search for: movmskp

...ht not produce anything good, the > free sext requires llvm 2.7 on x86 to work at all (certainly shouldn't > be a problem nowadays but on other backends it might be different) and > for the ptest sequence very recent svn is required. > I don't think the current code can generate movmskps + test (probably > the next best thing without sse41) instead of ptest though if you only > got sse. Thanks Roland, sign extending gets me part of the way at least. I'm on version 3.1 and as you say in bug report, there are a few extraneous instructions. For the record, casting to a &l...

...nything good, the >> free sext requires llvm 2.7 on x86 to work at all (certainly shouldn't >> be a problem nowadays but on other backends it might be different) and >> for the ptest sequence very recent svn is required. >> I don't think the current code can generate movmskps + test (probably >> the next best thing without sse41) instead of ptest though if you only >> got sse. > > > Thanks Roland, sign extending gets me part of the way at least. > I'm on version 3.1 and as you say in bug report, there are a > few extraneous instructions...

...nt a feature, so I want your advice on its feasibility and difficulty. I want the code generator to support such code : %1 = ; <4 x float> vector %2 = ; <4 x float> vector %3 = fcmp ogt <4 x float> %1, %2 %4 = bitcast <4 x i1> %3 to i4 %5 = zext i4 %4 to i32 and generate a movmskps instruction (correct me if it's not possible). So the code generator would have to find such pattern (fcmp followed by a bitcast and/or zext) and generate correct assembly (using movmskps or a slower code path). Currently, LLVM 3.3 seems to generate wrong code (see bug 17479 : http://llvm.org/...

...ng how it is consumed, it is best to look at what happens to v8i1. We can then let the same optimizer work to get the optimal ASM code out in the end, whether vectorization factor is 4 or 8. In the end, I may be agreeing to Rob, but not because of the reasons Rob mentioned. One of the headaches is movmskps/pmovmskb do not have a quick reverse instruction (MIC-AVX512 and below). I do not know LLVM's X86 CodeGen enough to say whether it internally has mask-to/from-vector nodes. If it has, I'd hope X86 CodeGen can cancel out such things in a peephole manner very efficiently so that blindly goin...

...ec this sequence might not produce anything good, the free sext requires llvm 2.7 on x86 to work at all (certainly shouldn't be a problem nowadays but on other backends it might be different) and for the ptest sequence very recent svn is required. I don't think the current code can generate movmskps + test (probably the next best thing without sse41) instead of ptest though if you only got sse. Roland

> > which goes through memory. Is there some idiom I'm missing so that it would use > > for instance movmsk for SSE or vcmpgt & cr6 for altivec? > > I don't think you are missing anything: LLVM IR has no support for horizontal > operations like or'ing the elements of a vector of boolean together. The code > generators do try to recognize a few idioms and

...rse method, the masks are kept in <4 x 32bit> registers, which are mapped to xmm registers. This is the ‘native’ way of using masks. In the second representation, the packed method, the MSB bits are collected from the xmm register into a packed general purpose register. Luckily, SSE has the MOVMSKPS instruction, which converts sparse masks to packed masks. I am not sure which representation is better, but both are reasonable. The former may cause register pressure in some cases, while the latter may add the packing-unpacking overhead. _Sparse_ After my discussion with Duncan, last week, I s...

"Rotem, Nadav" <nadav.rotem at intel.com> writes: > I can think of two ways to represent masks in x86: sparse and > packed. In the sparse method, the masks are kept in <4 x 32bit> > registers, which are mapped to xmm registers. This is the ‘native’ way > of using masks. This argues for the sparse representation, I think. > _Sparse_ After my discussion with

...alizing of <4 x i1> to <4 x i32> is the way to go. Cheers, Nadav -----Original Message----- From: Rotem, Nadav Sent: Thursday, March 10, 2011 11:04 To: 'David A. Greene' Cc: llvmdev at cs.uiuc.edu Subject: RE: [LLVMdev] Vector select/compare support in LLVM Hi David, The MOVMSKPS instruction is cheap (2 cycles). Not to be confused with VMASKMOV, the AVX masked move, which is expensive. One of the arguments for packing masks is that it reduces vector-registers pressure. Auto-vectorizing compilers maintain multiple masks for different execution paths (for each loop nesti...

On Dec 25, 2008, at 11:02 AM, Eli Friedman wrote: > On Thu, Dec 25, 2008 at 1:54 AM, Eli Friedman > <eli.friedman at gmail.com> wrote: >> On Thu, Dec 25, 2008 at 1:28 AM, Claudio Basile <cbasile at tempo- >> da.com> wrote: >>> Hi all, >>> >>> is there any way to compare two 128bit values? >>> I have tried 3 different approaches

On Thu, Dec 25, 2008 at 1:54 AM, Eli Friedman <eli.friedman at gmail.com> wrote: > On Thu, Dec 25, 2008 at 1:28 AM, Claudio Basile <cbasile at tempo-da.com> wrote: >> Hi all, >> >> is there any way to compare two 128bit values? >> I have tried 3 different approaches and they all fail with an internal >> assertion. >> I'm running llvm 2.4 on

...regs.eip = (unsigned long)&instr[0]; + regs.ecx = 0; + regs.edx = (unsigned long)res; + rc = x86_emulate(&ctxt, &emulops); + if ( rc != X86EMUL_OKAY ) + goto fail; + asm ( "cmpeqps %1, %%xmm7\n\t" + "movmskps %%xmm7, %0" : "=r" (rc) : "m" (res[8]) ); + if ( rc != 0xf ) + goto fail; + printf("okay\n"); + } + else + printf("skipped\n"); + for ( j = 1; j <= 2; j++ ) { #if defined(__i386__) --- a/xen/arch/x86/x...

...> > Cheers, > Nadav > > -----Original Message----- > From: Rotem, Nadav > Sent: Thursday, March 10, 2011 11:04 > To: 'David A. Greene' > Cc: llvmdev at cs.uiuc.edu > Subject: RE: [LLVMdev] Vector select/compare support in LLVM > > Hi David, > > The MOVMSKPS instruction is cheap (2 cycles). Not to be confused with VMASKMOV, the AVX masked move, which is expensive. > > One of the arguments for packing masks is that it reduces vector-registers pressure. Auto-vectorizing compilers maintain multiple masks for different execution paths (for each lo...

On Tue, Jun 28, 2016 at 2:45 AM, jingu kang via llvm-dev <llvm-dev at lists.llvm.org> wrote: > Hi All, > > Can someone comment below question whether it is wrong or not please? > > 2016-06-25 7:52 GMT+01:00 jingu kang <jaykang10 at gmail.com>: >> Hi All, >> >> I have a problem with VectorLegalizer::ExpandStore() with v4i1. >> >> Let's

Hi David, The MOVMSKPS instruction is cheap (2 cycles). Not to be confused with VMASKMOV, the AVX masked move, which is expensive. One of the arguments for packing masks is that it reduces vector-registers pressure. Auto-vectorizing compilers maintain multiple masks for different execution paths (for each loop nesti...