search for: afterloop

...ll ptx_device i32 @llvm.nvvm.read.ptx.sreg.ctaid.x() %PositionOfBlockInGrid.x = shl i32 %ctaid.x, 9 %BlockLB.Add.ThreadPosInBlock.x = add i32 %PositionOfBlockInGrid.x, %tid.x %isThreadLBgtLoopUB.x = icmp sgt i32 %BlockLB.Add.ThreadPosInBlock.x, 65535 br i1 %isThreadLBgtLoopUB.x, label %CUDA.AfterLoop.x, label %CUDA.LoopHeader.x.preheader CUDA.LoopHeader.x.preheader: ; preds = %"Loop Function Root" %1 = sext i32 %BlockLB.Add.ThreadPosInBlock.x to i64 store float 0.000000e+00, float* inttoptr (i64 47380979712 to float*), align 8192, !tbaa !0 %p_.moved.to.4.cl...

Dear all, Attached notunrolled.ll is a module containing reduction kernel. What I'm trying to do is to unroll it in such way, that partial reduction on unrolled iterations would be performed on register, and then stored to memory only once. Currently llvm's unroller together with all standard optimizations produce code, which stores value to memory after every unrolled iteration, which is

...hen11 %storemerge26 = phi double [ %phitmp, %then11 ], [ 1.000000e+01, %endif.endif15_crit_edge ] ; <double> [#uses=1] %lsr.iv.next = add i32 %lsr.iv, 1 ; <i32> [#uses=2] %exitcond = icmp eq i32 %lsr.iv.next, 10001 ; <i1> [#uses=1] br i1 %exitcond, label %afterloop, label %loop afterloop: ; preds = %endif15 %r_fadd19 = fadd double %anotherFloat.1, %storemerge26 ; <double> [#uses=1] %phitmp32 = fcmp ogt double %r_fadd19, 3.535200e+03 ; <i1> [#uses=1] %storemerge27 = zext i1 %phitmp32 to i32 ; &lt...

...erge26 = phi double [ %phitmp, %then11 ], [ 1.000000e+01, %endif.endif15_crit_edge ] ; <double> [#uses=1] > %lsr.iv.next = add i32 %lsr.iv, 1 ; <i32> [#uses=2] > %exitcond = icmp eq i32 %lsr.iv.next, 10001 ; <i1> [#uses=1] > br i1 %exitcond, label %afterloop, label %loop > > afterloop: ; preds = %endif15 > %r_fadd19 = fadd double %anotherFloat.1, %storemerge26 ; <double> [#uses=1] > %phitmp32 = fcmp ogt double %r_fadd19, 3.535200e+03 ; <i1> [#uses=1] > %storemerge27 = zext i1 %phi...

...fork label %task, label %latch task: %aptr = getelementptr i32, i32* %A, i32 0, i32 %i %aval = load i32* %aptr %cptr = getelementptr i32, i32* %C, i32 0, i32 %i store i32 %aval, i32* %aptr halt label %latch latch: %inc = add i32, i32 %i, i32 1 br label %header exit: join label %afterloop afterloop: ... (2) Reasoning: The proposed approach is crafted such that the semantics of the parallel program is represented correctly in almost native, low-level IR right after front-end and preserved at any point till the final lowering to sequential IR or parallel runtime library calls. To...

...adly get 3.000000 instead of 2.000000. This happens, I believe, because the instruction %faddtmp = fadd *double* %x1.0, 1.000000e+000 ; <*double*> [#uses=2] is being generated before %ltcmptmp = fcmp ult *double* %c.0, 2.000000e+000 ; <i1> [#uses=1] br i1 %ltcmptmp, label %loop, label %afterloop and therefore the loop body is emited first and only afterwards we determine whether the loop should exit. I was wondering if this is the intended behaviour, since the fibi(x) example in chapter 7 uses this extra loop to return correct values for fibonacci numbers, or perhaps a known bug in Kaleid...

...= load i32* %aptr > > %cptr = getelementptr i32, i32* %C, i32 0, i32 %i > > store i32 %aval, i32* %aptr > > halt label %latch > > > > latch: > > %inc = add i32, i32 %i, i32 1 > > br label %header > > > > exit: > > join label %afterloop > > > > afterloop: > > ... > > > > > > > > (2) Reasoning: > > The proposed approach is crafted such that the semantics of the parallel > > program is represented correctly in almost native, low-level IR right > > after front-end and pr...

...%cptr = getelementptr i32, i32* %C, i32 0, i32 %i >>> store i32 %aval, i32* %aptr >>> halt label %latch >>> >>> latch: >>> %inc = add i32, i32 %i, i32 1 >>> br label %header >>> >>> exit: >>> join label %afterloop >>> >>> afterloop: >>> ... >>> >>> >>> >>> (2) Reasoning: >>> The proposed approach is crafted such that the semantics of the >>> parallel program is represented correctly in almost native, >>> low-level IR...

...32, i32* %C, i32 0, i32 %i >>>> store i32 %aval, i32* %aptr >>>> halt label %latch >>>> >>>> latch: >>>> %inc = add i32, i32 %i, i32 1 >>>> br label %header >>>> >>>> exit: >>>> join label %afterloop >>>> >>>> afterloop: >>>> ... >>>> >>>> >>>> >>>> (2) Reasoning: >>>> The proposed approach is crafted such that the semantics of the parallel >>>> program is represented correctly in alm...

...2 %i > >>> store i32 %aval, i32* %aptr > >>> halt label %latch > >>> > >>> latch: > >>> %inc = add i32, i32 %i, i32 1 > >>> br label %header > >>> > >>> exit: > >>> join label %afterloop > >>> > >>> afterloop: > >>> ... > >>> > >>> > >>> > >>> (2) Reasoning: > >>> The proposed approach is crafted such that the semantics of the > >>> parallel program is represented correctly...

...t;>>>> halt label %latch >>>>>> >>>>>> latch: >>>>>> %inc = add i32, i32 %i, i32 1 >>>>>> br label %header >>>>>> >>>>>> exit: >>>>>> join label %afterloop >>>>>> >>>>>> afterloop: >>>>>> ... >>>>>> >>>>>> >>>>>> >>>>>> (2) Reasoning: >>>>>> The proposed approach is crafted such that the semantics of the >&...

...l, i32* %aptr > > >>> halt label %latch > > >>> > > >>> latch: > > >>> %inc = add i32, i32 %i, i32 1 > > >>> br label %header > > >>> > > >>> exit: > > >>> join label %afterloop > > >>> > > >>> afterloop: > > >>> ... > > >>> > > >>> > > >>> > > >>> (2) Reasoning: > > >>> The proposed approach is crafted such that the semantics of the > > >>&g...

...t;>>>> halt label %latch >>>>>> >>>>>> latch: >>>>>> %inc = add i32, i32 %i, i32 1 >>>>>> br label %header >>>>>> >>>>>> exit: >>>>>> join label %afterloop >>>>>> >>>>>> afterloop: >>>>>> ... >>>>>> >>>>>> >>>>>> >>>>>> (2) Reasoning: >>>>>> The proposed approach is crafted such that the semantics of the >...

...gt;>>> >>>>>>>> latch: >>>>>>>> %inc = add i32, i32 %i, i32 1 >>>>>>>> br label %header >>>>>>>> >>>>>>>> exit: >>>>>>>> join label %afterloop >>>>>>>> >>>>>>>> afterloop: >>>>>>>> ... >>>>>>>> >>>>>>>> >>>>>>>> >>>>>>>> (2) Reasoning: >>>>>>>> T...