similar to: view how instruction selection is happening

In an attempt to add vector registers to my target, I ran into a problem. LLVM started to complain about not being able to infer types from the provided DAG patterns for several classes of instructions. After a discussion on the llvm-dev mailing list and IRC channel the recommendation was to make DAG patterns for these classes of instructions more specific. Which is what was done. However after

On Thu, Feb 18, 2016 at 11:22 AM, Jonathan Roelofs < jonathan at codesourcery.com> wrote: > > > On 2/18/16 12:19 PM, Rail Shafigulin via llvm-dev wrote: > >> >> >> On Thu, Feb 18, 2016 at 10:50 AM, Krzysztof Parzyszek via llvm-dev >> <llvm-dev at lists.llvm.org <mailto:llvm-dev at lists.llvm.org>> wrote: >> >> On 2/18/2016

On Thu, Feb 18, 2016 at 10:50 AM, Krzysztof Parzyszek via llvm-dev < llvm-dev at lists.llvm.org> wrote: > On 2/18/2016 12:43 PM, Rail Shafigulin via llvm-dev wrote: > >> >> LLVM ERROR: Cannot select: 0x3284268: glue = EsenciaISD::SET_FLAG >> 0x3283608, 0x3283710, 0x3283e48 [ORD=3] [ID=11] >> 0x3283608: i32,ch = CopyFromReg 0x3257980, 0x3283500 [ORD=1] [ID=9]

I recently had to change some DAG patterns for my target. Conceptually I understand what DAG patterns are and what is their purpose, however given that I have quite vague understanding about how to LLVM implements and specifies them, naturally I got an error. Here it is Here is the error: LLVM ERROR: Cannot select: 0x3284268: glue = EsenciaISD::SET_FLAG 0x3283608, 0x3283710, 0x3283e48 [ORD=3]

> > > > > def SDTX86CmpPTest : SDTypeProfile<1, 2, [SDTCisVT<0, i32>, > SDTCisVec<1>, > SDTCisSameAs<2, 1>]>; > > This is confusing to me. This tells me that there is 1 result but and 2 > operands. But then it says that operands 2 and 1 are of the same type,

On Thu, Feb 18, 2016 at 11:34 AM, Krzysztof Parzyszek < kparzysz at codeaurora.org> wrote: > On 2/18/2016 1:32 PM, Rail Shafigulin wrote: > >> I think this is where I'm loosing the "thread". Based on what I'm seeing >> SET_FLAG has three operands, the first of which is a CopyFromReg. So how >> come the pattern is SET_FLAG %vreg5, 3, 20 and not

I have added new instructions to my target, unfortunately they are not being properly type legalized. The instructions I've added are a vector add, vector load and a vector store. Can anyone lend a hand on how to legalize them so that my target would be able to recognize them. Below is the output of llc with a -debug-only=isel. As you could see the output type for load, store, and add changes

> On Feb 1, 2016, at 16:53, Rail Shafigulin <rail at esenciatech.com> wrote: > > > > On Fri, Jan 29, 2016 at 10:03 PM, Matt Arsenault <arsenm2 at gmail.com <mailto:arsenm2 at gmail.com>> wrote: > > > On Jan 29, 2016, at 13:25, Rail Shafigulin via llvm-dev <llvm-dev at lists.llvm.org <mailto:llvm-dev at lists.llvm.org>> wrote: > > >

On Tue, Feb 2, 2016 at 1:41 AM, Rail Shafigulin <rail at esenciatech.com> wrote: > > Let me clarify. >> >> I'm not sure I understand what you are saying. Let me post more >> information. >> >> Here is what I have defined for Escalasetflag >> >> def Escalasetflag : SDNode<"EscalaISD::SET_FLAG", SDT_EscalaSetFlag, >>

I'm trying to figure out how to interpret DAG output that is shown below. If I were to draw it by hand, how would I do it? I tried to do it, but unfortunately I can't figure it out. I know there is a way to generate a graph, but I made some changes to my backend which hangs clang/llc. Would someone be kind enough to give a picture of what of a graph below? I'm sure once I see one I

On Tue, Feb 2, 2016 at 8:42 PM, Matt Arsenault <arsenm2 at gmail.com> wrote: > > On Feb 2, 2016, at 16:52, Rail Shafigulin <rail at esenciatech.com> wrote: > > def SDT_EscalaSetFlag : SDTypeProfile<0, 3, [SDTCisSameAs<0, 1>]>; > > > I think for setting an implicit register, you still need to have 1 result > here. > > If you look at

It does have an output register, it's just an implicit flag register. It still has a DAG output. I'm not sure if the allocatable bit matters at this point for selection purposes, but it does later. Not adding a type to the register class can also be problematic (e.g. a flag register should have i1 added to regTypes for its class). -Matt > Does LLVM make an assumption that there is an

On Fri, Jan 29, 2016 at 10:22 AM, Krzysztof Parzyszek via llvm-dev < llvm-dev at lists.llvm.org> wrote: > On 1/28/2016 8:11 PM, Rail Shafigulin via llvm-dev wrote: > >> >> Would anyone be able to figure out why this is happening? I can provide >> more code if needed. >> > > The error message should show what types have been inferred so far. > > You

> On Jan 29, 2016, at 13:25, Rail Shafigulin via llvm-dev <llvm-dev at lists.llvm.org> wrote: > > > > I think I understand it. But looks like I have everything labelled properly. Maybe I missed something. Here are more details: > > defm SFEQ : SF<0x0, "l.sfeq", Escala_CC_EQ>; > > multiclass SF<bits<5> op2Val, string asmstr, PatLeaf

On Fri, Feb 19, 2016 at 6:10 AM, Krzysztof Parzyszek via llvm-dev < llvm-dev at lists.llvm.org> wrote: > On 2/18/2016 6:01 PM, Rail Shafigulin via llvm-dev wrote: > >> [(set SR:$rD, (Esenciasetflag (i32 GPR:$rA), (i32 >> immSExt16:$imm), (i32 Cond)))]> { >> > > I suspect that the "set SR:$rD" is the problem here. The Esenciasetflag

--- .gitignore | 1 + mllib/Makefile.am | 30 ++++++++- mllib/getopt_tests.ml | 68 ++++++++++++++++++++ mllib/test-getopt.sh | 168 ++++++++++++++++++++++++++++++++++++++++++++++++++ 4 files changed, 265 insertions(+), 2 deletions(-) create mode 100644 mllib/getopt_tests.ml create mode 100755 mllib/test-getopt.sh diff --git a/.gitignore b/.gitignore index 000e984..c1ae484

On Sat, Mar 12, 2016 at 2:38 PM, Tim Northover <t.p.northover at gmail.com> wrote: > On 12 March 2016 at 11:51, Rail Shafigulin via llvm-dev > <llvm-dev at lists.llvm.org> wrote: > > I tried every possible combination of --target I could think of but > nothing > > worked. Would you mind helping me out? > > First, 64-bit x86 is "x86_64", and 32-bit

My target has an instruction that adds up all elements in the vector and stores the result in a register. I'm trying to implement it in my compiler but I'm not sure even where to start. I did look at other targets, but they don't seem to have anything like it ( I could be wrong. My experience with LLVM is limited, so if I missed it, I'd appreciate if someone could point it out ).

Hi Rail, Below 2 revisions might be of your interest which Detect SAD patterns and emit psadbw instructions on X86.: http://reviews.llvm.org/D14840 http://reviews.llvm.org/D14897 Intrinsics related to absdiff revisons : http://reviews.llvm.org/D10867 http://reviews.llvm.org/D11678 Hope this helps. Regards, Suyog On Sat, May 28, 2016 at 4:20 AM, Rail Shafigulin via llvm-dev < llvm-dev at

On Thu, Jan 28, 2016 at 8:34 PM, Dylan McKay <dylanmckay34 at gmail.com> wrote: > Try visualising the DAG like this. > > ``` > ---- GPR:$rA > / > set GPR:$rd ---- add > \ > ---- GPR:$rB > ``` > > Each instruction forms a DAG with its operands being subnodes. > >