search for: vtmp

Gracias, Carlos. Habia pensado en algo similar usando sapply(): sapply(seq(1, ncol(vtmp), by = 2), function(i) c(rbind(as.character(vtmp[, i]), as.character(vtmp[, i+1])))) Dependiendo de la dimension de los datos, quizas mapply() sea mas eficiente que sapply(). Saludos cordiales, Jorge.- 2015-02-25 1:01 GMT+11:00 Carlos Ortega <cof en qualityexcellence.es>: > Hola, >...

Excelente! Ahora corre muy rápido. No conocía ese método, creo que me va a resultar muy útil. Muchas gracias y saludos. Fernando Macedo El 24/02/15 a las 10:51, Jorge I Velez escribió: Fernando, Podrias intentar R> a <- rep('a', 5) R> b <- rep('b', 5) R> a [1] "a" "a" "a" "a" "a" R> b [1] "b"

V2: clarify code, commit msgs, add comments. Drop code to was supposed to make register assignment prefer SDST == SRC2 (patch 2) for now, because it didn't quite do what I intended.

...er allocation due to +// constraint SDST == SSRC2 +// TODO: +// Does NVC0+ have other situations where this pass makes sense? +class NV50PostRaConstantFolding : public Pass +{ +private: + virtual bool visit(BasicBlock *); +}; + +bool +NV50PostRaConstantFolding::visit(BasicBlock *bb) +{ + Value *vtmp; + Instruction *def; + + for (Instruction *i = bb->getFirst(); i; i = i->next) { + switch (i->op) { + case OP_MAD: + if(i->def(0).getFile() == FILE_GPR && + i->src(0).getFile() == FILE_GPR && + i->src(1).getFile() =...

...er allocation due to +// constraint SDST == SSRC2 +// TODO: +// Does NVC0+ have other situations where this pass makes sense? +class NV50PostRaConstantFolding : public Pass +{ +private: + virtual bool visit(BasicBlock *); +}; + +bool +NV50PostRaConstantFolding::visit(BasicBlock *bb) +{ + Value *vtmp; + Instruction *def; + + for (Instruction *i = bb->getFirst(); i; i = i->next) { + switch (i->op) { + case OP_MAD: + if(i->def(0).getFile() != FILE_GPR || + i->src(0).getFile() != FILE_GPR || + i->src(1).getFile() != FILE_GPR || +...

was "nv50/ir: PostRaConstantFolding improvements" before. nothing really changed from the last version, just minor things. Karol Herbst (5): nv50/ir: restructure and rename postraconstantfolding pass nv50/ir: implement mad post ra folding for nvc0+ gk110/ir: add LIMM form of mad gm107/ir: add LIMM form of mad nv50/ir: also do PostRaLoadPropagation for FMA

MAD IMM has a very specific SDST == SSRC2 requirement, so don't emit Signed-off-by: Roy Spliet <rspliet at eclipso.eu> --- .../drivers/nouveau/codegen/nv50_ir_emit_nv50.cpp | 18 ++++++++++++------ .../drivers/nouveau/codegen/nv50_ir_target_nv50.cpp | 2 +- 2 files changed, 13 insertions(+), 7 deletions(-) diff --git a/src/gallium/drivers/nouveau/codegen/nv50_ir_emit_nv50.cpp

Add emission rules for negative and saturate flags for MAD 4-byte opcodes, and get rid of constraints. Short MAD has a very specific SDST == SSRC2 requirement, and since MAD IMM is short notation + 4-byte immediate, don't have the compiler create MAD IMM instructions yet. V2: Document MAD as supported short form Signed-off-by: Roy Spliet <rspliet at eclipso.eu> ---

...er allocation due to +// constraint SDST == SSRC2 +// TODO: +// Does NVC0+ have other situations where this pass makes sense? +class NV50PostRaConstantFolding : public Pass +{ +private: + virtual bool visit(BasicBlock *); +}; + +bool +NV50PostRaConstantFolding::visit(BasicBlock *bb) +{ + Value *vtmp; + Instruction *def; + + for (Instruction *i = bb->getFirst(); i; i = i->next) { + switch (i->op) { + case OP_MAD: + if (i->def(0).getFile() != FILE_GPR || + i->src(0).getFile() != FILE_GPR || + i->src(1).getFile() != FILE_GPR || +...

...er allocation due to +// constraint SDST == SSRC2 +// TODO: +// Does NVC0+ have other situations where this pass makes sense? +class NV50PostRaConstantFolding : public Pass +{ +private: + virtual bool visit(BasicBlock *); +}; + +bool +NV50PostRaConstantFolding::visit(BasicBlock *bb) +{ + Value *vtmp; + Instruction *def; + + for (Instruction *i = bb->getFirst(); i; i = i->next) { + switch (i->op) { + case OP_MAD: + if (i->def(0).getFile() != FILE_GPR || + i->src(0).getFile() != FILE_GPR || + i->src(1).getFile() != FILE_GPR || +...

Add emission rules for negative and saturate flags for MAD 4-byte opcodes, and get rid of some of the constraints. Obviously tested with a wide variety of shaders. V2: Document MAD as supported short form V3: Split up IMM from short-form modifiers Signed-off-by: Roy Spliet <rspliet at eclipso.eu> --- src/gallium/drivers/nouveau/codegen/nv50_ir_emit_nv50.cpp | 10 ++++------

...0 10 11 8 10 12 > 9 8 8 9 9 9 9 9 5 9 7 13 10 11 8 10 9 10 12 8 12 > [441] 9 10 4 7 11 7 10 4 6 13 8 7 10 9 7 6 8 9 7 11 8 8 9 > 10 5 8 11 12 6 5 10 10 6 10 10 5 10 13 9 13 10 10 6 12 > [485] 8 7 9 12 10 9 7 7 14 6 9 6 6 8 10 6 >> vtmp <- treesize(iris.rf) >> sum(vtmp) > > Por defecto al no especificar nada, el "ntrees" de randomForest() es 500. > Efectivamente generas 500 árboles como ves en el número de elementos que > devuelve "treesize(iris.rf)". > > Y cada árbol, tiene el número...

Gracias Carlos y Javier, ntrees es el nº de árboles y treesize sus respectivos tamaños (nº de nodos) ntree: Number of trees to grow. This should not be set to too small ...... treesize: Size of trees (number of nodes) in and ensemble. Puse 1000 árboles (ntree=1000), si, pero la función treesize te da el nº de nodos: treesize(RFfit, terminal=TRUE) me da un vector de 1000 elementos (uno

...5 9 7 13 10 11 8 10 9 10 12 8 12 >>> [441] 9 10 4 7 11 7 10 4 6 13 8 7 10 9 7 6 8 9 7 11 8 8 9 >>> 10 5 8 11 12 6 5 10 10 6 10 10 5 10 13 9 13 10 10 6 12 >>> [485] 8 7 9 12 10 9 7 7 14 6 9 6 6 8 10 6 >>> >>>> vtmp <- treesize(iris.rf) >>>> sum(vtmp) >>>> >>> >>> Por defecto al no especificar nada, el "ntrees" de randomForest() es 500. >>> Efectivamente generas 500 árboles como ves en el número de elementos que >>> devuelve "treesiz...

...id = 4, l_type = 0, l_whence = 0, l_sysid = 0} #13 0xffffffff805c4945 in exit1 (td=0xfffffe000b0af000, rv=0) at /usr/home/julian/src/freebsd/sys/kern/kern_exit.c:301 id = <value optimized out> p = (struct proc *) 0xfffffe000b181950 nq = <value optimized out> q = (struct proc *) 0x4 vtmp = <value optimized out> ttyvp = <value optimized out> plim = <value optimized out> reason = <value optimized out> #14 0xffffffff805c5d0e in sys_sys_exit (td=<value optimized out>, uap=<value optimized out>) at /usr/home/julian/src/freebsd/sys/kern/kern_exi...