pcordes

// https://github.com/bittnkr/uniq/issues/2

// compile with g++ -m32 -O1 -pthread -Wall
// based on sandbox/hello.cpp but parameterized for qtype instead of int
// and queuing values that have non-zero high halves (and aren't all the same as each other!!!)

#include <pthread.h>
#include <stdio.h>
#include <thread>
#include <vector>

pcordes

    ;; 3 versions, only one enabled with %if 1  NASM/YASM preprocessor stuff, like C #if 0 / #if 1
    ;; shuffle on the fly
    ;; copy + in-place
    ;; read-only up/down/bidir

;;; ~/bin/asm-link loop-up-down.asm && disas loop-up-down && ocperf.py stat -etask-clock,page-faults,cycles,L1-dcache-loads,L1-dcache-load-misses,LLC-loads,LLC-load-misses,instructions,dtlb_store_misses.miss_causes_a_walk  -r4 ./loop-up-down

;;;~/bin/asm-link loop-up-down.asm && rm loop-up-down.o && disas loop-up-down && nice ocperf.py stat -etask-clock,page-faults,cycles,L1-dcache-loads,LLC-loads,LLC-load-misses,instructions,dtlb_store_misses.miss_causes_a_walk,dtlb_load_misses.stlb_hit,dtlb_load_misses.miss_causes_a_walk -r3 ./loop-up-down

default rel

pcordes

// x86 SIMD string to uppercase
// See http://stackoverflow.com/questions/735204/convert-a-string-in-c-to-upper-case

#include <stdio.h>
#include <stdint.h>
#include <string.h>
#include <strings.h>  // for ffs
#include <ctype.h>
#include <immintrin.h>