similar to: [LLVMdev] problem using 128-bit integer on x86-32

Hi Jeff, > I'm seeing LLVM (v 3.1) abort when trying to generate code that > multiplies or divides 128-bit integers on x86-32. this is a known issue. The runtime library (libgcc) has routines for dividing two 64-bit integers on x86-32, but not two 128-bit integers. At least, that is how it was last time I looked. To overcome this libgcc or LLVM's compiler-rt would need to get

there is my code,  expect return value  is a data frame but R say it is list: SEXP Julia_R_MD_NA_DataFrame(jl_value_t* Var) {  SEXP ans,names,rownames;  char evalcmd[4096];  int i;  const char* dfname="DataFrameName0tmp";  jl_set_global(jl_main_module, jl_symbol(dfname), (jl_value_t*)Var);  //Get Frame cols   sprintf(evalcmd,"size(%s,2)",dfname);  jl_value_t*

After writing a simple FPRange, I've hit a stumbling block. I don't know what LLVM code should be extended to use it. I was initially thinking of extending LazyValueInfo, but it appears to be used for passes that don’t address the case that I need for Julia. I’m now wondering if I’m better off extending SimplifyFCmpInst to handle the few cases in question instead of trying to be more

How is ccall() implemented? If it manually sets up a stack frame, then it also needs to store argument shadow values in paramtls. I don't think there is an overflow, unless you have a _lot_ of arguments in a function call. On Sun, Feb 2, 2014 at 9:26 AM, Keno Fischer <kfischer at college.harvard.edu> wrote: > Also, I was looking at the instrumented LLVM code and I noticed that the

The code for ccall looks right. Sounds like you have a very small range of instructions where an uninitialized value appear. You could try debugging at asm level. Shadow for b should be passed at offset 0 in __msan_param_tls. MSan could propagate shadow through arithmetic and even some logic operations (like select). It could be that b is clean on function entry, but then something uninitialized

Most probably you need to properly specify the calling convention the backend is using for calling the runtime functions. Or implement the stub for udivti3 that performs the necessary argument lifting. I guess there is no standard ABI document describing the intended calling convention here, so I'd just do what mingw64 does here and make everything here compatible. On Thu, Apr 26, 2018 at

I'm trying to use LLVM to create compiler-rt.o on Windows. I use this command from the compiler-rt project: [nix-shell:~/downloads/llvm-project/compiler-rt]$ clang -nostdlib -S -emit-llvm lib/builtins/udivti3.c -g -target x86_64-windows -DCRT_HAS_128BIT The resulting LLVM IR is: ================================================================= ; ModuleID = 'lib/builtins/udivti3.c'

On Thu, Apr 26, 2018 at 3:44 AM, Anton Korobeynikov <anton at korobeynikov.info > wrote: > Most probably you need to properly specify the calling convention the > backend is using for calling the runtime functions. Thanks for the tip. Can you be more specific? Are you suggesting there is some config parameter I can set before running TargetMachineEmitToFile? Do you know what

I think from the C compiler's point of view, it is going to want it to work for any size above an i64, i.e. all the way up to an i128 so that if the user of the C compiler does this computation with __uint128_t's then it will Do The Right Thing TM. Basically, you want unsigned long a, b, c, d; .... const __uint128_t u = (__uint128_t) a + b; const unsigned long v = u >> 64; const

Looks like when you materialize the stores, you should check the size of the the store and emit an appropriate amount of stores to the origin shadow (or just a memset intrinsic?). On Wed, Feb 5, 2014 at 2:13 PM, Keno Fischer <kfischer at college.harvard.edu>wrote: > The @entry stuff is just a gdb artifact. I've been tracking this back a > little further, and it seems there's

Yeah I had a think about it, and I think intrinsics are the wrong way to do it. So I'd say you are likely right. Bill. On 13 June 2010 04:33, Alistair Lynn <arplynn at gmail.com> wrote: > Hi Bill- > > I think, ideally, the backend would be able to match arbitrary-precision arithmetic to add-with-carry or subtract-with-borrow through i65/i33. That would remove the need for the

On 12 June 2010 00:51, Eli Friedman <eli.friedman at gmail.com> wrote: > On Fri, Jun 11, 2010 at 3:28 PM, Bill Hart <goodwillhart at googlemail.com> wrote: >> Hi Eli, >> >> On 11 June 2010 22:44, Eli Friedman <eli.friedman at gmail.com> wrote: >>> On Fri, Jun 11, 2010 at 10:37 AM, Bill Hart <goodwillhart at googlemail.com> wrote:

Yes, it would be great to get that fixed. On Wed, Feb 5, 2014 at 4:09 PM, Evgeniy Stepanov <eugeni.stepanov at gmail.com>wrote: > On Thu, Feb 6, 2014 at 12:21 AM, Keno Fischer > <kfischer at college.harvard.edu> wrote: > > Looks like when you materialize the stores, you should check the size of > the > > the store and emit an appropriate amount of stores to the

I was able to get the loop to increment from -999 to 0 using IR directly. That got rid of the cmpq. The carry i was after was able to be obtained using the intrinsic @llvm.uadd.with.overflow.i64, however there is no way to add with carry and have it realise that the resulting *carry out* cannot exceed 1. It actually writes the carry to a byte, and then uses logical operations on it, which slows

Hi Bill- I think, ideally, the backend would be able to match arbitrary-precision arithmetic to add-with-carry or subtract-with-borrow through i65/i33. That would remove the need for the overflow intrinsics entirely. Alistair On 13 Jun 2010, at 02:27, Bill Hart wrote: > I was able to get the loop to increment from -999 to 0 using IR > directly. That got rid of the cmpq. > > The

In both LLVM 2.9 and the current svn head, I get the following error when running llc % llc < fxp2.ll LLVM ERROR: Cannot select: 0xa5302b0: glue = carry_false [ID=7] on this code: target triple = "i386-pc-linux-gnu" define i32 @fxpadd(i32 %cl) { entry: %0 = zext i32 %cl to i128 %1 = zext i32 %cl to i128 %2 = add i128 %1, %0 br label %L1001510 L1001510:

I am seeing i128 from llvm-gcc on Alpha.  I know the calling convention for them, they are split into two registers, but I don't know if that should be handled in the frontend or the backend.  I would just as soon do it in the backend, but I didn't see any support in the new calling convention work for automatically splitting an argument into multiple registers. Is the backend the best

Hi everyone! Here, I have a question and am curious about i128. I want to know how the LLVM handle i128, because many compiler backend doesn't support i128 directly. So I am very curious and want to how the llvm handle this situation? Besides i128, such as i256, i512, even i24? Thanks. Best Regards Wu Zhao -------------- next part -------------- An HTML attachment was scrubbed...

Hi all, I have a bunch of new registers set up in X86RegisterInfo.td, the important part being def PR128 : RegisterClass<"X86", [i128], 128, (sequence "POI%u", 0, 7)>; def VR128 : RegisterClass<"X86", [v4f32, v2f64, v16i8, v8i16, v4i32, v2i64], 128, (add PR128, FR32)>; I have an entry in

It seems that one of the optimization passes (it seems to be SROA) sometimes transforms computations on vectors of ints to computations on wide integer types; for example, I'm seeing code like the following after optimizations(*): %0 = bitcast <16 x i8> %float2uint to i128 %1 = shl i128 %0, 8 %ins = or i128 %1, 255 %2 = bitcast i128 %ins to <16 x i8> The back end I'm