similar to: [LLVMdev] MMX vs SSE

This thread confuses me. I thought Chris said that LLVM 2.8 will not lower generic vectors to MMX because it breaks x87 code, and I didn't see an answer to your question about a switch to tell the code generator otherwise. However, you're complaining that MMX performance is subpar, even though LLVM 2.8 isn't supposed to generate MMX instructions. Can someone clarify the situation

On Sep 21, 2010, at 5:30 PMPDT, Bill Wendling wrote: > LLVM isn't going to stop generating MMX instructions all together. We can't do that. :-) If the user specifically wants MMX (by, say, using the builtins), we have to support that still. The plan to cease generating MMX for generic vectors is a work-in-progress right now. It's not in 2.8. > > -bw Right, early on there

Assign the bug to me and I'll fix it in TOT next week! Thanks for narrowing it down! On Wednesday, September 22, 2010, Nicolas Capens <nicolas.capens at gmail.com> wrote: > Hi all, > > I think I figured it out: > 112804 causes 64-bit UNPCKLBW to no longer be selected for certain cases. > 112805 is benign. > 112806 causes 64-bit UNPCKHBW to no longer be selected for

LLVM isn't going to stop generating MMX instructions all together. We can't do that. :-) If the user specifically wants MMX (by, say, using the builtins), we have to support that still. The plan to cease generating MMX for generic vectors is a work-in-progress right now. It's not in 2.8. -bw On Sep 21, 2010, at 4:24 PM, Reid Kleckner wrote: > This thread confuses me. I thought

On Jul 1, 2011, at 2:43 PM, Alistair Lynn wrote: > Hi Andrew- > >> fatal error: error in backend: SSE2 register return with SSE2 disabled > > Is this for 32-bit or 64-bit x86? 64-bit x86. > If it's the latter, the ABI demands > that the return value in this case is in xmm0 - SSE is required. > Well -no-sse -mno-mmx works for EFI as it is pre-boot firmware and

On Wed, Sep 8, 2010 at 12:35 AM, Nicolas Capens <nicolas.capens at gmail.com> wrote: > Hi Chris, > > It's not broken, but the performance is crippled. > > I noticed that the code still contains some MMX instructions, but several > operations get expanded (apparently swizzling and such get expanded to a > large number of byte moves). I think some changes related to

Hi Dale, I will not change the default. I would dislike to see any regressions due to this type of change. -- Mon Ping On Nov 20, 2008, at 10:12 AM, Dale Johannesen wrote: > > On Nov 19, 2008, at 11:57 PMPST, Mon Ping Wang wrote: > >> Hi, >> >> When setting -mattr option on X86, I would like to treat MMX >> separately from SSE levels. This would allow a

Your clang invocation below works for me, and generates target triple in the llvm IR of i386. And then in the specific options for the functions it generates the following: ; Function Attrs: nounwind define float @foo() #0 { entry: ret float 1.000000e+00 } attributes #0 = { nounwind "less-precise-fpmad"="false" "no-frame-pointer-elim"= "true"

Hi Andrew- > Well -no-sse -mno-mmx works for EFI as it is pre-boot firmware and does not have any floating point C code. We use -no-sse and -mno-mmx code to prevent optimized code gen using these registers for optimizations. Whether it's optimised or not doesn't particularly matter, the x86_64 ABI says that floating-point return values go into SSE registers, so that is where LLVM is

On Nov 19, 2008, at 11:57 PMPST, Mon Ping Wang wrote: > Hi, > > When setting -mattr option on X86, I would like to treat MMX > separately from SSE levels. This would allow a client who sets the > attributes directly to set the SSE level independent of MMX, e.g., llc > -march=x86 -mattr=sse41, one would get sse4.1 with mmx disabled while > llc -march=x86 -mattr=mmx

Hi, I’ve come back and updated my llvm toolset with modern code (my branch was about 1-2 years old) and now the llvm IR files produced by my front end no longer compile with llc. Here is a sample of llvm ir produced by my front end (it’s a standard version 3.1 build of swift from the swift.org open source website). ; ModuleID = 'main.ll' source_filename = "main.ll" target

Hi, When setting -mattr option on X86, I would like to treat MMX separately from SSE levels. This would allow a client who sets the attributes directly to set the SSE level independent of MMX, e.g., llc -march=x86 -mattr=sse41, one would get sse4.1 with mmx disabled while llc -march=x86 -mattr=mmx -mattr=sse42 will get mmx and sse42. If anyone objects to this change, please let me

Thanks Kevin for the reply. I got the point now :) On 10 Apr 2015 00:18, "Smith, Kevin B" <kevin.b.smith at intel.com> wrote: > For x86_64 ABI, a minimum feature set of SSE2 is required. > > > > Kevin > > > > *From:* llvmdev-bounces at cs.uiuc.edu [mailto:llvmdev-bounces at cs.uiuc.edu] *On > Behalf Of *suyog sarda > *Sent:* Thursday, April 09,

That’s useful info, thanks. I think it will be useful for me to understand the connection, why this type of pointer is being emitted now. Do you have any suggestions where i can look to find the platform specific code that is making function pointers go into addrspace? Carl p.s. I am also working on passing the avr target flag to swift, but swift itself had (has?) limitations that make it

0x0f 0x6f 0xc8 And 0x0f 0x7f 0xc1 Should both be movq % mm0, % mm1. (AT&T) However, llvm 3.4 at least does not recognise the second variant as being a valid instruction. We are currently compiling up latest src incase it has been fixed. If not, could someone take a look or recommend how to fix? Lee -------------- next part -------------- An HTML attachment was scrubbed... URL:

Hi Andrew- > fatal error: error in backend: SSE2 register return with SSE2 disabled Is this for 32-bit or 64-bit x86? If it's the latter, the ABI demands that the return value in this case is in xmm0 - SSE is required. Alistair

Cool. That explains a lot! Sorry if this is a total n00b question, but… how does the datalayout string get overridden? in llvm/lib/Target/AVR/AVRTargetMachine.cpp I can see the code that determines the default datalayout for AVR… static const char *AVRDataLayout = "e-P1-p:16:8-i8:8-i16:8-i32:8-i64:8-f32:8-f64:8-n8-a:8”; However in the LLVM iR below, the target datalayout was present and

I recently diagnosed a bug in someone else's software, which turned out to be due to incorrect MMX intrinsics usage: if you use any of the x86 intrinsics that accept or return __m64 values, then you, the *programmer* are required to call _mm_empty() before using any x87 floating point instructions or leaving the function. I was aware that this was required at the assembly-level, but not that

On Mon, Aug 31, 2020 at 3:02 PM Eli Friedman <efriedma at quicinc.com> wrote: > Broadly speaking, I see two problems with implicitly enabling MMX > emulation on a target that has SSE2: > > > > 1. The interaction with inline asm. Inline asm can still have MMX > operands/results/clobbers, and can still put the processor in MMX mode. If > code is mixing MMX

On Oct 20, 2011, at 8:42 AM, Nicolas Capens wrote: > Hi all, > > I'm working on a graphics project which uses LLVM for dynamic code > generation, and I noticed a major performance regression when upgrading > from LLVM 2.8 to 3.0-rc1 (LLVM 2.9 didn't support Win64 so I skipped it > entirely). > > I found out that the performance regression is due to removing