similar to: [LLVMdev] Question to Chris

Dear Dr.Lattner Hello, Dr.Lattner. You may find your reply at http://lists.cs.uiuc.edu/pipermail/llvmdev/2007-August/010479.html and other your replies to me right up and down at the list. You had suggested me to read the "structural analysis" section in Muchnick's book. Thank you for this. I bought and read it, which was very helpful but... I still don't have any idea

Hi Seung, It should be fairly straight-forward to do in LLVM. Once you identify the loops, then identify the PHI nodes that you need to convert, then apply the transformation below. The fine details on how to create an instruction and replace one instruction with another are documented in the docs section and in other code. :-) One thing to be careful of, if you convert a variable like

Bill, Depending on what Seung's problem is, converting *out* of SSA form may actually be the wrong thing to do. Seung needs to explain exactly problem he is unable solve using SSA form. --Vikram http://www.cs.uiuc.edu/~vadve http://llvm.org/ On Jan 28, 2008, at 1:10 AM, Bill Wendling wrote: > Hi Seung, > > It should be fairly straight-forward to do in LLVM. Once you

Hello, guys. I am trying to construct higher-level 'for' from the low-level LLVM bitcode(ver 1.9). It's partly successful thanks to David A. Greene's advice suggested to use Control Dependence Graph(CDG). I could find which BB contributes to form which loop with CDG. For example, for this simple function: ----------------------------------------------------------- void bsloop(int

On May 22, 2014, at 3:51 PM, Chandler Carruth <chandlerc at google.com> wrote: > > On Thu, May 22, 2014 at 4:42 PM, Louis Gerbarg <lgg at apple.com> wrote: > The problem that the above transform is technically illegal because “When indexing into a (optionally packed) structure, only i32 integer constants are allowed (when using a vector of indices they must all be the same

---------- Forwarded message ---------- From: Ryan Taylor <ryta1203 at gmail.com> Date: Mon, Feb 6, 2012 at 10:36 AM Subject: Re: [LLVMdev] Updating PHI for Instruction Domination? To: Eric Christopher <echristo at apple.com> Since I'm not sure which instructions I might want to replicate, would it be possible to cast the Value from "PHINode::getIncomingValue" to

Dear All, Is there a pass (or set of passes) that will replace a phi whose input operands all compute the same value with an instruction that computes that value? In other words, something that will convert: define internal i32 @function(i32 %x) { ... bb1: %y = add %x, 10 ... bb2: %z = add %x, 10 ... bb3: %phi = [bb1, %y], [bb2, %z] into define internal i32 @function(i32 %x) { ... bb1: ...

I guess not since Value is a superclass of Instruction. On Mon, Feb 6, 2012 at 10:36 AM, Ryan Taylor <ryta1203 at gmail.com> wrote: > > > ---------- Forwarded message ---------- > From: Ryan Taylor <ryta1203 at gmail.com> > Date: Mon, Feb 6, 2012 at 10:36 AM > Subject: Re: [LLVMdev] Updating PHI for Instruction Domination? > To: Eric Christopher <echristo at

You're creating a new path that doesn't include L. For all values defined in L and used outside of L, you need to determine the new reaching def. That's specific to your transformation and can't be automated. Once you do that, creating the phi in F is natural. -Andy On Feb 6, 2012, at 11:51 AM, Ryan Taylor <ryta1203 at gmail.com> wrote: > I guess not since Value is a

On Fri, Jun 28, 2019 at 3:35 PM James Y Knight <jyknight at google.com> wrote: > On Fri, Jun 28, 2019 at 5:53 PM Bill Wendling <isanbard at gmail.com> wrote: > >> On Fri, Jun 28, 2019 at 1:48 PM James Y Knight <jyknight at google.com> >> wrote: >> >>> On Fri, Jun 28, 2019 at 3:00 PM Bill Wendling <isanbard at gmail.com> >>> wrote:

On 01/26/2011 07:50 PM, Vu Le wrote: > I guess I didn't have a clear question. > > Suppose we have BB1 and BB2 both point to BB3. > BB1 has variable x. BB2 also as variable x. > BB3 will have PHINode for x with 2 value from BB1 and BB2. > BB1 BB2 > \ / > BB3 > > Now if BB1 and BB2 is extracted into a function > (using ExtractCodeRegion), they

On Mon, Jul 1, 2019 at 6:25 PM Finkel, Hal J. <hfinkel at anl.gov> wrote: > On 7/1/19 1:38 PM, Bill Wendling via llvm-dev wrote: > > On Fri, Jun 28, 2019 at 3:35 PM James Y Knight <jyknight at google.com> > wrote: > >> On Fri, Jun 28, 2019 at 5:53 PM Bill Wendling <isanbard at gmail.com> wrote: >> >>> On Fri, Jun 28, 2019 at 1:48 PM James Y

Hi John, On 30/07/13 16:12, John Criswell wrote: > Dear All, > > Is there a pass (or set of passes) that will replace a phi whose input operands > all compute the same value with an instruction that computes that value? In > other words, something that will convert: > > define internal i32 @function(i32 %x) { > ... > bb1: > %y = add %x, 10 > ... > bb2: > %z

On 7/30/13 9:46 AM, Duncan Sands wrote: > Hi John, > > On 30/07/13 16:12, John Criswell wrote: >> Dear All, >> >> Is there a pass (or set of passes) that will replace a phi whose >> input operands >> all compute the same value with an instruction that computes that >> value? In >> other words, something that will convert: >> >>

Dear all, I'm working on an exploratory backend on llvm. In the instruction set I'm using I have an instruction (called DECJNZ) that decrements a register and, if the decremented value is not zero, jumps (with a relative jump) to a given offset. I've described in tablegen this instruction as follow: def DECJNZ : Instruction { let Namespace = "MyTarget"; let

Dear Prof.Adve and Bill, I deeply appreciate your comments and concerns. (Please forgive my late response. I've tried some more cases to make this issue) As Prof.Adve mentioned, I need to explain exactly what my problem is, but I have no good ability that I can explain it in this plain text space. For this reason, I made a .pdf file and linked it as follows:

So my best bet is to try and work in reg2mem mode and then go back to mem2reg? I'm curious, it seems though when you split a block that the phis get updated, right? On Thu, Feb 2, 2012 at 5:05 PM, Eric Christopher <echristo at apple.com> wrote: > Not that I'm aware of. > > -eric > > On Feb 2, 2012, at 3:47 PM, Ryan Taylor wrote: > > So essentially I'm

Recently I posted a patch to migrate certain GEPs between basic blocks in cases where doing so would improve the ability of instcombine to merge into more complicated addressing mode (r209049 and r209065). After some build to failures it was rolled back. I now have a patch that no longer causes the regressions I was seeing, but it also no longer can optimize the case I was trying to optimize. As

On Fri, Jun 28, 2019 at 1:48 PM James Y Knight <jyknight at google.com> wrote: > On Fri, Jun 28, 2019 at 3:00 PM Bill Wendling <isanbard at gmail.com> wrote: > >> On Thu, Jun 27, 2019 at 1:44 PM Bill Wendling <isanbard at gmail.com> wrote: >> >>> On Thu, Jun 27, 2019 at 1:29 PM James Y Knight <jyknight at google.com> >>> wrote:

I guess I didn't have a clear question. Suppose we have BB1 and BB2 both point to BB3. BB1 has variable x. BB2 also as variable x. BB3 will have PHINode for x with 2 value from BB1 and BB2. BB1 BB2 \ / BB3 Now if BB1 and BB2 is extracted into a function (using ExtractCodeRegion), they will be replaced by a basic block called codeRepl (which has a call to the extracted