similar to: Pattern transformation between scalar and vector on IR.

Hi, We have implemented the rsqrt instruction generation for X86 target architecture. We have introduced a flag -fp-rsqrt flag which controls the generatation of X86 rsqrt instruction generation. We have observed minor effects on precision due to rsqrt and hence has put these transformations under the mentioned flag. Note that -fp-rsqrt is only enabled with -enable-unsafe-fp-math flag presently.

Hi, Please find attached the modified patch and description. We have modified and retested the patch taking into consideration the comments and inputs provided earlier. Thanks & Regards, soham -----Original Message----- From: Eli Friedman [mailto:eli.friedman at gmail.com] Sent: Thursday, November 15, 2012 12:59 PM To: Chakraborty, Soham Cc: llvmdev at cs.uiuc.edu Subject: Re: [LLVMdev]

On Wed, Nov 14, 2012 at 10:43 PM, Chakraborty, Soham <Soham.Chakraborty at amd.com> wrote: > Hi, > > > > We have implemented the rsqrt instruction generation for X86 target > architecture. We have introduced a flag -fp-rsqrt flag which controls the > generatation of X86 rsqrt instruction generation. > > We have observed minor effects on precision due to rsqrt and

I'm getting this on LLVM trunk: SplitVectorResult #0: 0x27e6250: v8f32 = llvm.x86.avx.rsqrt.ps.256 0x2739310, 0x2739420 [ORD=16] [ID=0] LLVM ERROR: Do not know how to split the result of this operator! clang: error: linker command failed with exit code 1 (use -v to see invocation) Oddly, when I build the same code without -flto I don't see this issue. I see a similar bug was reported

All llvm-devs, We're going to introduce the new possible implementation for such optimizations as reciprocal estimation instead of fdiv. In short it's a replacement of fdiv instruction (which is very expensive in most of CPUs) with alternative sequence of instructions which is usually cheaper but has appropriate precision (see genReciprocalDiv in lib/Target/X86/X86InstrInfo.cpp for

In fact to commit the change before dealing with worst-case performance is a good idea because here we have 2 different issues. But the main idea of this RFC is an attempt to show the better approach to to these kinds of transformations and to suggest to use this approach in the future. At the same time, I'm trying to explain that this patch is not the performance one because the

What is the state of sqrt in LLVM? It was an intrinsic but there are no OCaml bindings for it and, last I looked, it generated inefficient code on Linux due to this bug: http://www.llvm.org/PR3219 Is the intrinsic deprecated? Am I losing a lot of performance by calling sqrt from libm instead of using the intrinsic? -- Dr Jon Harrop, Flying Frog Consultancy Ltd.

On Jan 7, 2010, at 7:06 AM, Jon Harrop wrote: > > What is the state of sqrt in LLVM? > > It was an intrinsic but there are no OCaml bindings for it and, last > I looked, > it generated inefficient code on Linux due to this bug: > > http://www.llvm.org/PR3219 > > Is the intrinsic deprecated? Am I losing a lot of performance by > calling sqrt > from libm

For background on these calls, see <http://lists.llvm.org/pipermail/llvm-dev/2019-September/135087.html>. Reminder: the purpose is to co-ordinate between active contributors. If you have support questions etc then it's best to post to llvm-dev. We have a call every alternate Thursday at 4pm BST, via <https://meet.google.com/ske-zcog-spp>. We have created a shared calendar which

Dear all, I have a client who uses Samba 3.0.28 on AIX 5.3 and he is running into problems with the cpu utilization which is 100% caused by one or more smbd processes. I found out that this problem should be solved in version 3.0.37 and want to know if there is a way to get the bff-file of this version. I am not able to compile samba myself on the machine which has the problem and I am

Hi, We have implemented the rcp instruction generation for X86 target architecture. We have introduced a flag -fp-rcp flag which controls the generatation of X86 rcp instruction generation. We have observed minor effects on precision and hence hve put these transformations under the mentioned flag. Note that -fp-rcp is only enabled with -enable-unsafe-fp-math flag presently. Moreover we have

On 8 December 2016 at 02:34, Jojo Ma <jojo.ma at linaro.org> wrote: > It would be profitable as well if we could enable the canonicalisation on > it. > sequence before this canonicalisation (ARM): > test: > .fnstart > @ BB#0: @ %entry > movw r1, #65534 > and r1, r0, r1 > ubfx r0, r0, #1, #15 > add r0, r0, r1, lsr #1 > bx lr

On 19 December 2016 at 09:58, Jojo Ma <jojo.ma at linaro.org> wrote: > /home/likewise-open/SPREADTRUM/jojo.ma/jojoma/source/llvm/llvm-linaro/llvm/llvm/test/CodeGen/Thumb2/machine-licm.ll:88:10: > error: expected string not found in input > ; CHECK: movw {{(r[0-9])|(lr)}}, #32768 > ^ > <stdin>:56:2: note: scanning from here > movw r12, #32768 > ^ Hi

If that doesn''t work, then you''ve done something wrong somewhere. ''Cause it works for IE for me and everyone else. The only advantage I see with the YUI toolkit is that it''ll defer attaching events to elements that do not yet exist in the DOM. Big whoop, only a sloppy programmer would try to attach an event to a non-existant element. With Prototype, you can

Hi All I'm trying to intergrate samba server with ADS on AIX 6.1 TL04, for last one week, with idmap / winbind but no satisfactory results. I have gone through various links at samba.org relating to winbind, idmapper and followed http://pware.hvcc.edu/ for precompiled binaries and http://pware.hvcc.edu/AIX-Samba.pdf which is for AIX 6.1 TL03 though. I have found the samba which is provided

Hi all, Actually,I found there is a same problem for arm.For this case,I think > maybe we can play a trick in the clang. > Checking whether the given arch valid or not,before we throw it to the > parser,which can be used for both arm > and aarch64. For the actions I mentioned above,I wrote a check function as below, basing on the naming rules of the arm architecture. +//Only if

Thanks very much, Renato! I looked into the regressions, there're many cases have similar context as the case I mentioned but N just has only one use. In this case, the canonicalisation won't make it more profitable, will only prevent the possible folding or make the sequence is not expected. The regressions be eliminated, after I limit the expanding to not "N->hasOneUse()".

Hi everyone, I'm a member engineer of linaro's llvm team,coming from Spreadtrum.I am a new person on LLVM.Now I'm writing a Target Parser for AArch64,so options parsing of AArch64 about cpu & arch & fpu can be summary to one place. In the TargetParser,we assume "aarch64" and "arm64" are synonyms of armv8a(as they are only for armv8a,people usually do

Hi R people, I have a very basic question to ask - I'm sorry if it's been asked before, but I searched the archives and could not find an answer. All the examples I found were much more complicated/nuanced versions of the problem - my question is much more simple. I have data with multiple, nested fixed effects (as I understand it, fixed effects are specified by the experimental design

When I try to join the Win2k machine to a domain, I get the following message: The following error occurred attempting to join the domain "WIN2K": The procedure number is out of range. This is a Win2k PC attempting to join to AIX 4.3.2 running samba 2.0.7. Any ideas would be greatly appreciated. Thanks, Bill -- /------------------------------------------------------\