search for: unfolding

I have gone ahead and implemented the unfold distance option as mentioned in the source. the patch is here http://home.comcast.net/~moppsy/compiz/cube-unfold-distance.patch

Hello, I am looking for a loop unfolding procedure implemented in LLVM that helps to transform a while-loop to n-layer If-statements. The transformation should be on IR, although the example below is illustrated on the source level. original loop: * WHILE (condition) DO action ENDWHILE* Expected unfolded loop (2-layer): *...

Here's the code and compilation steps: #include <stdint.h> typedef unsigned int uint128_t __attribute__((mode(TI))); typedef struct{ uint64_t l[5]; } s; void f(s * restrict r, const s * restrict x, const s * restrict y) { uint128_t t[5] = {0, 0, 0, 0, 0}; #define BODY(i,j) { int i_ = i < j ? i : j; int j_ = i < j ? j : i; uint128_t m = (uint128_t) x->l[i_] *

Dear all, I am using the function ?unfold? from the ?RcmdrPlugin.survival? to convert my time-varying covariates dataset from wide to long. I managed to have it working for my data. However, the problem I have is that the observations after an event, won?t be dropped from the dataset. For example, see the dataframe below: the event occurs at 1.2 (event.time=1), but the 1.3 to 1.6 will remain in

Hello All, I am trying to use the unfold function in RcmdrPlugin.survival library, which converts the survival data with time varying covariates to the counting process notation. The problem is somehow, the event indicator created is not correct. Below is the data, I am trying to convert: CASE TRT FAILTIME FAILCENS SEX AGE IGG0 IGG28 IGG42 IGG84 IGG364 26003 A 11.2033

...t;getName()); Value *RCast = Builder->CreateZExt(RHS, CI.getType(), RHS->getName()); return BinaryOperator::Create(Instruction::Or, LCast, RCast); } } ``` The first two lines do the actual `zext(icmp)` optimization. The remaining lines have been added by r48715 and do the mentioned unfolding. However, in `foldCastedBitwiseLogic()` there happens a reverse of the above unfolding, which transforms `logic(cast(A), cast(B))` to `cast(logic(A, B))`: ``` if ((!isa<ICmpInst>(Cast0Src) || !isa<ICmpInst>(Cast1Src)) && shouldOptimizeCast(Cast0) && shouldOptimizeCa...

I'm trying to put together an R routine to conduct unidimensional unfolding scaling analysis using maximum likelihood. My problem is that ML optimization will get stuck at latent scale points that are far from optimal. The point optimizes on one of the observed variables but not others and for ML to move away from this 'local optimum', it has to move in a directi...

...nd to i8 ret i8 %zext } ``` That means as long as the `zext` in zext-or-icmp.ll isn't pushed in front of the `icmp` operations it will not be possible to remove one of them via `transformZExtICmp()` to get to something like `@goo`. So if I'm not mistaken we would still be in need of this unfolding for `zext-or-icmp.ll` if we want it to be optimized equivalently to `foo`, which we would also have to take into account when adapting it for DeMorgan. In general, and independently from my concerns above, I would think that when we want to get from `(((a & 1) == 0) | (b == 0))` to `!(((a &amp...

Hi, SystemZ supports @llvm.ctlz.i64() natively with a single instruction (FLOGR), and lesser bitwidth versions of the intrinsic are promoted to i64. For some reason, this leads to unfolded additions of constants as shown below: This function: define i16 @fun(i16 %arg) {   %1 = tail call i16 @llvm.ctlz.i16(i16 %arg, i1 false)   ret i16 %1 } ,gives this optimized DAG as input to instruction

Hello, I used R a lot one year ago and now I am a bit rusty :) I have my raw data which correspond to the list of runtimes per minute (minute "1" "2" "3" in two database modes "sharding" and "query" and two workload types "query" and "refresh") and as a list of char arrays that looks like this: > str(data) List of 122 $ :

Hi, I am attempting to modify LLVM to generate code for an architecture which is nearly identical to X86-64, but with a few minor differences. In particular, "call" must always have a register operand, and cannot have a memory operand. Any ideas on how I can express this rule? Thanks, - David -------------- next part -------------- An HTML attachment was scrubbed... URL:

On Tue, Jun 8, 2010 at 2:05 PM, David Meyer <pdox at google.com> wrote: > Hi, > I am attempting to modify LLVM to generate code for an architecture which is > nearly identical to X86-64, but with a few minor differences. > In particular, "call" must always have a register operand, and cannot have a > memory operand. Any ideas on how I can express this rule? Just get

...ext > } > ``` > > That means as long as the `zext` in zext-or-icmp.ll isn't pushed in front of the `icmp` operations it will not be possible to remove one of them via `transformZExtICmp()` to get to something like `@goo`. So if I'm not mistaken we would still be in need of this unfolding for `zext-or-icmp.ll` if we want it to be optimized equivalently to `foo`, which we would also have to take into account when adapting it for DeMorgan. > > In general, and independently from my concerns above, I would think that when we want to get from `(((a & 1) == 0) | (b == 0))` to `...

I was always annoyed that you couldn't terminate an action with dbus so I have written this patch which fixes that. Hopefully this can be included. I wrote the patch trying to repeat as little code as possible so I just added a boolean parameter to the dbusHandleActivateMessage and changed the name to dbusHandleActivateDeactivateMessage. You could probably change the deactivate parameter

On Tue, Jun 8, 2010 at 4:20 PM, David Meyer <pdox at google.com> wrote: > Hi Eli, > I have tried this, but the resulting tool-chain was broken. > There are only two references to "CALL64m": the definition in > X86Instr64bit.td, and an entry in X86InstrInfo.cpp. > After commenting both out, compilation of a large application fails with: > llc: ScheduleDAG.cpp:462:

...mal point on the other side of the normal curve. > > > > SANN might work, but I kind of wonder how useful it'd be in > estimating > > hundreds of parameters--thanks to that latent scale. > > > > My (possibly harebrained) thought for how to estimate this > unfolding using > > some gradient-based method would be to run through some > iterations and then > > check to see whether a better solution exists on the 'other > side' of the > > normal curves. If it does, replace those parameters with > the better ones. > > Bec...

After removing CALL64m, the resulting DAG has a cycle that cannot be scheduled. I've attached a PDF of the DAG before instruction selection (-view-isel-dags), and after instruction select (-view-sched-dags). Notice how the flag/chain relationships between MOV64rm and CALL64r make it impossible to schedule. Here's the code being compiled: define ccc void @ArgsFree() nounwind { entry:

Hi Eli, I have tried this, but the resulting tool-chain was broken. There are only two references to "CALL64m": the definition in X86Instr64bit.td, and an entry in X86InstrInfo.cpp. After commenting both out, compilation of a large application fails with: llc: ScheduleDAG.cpp:462: void llvm::ScheduleDAGTopologicalSort::InitDAGTopologicalSorting(): Assertion `Node2Index[SU->NodeNum]

Hi, I find that LSR is not helping enough on avoiding unfoldable offsets for SystemZ. When the loop has three stores with unfoldable offsets, LSR rewrites the IV in a good way. However, if adding another store with a foldable offset that fits already, LSR fails to rewrite the three stores. And if I happen to add a too big *positive* offset (the first three were negative) instead of a

i'm somehow embarrassed to even ask this, but is there any built-in method for doing this: my_list <- list() my_list[[1]] <- matrix(1:20, ncol = 5) my_list[[2]] <- matrix(20:1, ncol = 5) now, knowing that these matrices are identical in dimension, i'd like to unfold the list to a 2x4x5 (or some other permutation of the dim sizes) array. i know i can initialize the array, then