similar to: [SPIR/PTX] Divergence analysis for BasicBlocks

Hello, Yes? Where is allActive defined, I couldn't find it. Basically, a BB is control divergent if it's execution depends on a branch that itself depends on a divergent ssa value. On Fri, Jul 21, 2017 at 4:13 PM, Zaks, Ayal <ayal.zaks at intel.com> wrote: > What would be the definition of “isControlDivergent(BasicBlock*)”; the > complementary of “allActive(BasicBlock*)” –

TL;DR This RFC is a joint effort by Intel and Saarland University to bring the divergence analysis of the Region Vectorizer [1,2,3,4,5] (dubbed the vectorization analysis of RV) to LLVM. The implementation is available on github for feedback [0]. The existing divergence analysis infrastructure in LLVM has conceptual limitations (structured control, SCEV based). The new analysis resolves bugs

On 25.10.2016 16:28, Nicolai Hähnle wrote: > But I fear that this path leads to eternal fuzziness. Let me try a > completely different approach to define what we need by augmenting the > semantics of IR with "divergence tokens". In addition to its usual > value, every IR value carries a "divergence set" of divergence tokens. > > The basic rule is: the

> On Oct 24, 2016, at 4:15 PM, Nicolai Hähnle <nhaehnle at gmail.com> wrote: > > On 25.10.2016 01:11, Nicolai Hähnle wrote: >> On 24.10.2016 21:54, Mehdi Amini wrote: >>>> On Oct 24, 2016, at 12:38 PM, Nicolai Hähnle via llvm-dev >>>> <llvm-dev at lists.llvm.org> wrote: >>>> Some brain-storming on an issue with SPMD/SIMT backend

I tried to reduce the test case in https://bugs.llvm.org/show_bug.cgi?id=42706. Here it is crashing opt: $ ~/llvm-debug/bin/opt -use-gpu-divergence-analysis -divergence stripped.ll WARNING: You're attempting to print out a bitcode file. This is inadvisable as it may cause display problems. If you REALLY want to taste LLVM bitcode first-hand, you can force output with the `-f' option.

And again... LLVM_BUILD_LLVM_DYLIB:BOOL=ON LLVM_LINK_LLVM_DYLIB:BOOL=ON This one is the good one... maybe. On Tue, Sep 6, 2016 at 11:35 PM, Alexandre Isoard < alexandre.isoard at gmail.com> wrote: > That is because I mistyped it: > LLVM_ENABLE_LLVM_DYLIB:BOOL=ON > LLVM_LINK_LLVM_DYLIB:BOOL=ON > > On Tue, Sep 6, 2016 at 11:31 PM, Wink Saville <wink at saville.com>

Hello, I assumed SCEV purpose was to be a normal form, but then I wondered which one of those is the normal form: (zext i16 (trunc i32 %a to i16) to i32) vs (-((%a /u 65536) *u 65536) + %a) The first one is nice for interval analysis, and known bit analysis. The second one is nice when plugged into gep of 2d arrays. -- *Alexandre Isoard* -------------- next part -------------- An HTML

On 07/05/2017 03:10 PM, Alexandre Isoard wrote: > Ah, ok. I read it wrong. In *neither* case it is UB. > > Hum, can an implementation define it as UB? :-) Nope :-) The only case I've thought of where we could add these for C++ would be on conversions to (most) enums (because they used signed underlying types and the out-of-bounds mapping won't generally be one of the allowed

On Tue, Sep 6, 2016 at 3:05 PM, Alexandre Isoard <alexandre.isoard at gmail.com> wrote: > LLVM_ENABLE_DYLIB Where/when/how do you specify LLVM_ENABLE_DYLIB and LLVM_LINK_DYLIB? I tried the following on the cmake command line: $ cmake -G Ninja .. -DCMAKE_INSTALL_PREFIX=/home/wink/opt/llvm -DLLVM_ENABLE_DYLIB=true -DLLVM_LINK_DYLIB=true And got: ... -- Performing Test

Hi, Even if there are no ways in which a *frontend* can produce nsw truncs, it may still be useful to have if optimization passes can usefully attach nsw to truncates (after proving the truncates don't "overflow"). For instance in %a = ashr i64 %v, i32 33 %t = trunc %a to i32 the trunc can be marked nsw. However, the burden of proof here is to show that we can do some useful

On Wed, Jul 5, 2017 at 3:59 PM, Hal Finkel via llvm-dev < llvm-dev at lists.llvm.org> wrote: > > On 07/04/2017 01:41 AM, Dr.-Ing. Christoph Cullmann via llvm-dev wrote: > >> Hi, >> >> Hi Alexandre, >>> >>> LLVM currently doesn't have trunc nsw/nuw, no. >>> Which frontend would emit such instructions? Any application in mind?

According to 6.3.1.3/3 of the C standard (I didn't check C++): "3 Otherwise, the new type is signed and the value cannot be represented in it; either the result is implementation-defined or an implementation-defined signal is raised." I *think* that means that IF a signal is raised then the signal raised could be one that you can't guarantee to be able to return from

Note that there is a slight difficulty due to the fact that we "sink" the trunc: (zext i16 {0,+,1}<%bb> to i32) + (65536 * ({0,+,1}<nuw><%bb> /u 65536) Here the recurrence lost it's <nuw> and got reduced to a i16 (on the left), but not on the right. But we can prove: - that (zext i16 {0,+,1}<%bb> to i32) has the same 16 LSB than (i32

Ok. To go back to the original issue, would it be meaningful to add a SCEVUMax(0, BTC) on the final BTC computed by SCEV? So that it does not use "negative values"? On Wed, Aug 15, 2018 at 2:40 PM Friedman, Eli <efriedma at codeaurora.org> wrote: > On 8/15/2018 2:27 PM, Alexandre Isoard wrote: > > I'm not sure I understand the poison/undef/UB distinctions. >

I did not observe any assertion. In addition, the documentation ( https://llvm.org/docs/LangRef.html#undefined-values) says: The string ‘undef’ can be used anywhere a constant is expected, and indicates that the user of the value may receive an unspecified bit-pattern. Undefined values may be of any type (other than ‘label’ or ‘ void’) and be used anywhere a constant is permitted. Either way,

On Wed, Aug 24, 2016 at 10:11 AM, Renato Golin <renato.golin at linaro.org> wrote: > On 24 August 2016 at 08:48, Alexandre Isoard via llvm-dev > <llvm-dev at lists.llvm.org> wrote: > > I am probably stating the obvious, but if the function is side-effect > free > > (onlyReadsMemory) it is valid to remove it. > > > > But I am guessing that does not

Yes, it's still respected in this case, as the only instructions that will be deleted have been RAUW with undef. Originally, all instructions where RAUW but only non-EHPad were deleted (that means EHPad were RAUW but not deleted). Then it was later patched by not RAUW token instructions and now not deleting EHPad nor token instructions. My assumption is that the instructions we wanted to

Hi Owen and Vinod, Thanks for sharing the paper! I like the idea a lot. Regarding the paper itself, Vinod, are the consensual branches (e.g., cbranch.ifnone) you mentioned in the paper publicly available in PTX ISA? Owen, could you explain more on the approach of using branch-if-none instructions in your mind? I believe you have lots of great insights, but I don't see how cbranch.ifnone

Hello, I was thinking about printing SCEV into DiagnosticInfo messages, an example would be to print the loop trip count of loops, or the stride of memory accesses. I ran into two problems: - DiagnosticInfo is in Core, SCEV is in Analysis, so it is a little bit weird (I declare the operator<< overload for SCEV* in DiagnosticInfo, but only define it in ScalarEvolution) - I would like to

On Fri, Nov 3, 2017 at 5:39 PM, Hal Finkel <hfinkel at anl.gov> wrote: > > On 11/03/2017 07:20 PM, Alexandre Isoard via llvm-dev wrote: > > Hello, > > I am not sure about the semantic (if any) of returns_twice and noreturn > attributes. > > int fork() __attribute__((returns_twice)); > void join(int) __attribute__((noreturn)); > > int f(int n) { > int