search for: conv22

Displaying 5 results from an estimated 5 matches for "conv22".

Did you mean: conv2
2015 Jan 23
2
[LLVMdev] question about enabling cfl-aa and collecting a57 numbers
...> LICM: Promoting value stored to in loop: i16* getelementptr inbounds ([16 > x i16]* @pA, i64 0, i64 15) > > LICM sinking instruction: %conv32 = trunc i32 %add31 to i16 > > LICM sinking instruction: %add31 = sub nsw i32 %conv18, %conv17 > > LICM sinking instruction: %conv22 = trunc i32 %sub to i16 > > LICM sinking instruction: %sub = sub nsw i32 %conv17, %conv18 > > LICM sinking instruction: %conv19 = trunc i32 %add to i16 > > > > clang -c --target=aarch64-linux-gnu -mcpu=cortex-a57 -Ofast -mllvm > –use-cfl-aa -S test2.c -o test2out...
2015 Jan 24
2
[LLVMdev] question about enabling cfl-aa and collecting a57 numbers
No, i mean the actual store instruction looks like "store i16 %conv22, i16* getelementptr inbounds ([16 x i16]* @pA, i64 0, i64 12), align 2, !tbaa !1" Not that the pointer operand comes from a GEP, but it is a constantexpr, whose opcode is GEP. It sucks that there is such a complex thing to be handled as a store operand directly , but such is life ... CFL-A...
2015 Jan 26
2
[LLVMdev] question about enabling cfl-aa and collecting a57 numbers
...lt; > george.burgess.iv at gmail.com> > > Sent: Friday, January 23, 2015 7:14:24 PM > > Subject: Re: [LLVMdev] question about enabling cfl-aa and collecting a57 > numbers > > > > > > No, i mean the actual store instruction looks like "store i16 > > %conv22, i16* getelementptr inbounds ([16 x i16]* @pA, i64 0, i64 > > 12), align 2, !tbaa !1" > > > > > > Not that the pointer operand comes from a GEP, but it is a > > constantexpr, whose opcode is GEP. > > Ah, yep. We have ConstantExpr analogs for just about every...
2015 Jan 23
2
[LLVMdev] question about enabling cfl-aa and collecting a57 numbers
Works for me On Thu, Jan 22, 2015 at 8:27 PM, Daniel Berlin <dberlin at dberlin.org> wrote: > We should use graph edges, so we can do something better at set build time > :) > > > On Thu Jan 22 2015 at 5:20:46 PM George Burgess IV < > george.burgess.iv at gmail.com> wrote: > >> > Should we be added an edge from the inttoptr to all other pointer >>
2013 Feb 14
1
[LLVMdev] LiveIntervals analysis problem
...end219.i: ; preds = %while.end212.i %sub220.i = sub nsw i32 16526, %call214.i call fastcc void @emdnorm(i16* %arraydecay15.i, i32 %lost.0.i, i32 0, i32 %sub220.i, i32 64, %struct.LDPARMS* %rnd) nounwind %66 = load i16* %incdec.ptr.i.i, align 2, !tbaa !5 %conv223.i = zext i16 %66 to i32 %cmp224.i = icmp eq i32 %nexp.4.lcssa.i, 0 br i1 %cmp224.i, label %if.then284.i, label %if.end227.i if.end227.i: ; preds = %if.end219.i %cmp228.i = icmp slt i32 %nexp.4.lcssa.i, 0 br i1 %cmp228.i, label %if.then230.i, label %if....