similar to: Error using ddply inside user-defined function

Hi, there I am writing an backend, and I met a problem. We don't have load/store instructions for vector predicate registers(vpr for short). The hardware has 64 vector registers(vr for short) and 8 vector predicate registers. And there is no move instructions between vr and vpr. vr supports many operations, and vpr supports vpror, vprxor, vprand and vprinv operations. A vr has 512 bits, and

Hi, I am planning to expanding the pseudo instructions in XXXTargetLowering::EmitInstrWithCustomInserter(), and use temporary virtual registers as operands. If I use virtual registers, do I need to mark them as "early clobber"? I saw that sometimes they marked virtual register as "early clobber" in EmitInstrWithCustomInserter() in MIPS backend. What is the effect of marking a

On May 9, 2013, at 3:05 PM, Jeff Bush <jeffbush001 at gmail.com> wrote: > On Thu, May 9, 2013 at 8:10 AM, <dag at cray.com> wrote: >> Jeff Bush <jeffbush001 at gmail.com> writes: >> >>> %tx = select %mask, %x, <0.0, 0.0, 0.0 ...> >>> %ty = select %mask, %y, <0.0, 0.0, 0.0 ...> >>> %sum = fadd %tx, %ty >>> %newvalue

Hi,   For a custom target, there is a pass to perform memory dependence analysis, where, i need to get memory pointer for "load instruction". I want to check the pointer alias behavior. I am getting this by considering the memoperands for the load instruction.   For "load instruction", Machine Instruction dumps as below:   vr12<def> = LD_Iri %vr2<kill>, 0;

On Thu, May 9, 2013 at 8:10 AM, <dag at cray.com> wrote: > Jeff Bush <jeffbush001 at gmail.com> writes: > >> %tx = select %mask, %x, <0.0, 0.0, 0.0 ...> >> %ty = select %mask, %y, <0.0, 0.0, 0.0 ...> >> %sum = fadd %tx, %ty >> %newvalue = select %mask, %sum, %oldvalue >> >> I believe the generated instructions depend on whether

I'm trying to understand how predicated/masked instructions can be generated in llvm, specifically an instruction where a set bit in the mask will write the new result into the corresponding vector lane in the destination and a clear bit will cause the lane in the destination to remain what it was before the instruction executed. I've seen a few places that suggest 'select' is the

Machine Instruction Bundle in LLVM Hi all, There have been quite a bit of discussions about adding machine instruction bundle to support VLIW targets. I have been pondering what the right representation should be and what kind of impact it might have on the LLVM code generator. I believe I have a fairly good plan now and would like to share with the LLVM community. Design Criteria 1. The

Hi Evan, I just read your proposal and the following discussion for VLIW support and want to share my experience of writing a VLIW back-end for LLVM. I would not integrate the packetizer into the register allocator super class since it would reduce the flexibility for the back-end developer to add some optimization passes after the packetizer. Instead, I would add the packetizer as a separate

. and yes, one more thing. On some architectures it might be desirable to know the _order_ of instructions in the packet. That is a bit trickier.. -- Qualcomm Innovation Center, Inc. is a member of Code Aurora Forum. From: llvmdev-bounces at cs.uiuc.edu [mailto:llvmdev-bounces at cs.uiuc.edu] On Behalf Of Evan Cheng Sent: Friday, December 02, 2011 2:40 PM To: LLVM Dev Subject: [LLVMdev]

On Dec 2, 2011, at 2:41 PM, Sergei Larin wrote: > … and yes, one more thing. On some architectures it might be desirable to know the _order_ of instructions in the packet. That is a bit trickier…. Isn't that just the order of the instructions in the list? I don't see anything that prevents getting the order of instructions. It might require iterator over MIs in the packet. But for

Evan, I will need to comprehend it better, but one small comment right away. Did we not discuss one more option for bundle implementation - global cycle ID. We would add an unsigned int field to MI definition representing "global scheduling cycle". All MIs with the same global cycle value belong to one group/packet. Zero means unscheduled MI. That is light weight,

Hello, Freebsd-stable. BIND on my new router (7.1-STABLE, BIND 9.4.3-P1) shows bunch of errors on every start and doesn't answer on requests for 30-60 seconds after that. Errors are like this: Jan 24 12:18:12 gateway named[1455]: /usr/src/lib/bind/isc/../../../contrib/bind9/lib/isc/unix/socket.c:1567: unexpected error: Jan 24 12:18:12 gateway named[1455]: internal_send: 193.0.14.129#53:

Ah, I think I get it now. This was mentioned earlier in the thread, but it didn't click at the time. It sounds like I can do instruction selection with a pattern like (omitting selection of the sources): let Constraints = "$dst = $oldvalue" in { def MASKEDARITH : MyInstruction< (outs VectorReg:$dst), (ins MaskReg:$mask, VectorReg:$src1, VectorReg:$src2,

Hi, I am writing an analysis pass for a custom processor. I get an unusual situation where the code generated for a BB is BB#23: derived from LLVM BB %sw.bb99 Live Ins: %vr2 %vr0 %vr1 %vr9 %vr3 %vr8 %vr4 %vr5 %vr6 Predecessors according to CFG: BB#22 %vr46<def> = LD_Iri %LV, -4; mem:LD4[FixedStack0] %vr7<def> = ADDri %vr9, 1 %vr47<def> = ADDri

On Dec 2, 2011, at 12:40 PM, Evan Cheng wrote: > There have been quite a bit of discussions about adding machine instruction bundle to support VLIW targets. I have been pondering what the right representation should be and what kind of impact it might have on the LLVM code generator. I believe I have a fairly good plan now and would like to share with the LLVM community. Let me add some

Dear list, ? I am intersted in estimating movement based kernel densities for fish that were relocated at fixed receivers positioned along the coast. These data tend to display both a drift movment between receivers and a random movement component that can be estimated from the mean and the variance of the transit time between receivers. ?If I obtain an estimate of the diffusion coefficient from

R developers, I have a first attempt to make an aov function. Eventually I want to build in Error() structure, but first I am trying to get this presentable for balanced data with only a single stratum, just using residual error. I am following R. M. Heiberger's Computation for the Analysis of Designed Experiments, Wiley (1989) I a using a wrapper (aov.bal) to call the

I frequently get data sets with cryptically-named variables. The datasets are more useful to me with informative variable names. I know that I can rename variables using the following command: dimname(dataset[[2]][index.of.variable.to.be.renamed]<-new.variable.name If I want to do this inside a function (say something I call RenameCol) what is the best way to communicate the

The Asterisk.org development team has announced the release of Libpri versions 1.2.6 and 1.4.2. These releases fix the 2 B-channel transfer feature in both 1.2 and 1.4 on the NI2 and 5ESS switch types as well as a handful of other fixes. See the respective Changelogs for more details. Both releases are available as a tarball as well as a patch against the previous release. They are available

Hi, I've raised an assertion when I tried to remove an instruction within a bundle by calling eraseFromParent(). It seems that in order to erase the MI, a bundle iterator is created but it's forbidden if the MI is not the bundle's header. The attached patch fix it. As a test case, call eraseFromParent() on a machine instruction with its isInsideBundle flag activated. Ivan