similar to: [LLVMdev] Predication on SIMD architectures and LLVM

We are currently doing something similar to your third option in Hexagon backend. But it is a VLIW so predication is not the only reason for that. Sergei --- Qualcomm Innovation Center, Inc. is a member of Code Aurora Forum, hosted by The Linux Foundation > -----Original Message----- > From: llvmdev-bounces at cs.uiuc.edu [mailto:llvmdev-bounces at cs.uiuc.edu] > On Behalf Of Marcello

On Fri, Oct 19, 2012 at 04:38:29PM +0100, Marcello Maggioni wrote: > Hello, > I'm working on a compiler based on LLVM for a SIMD architecture that > supports instruction predication. We would like to implement > branching on this architecture using predication. > As you know the LLVM-IR doesn't support instruction predication, so > I'm not exactly sure on what is the

Hi Marcello, I am sure I've seen some postings on the list concerning architectures that support predicated execution and how to map that to LLVM IR, I'm just not sure anymore when and who was involved :). I have implemented your first suggestion for targets that do not have predicated instructions (where control flow to data flow conversion with explicit maintaining of masks and

Hi Marcello, On Fri, Oct 19, 2012 at 04:38:29PM +0100, Marcello Maggioni wrote: > Hello, > I'm working on a compiler based on LLVM for a SIMD architecture that > supports instruction predication. We would like to implement branching > on this architecture using predication. > As you know the LLVM-IR doesn't support instruction predication, so I'm > not exactly

Hi, while using this function I noticed that Type::getScalarSizeInBits() is not marked as const even if inside its body it just uses only other "const" functions and types. This prevents the usage of the function through const pointers/references to Type without any apparent reason. I propose of marking the function as const. (attached is a patch that does so). Cheers, Marcello --

Hi all, I recently created a version of llvm/clang for an architecture with some odd properties: - front end side: -- char,short,int : 20 bits -- long, long long: 40 bits -- pointer: 20 bits On the backend side, I have hardware support for 20-bit registers (and load/stores). (So, I have support for 'i20' and 'i40', but 'i40' is lowered into operations on 'i20')

Hi, In X86MachObjectWriter::RecordX86Relocation I found the comment if (Target.isAbsolute()) { // constant // SymbolNum of 0 indicates the absolute section. // // FIXME: Currently, these are never generated (see code below). I cannot // find a case where they are actually emitted. Type = macho::RIT_Vanilla; } Is the FIXME still true? I've got some code that

Hi, I've done work on predicated SIMD representations for LLVM. If you search through the archives, you may find my "applymask" proposal, which is an attempt at representing predication in a very comprehensive way. I've since stopped pushing the proposal in part because Larrabee's changing fortunes led to a decline of interest at the time, in part because the proposal

I have an instruction that takes no operands, and produces two results, in two consecutive cycles. I tried both of the following to my Schedule.td file: InstrItinData<IIMyInstr, [InstrStage<2, [FuncU]>], [1, 2]>, InstrItinData<IIMyInstr, [InstrStage<1, [FuncU]>, InstrStage<1, [FuncU]>], [1, 2]>, From what I can see in examples, these say that the first

Hi there, I wondered whether anyone could give me any advice about counting assembly instructions when using MCJIT? For performance regression testing I would like to be able to count the number of instructions generated during the jit compilation of a given module. The Statistic class, as far as I understand, cannot collect this data per-module (per-ExecutionEngine/per-MCJIT), and there is

Thanks Elior, Sorry for the delay - emergencies never seem to go away in compiler development ;-) That sounds like an interesting approach, a kind-of pre-processor for TD files. But will this still not result in a constant, although externally provided, or have I misunderstood? What I would like to do is compute a different value depending on which '-target-cpu <cpu>' option was

Can I assume we're talking about MCJIT since a file format and relocations are involved? Some changes are required in order to get MCJIT to generate ELF object files. Eli Bendersky submitted a patch some time ago to modify the target triple code to enable this, but after a bit of discussion there didn't seem to be a consensus for accepting this proposal. If you want to try it out, the

Hi Verena, Windows + MachO is likely to run into lots of problems once you start passing in non-trivial code. The relocation model is very tied to Darwin. I believe some folks doing JIT on Windows have had some success w/ ELF. Hopefully someone more familiar with the specifics of that will chime in. -Jim On Jun 29, 2012, at 8:58 AM, Verena Beckham <verena at codeplay.com> wrote: >

On Mar 3, 2014, at 8:53 AM, Pierre-Andre Saulais <pierre-andre at codeplay.com> wrote: > Hi Andrew, > > We are currently using a custom model where scheduling information is attached to each MCInstrDesc through tablegen, and we're trying to move to one of LLVM's models. > > To expand on what JinGu mentioned, our target has explicit ports that are used to read and

On Feb 19, 2014, at 1:54 PM, jingu <jingu at codeplay.com> wrote: > Hi Andy, > > I am trying to schedule and packetize instructions for VLIW at post-RA > stage or final codegen stage, where code transformations are not allowed > any more, because hardware can not resolve resource conflict. There is a > simple example as following: > > ADD dest_reg1, src_reg1,

Hi Jim, Andrew, Thanks, I read the discussions about using ELF + MCJIT + Windows and was hoping MachO would work just as well. Since you don't need to modify LLVM to be able to output MachO this was my first choice. But are you saying MachO will not work as well as ELF? Thanks for the patch Andrew, that will have to be my plan B in this case. Verena On 29/06/2012 23:47, Kaylor, Andrew

On Mar 4, 2014, at 10:05 AM, Pete Cooper <peter_cooper at apple.com> wrote: > > On Mar 3, 2014, at 2:21 PM, Andrew Trick <atrick at apple.com> wrote: > >> >> On Mar 3, 2014, at 8:53 AM, Pierre-Andre Saulais <pierre-andre at codeplay.com> wrote: >> >>> Hi Andrew, >>> >>> We are currently using a custom model where

Hi Verena, I think that we can count the number of instructions with "-stats" command line option. As you mentioned, this option uses Statistic class like "STATISTIC(EmittedInsts, "Number of machine instrs printed");" I don't know exactly about parallel code generation environment but this option seems like to work correctly in common case as following. This is

Hi list, I happened to read below thread (written in 3 years ago). I think I may need this ReadAdvance feature to work with my ARCH. It is about the scheduler info which describes reading my ARCH's vector register. There are different latencies since forwarding/bypass appears. I give it as below example: def : WriteRes<WriteVector, [MyArchVALU]> { let Latency = 6; } ... def

Thanks Andrew. I have tried with recent tblgen, ReadAdvance would not work for multiple latencies. Maybe I should make improvement into tblgen if Pierre-Andre does not have the change anymore. However, I just a little curious about the situation I met. The hardware forwording may fail for different reasons, which different register read may have different latencies, depending both on the register