best way to fix such kind of problem. The testcase:

Hi, I experienced a crash in code compiled with Clang 10.0.0 due to a misaligned 64-bit data access. The (ARMv8) CPU is configured with SCTL.A == 1 (alignment check enable). With SCTLR.A == 0 the code runs as expected. After some investigation I came up with the following reproducer: ---8<-------8<-------8<-------8<-------8<-------8<-------8<------- $ cat test.c extern char