[B! SIMD] yuroyoroã®ãƒ–ãƒƒã‚¯ãƒžãƒ¼ã‚¯

yuroyoro id:yuroyoro

SIMDã«é–¢ã™ã‚‹yuroyoroã®ãƒ–ãƒƒã‚¯ãƒžãƒ¼ã‚¯ (6)

${{author_name}}$

{{{comment_expanded}}}

{{label}}

{{#is_bookmark}}ãƒªã‚¹ãƒˆ{{/is_bookmark}}{{^is_bookmark}}ãƒªãƒ³ã‚¯{{/is_bookmark}}

${{author_name}}$
{{author_name}}{{created}}
{{ #comment }}{{ comment }}{{ /comment }}
- {{ label }}

${{author_name}}$

{{{comment_expanded}}}

{{label}}

{{#is_bookmark}}ãƒªã‚¹ãƒˆ{{/is_bookmark}}{{^is_bookmark}}ãƒªãƒ³ã‚¯{{/is_bookmark}}

SSEã¨AVXã§é«˜æ¬¡å…ƒãƒ™ã‚¯ãƒˆãƒ«ã®å†…ç©è¨ˆç®—ã‚’é«˜é€ŸåŒ–ã—ã¦ã¿ãŸ | ã•ã‹ãªå‰ç·š
ä¸–ç•Œæœ€é€Ÿã®ãŠéšã¨è¨€ãˆã°ã‚«ã‚¸ã‚é¡žã§ï¼Œæ³³ãé€Ÿåº¦ã¯æ™‚é€Ÿ100km/hã‚’è¶…ãˆã‚‹ã¨ã‹ï¼Ž55ãƒŽãƒƒãƒˆç¨‹ã«ãªã‚‹ã®ã§ã“ã‚Œã¯Mk-48éšé›·ã«ã‚‚åŒ¹æ•µã™ã‚‹ã»ã©ã§ã™ï¼Ž ä¸€æ–¹ã¡ã‚‡ã£ã¨ãƒãƒ¼ãƒˆãªé«˜é€ŸãŠéšã¨ã—ã¦ã¯ï¼ŒãŠé¦´æŸ“ã¿ãƒˆãƒ“ã‚¦ã‚ªãŒé£›è¡Œä¸ã«æœ€å¤§70km/hã»ã©ã«é”ã™ã‚‹ã¨ã‹ï¼Ž ä»Šæ—¥ã¯ãã‚“ãªè‹¥å¹²ãƒãƒ¼ãƒˆãªé«˜é€ŸåŒ–ã®ãŠè©±ï¼ˆï¼Ÿï¼‰ã¨ã„ã†ã“ã¨ã§ï¼ŒSSEçµ„ã¿è¾¼ã¿å‘½ä»¤ã«ã¤ã„ã¦ï¼Ž SSEã‚„AVXã¨ã„ãˆã°ãŠé¦´æŸ“ã¿SIMDå‘½ä»¤ã§ï¼Œãã‚Œã‚’ãƒ—ãƒã‚°ãƒ©ãƒ ã‹ã‚‰æ§‹é€ ä½“ã¨é–¢æ•°ã®å½¢å¼ã§é«˜ç§»æ®–ã«è¨˜è¿°ã™ã‚‹æ–¹æ³•ãŒSIMDçµ„ã¿è¾¼ã¿é–¢æ•°(SIMD Intrinsic)ãªã‚ã‘ã§ã™ãŒï¼Œã“ã‚Œã‚’ä½¿ã£ã¦ã”ãå…¸åž‹çš„ãªãƒ™ã‚¯ãƒˆãƒ«ã®å†…ç©è¨ˆç®—ã‚’é«˜é€ŸåŒ–ã—ã¦ã¿ã¾ã—ãŸï¼Ž ãƒ™ã‚¯ãƒˆãƒ«ã®å†…ç©ã®é«˜é€ŸåŒ–ã¨è¨€ãˆã°æ˜Ÿã®æ•°ã»ã©ã‚‚ã•ã‚Œã¦ã‚‹è©±ãªã‚ã‘ã§ï¼Œã„ã¾ã•ã‚‰éšã®æƒ…å ±ãªã‚“ã‹å½¹ã«ç«‹ã¤æ°—ã¯å…¨ãã—ãªã„ã®ã§ã™ãŒï¼Œç´”ç²‹ã«è‡ªåˆ†ã§ã‚„ã‚‰ãªã„ã¨ã‚ã‹ã‚“ãªã„ï¼žï¼œã€€ã¨ã„ã†ã“ã¨ã§ï¼Œ ã¨ã«ã‹ãæ›¸ã„ã¦ã¿ã‚ˆã† åŠ¹æžœã®ç¨‹ã¯ã„ã‹ã«ï¼Ÿ ã‚’èª¿ã¹ã¦ã¿ãŸãã¦ï¼Œã‚„
yuroyoro 2015/06/25
c++

SIMD
ãƒªãƒ³ã‚¯
AVX ã§éŠã‚“ã§ã¿ã‚‹ - Algorithmerâ€™s note
ä»Šå¹´ã®ï¼‘æœˆã«ç™ºè¡¨ã•ã‚ŒãŸIntelã®æ–°CPUã€ŒSandy Bridgeã€ã‹ã‚‰æè¼‰ã•ã‚ŒãŸæ–°æ©Ÿèƒ½ã«ã€ã€ŒAVXã€ã¨ã„ã†ã‚‚ã®ãŒã‚ã‚Šã¾ã™ã€‚AVXã‚’ç”¨ã„ã‚‹ã¨ã€256ãƒ“ãƒƒãƒˆã®ãƒ¬ã‚¸ã‚¹ã‚¿ã‚’ç”¨ã„ã¦ã€ä¾‹ãˆã°doubleåž‹ã®è¨ˆç®—ã‚’4ã¤åŒæ™‚ã«è¡Œã†ãªã©é«˜é€Ÿã«æ¼”ç®—ã‚’ã™ã‚‹ã“ã¨ãŒã§ãã¾ã™ã€‚ AVXã‚’ä½¿ã†ãŸã‚ã«ã¯ã€å°‚ç”¨ã®ã‚³ãƒ¼ãƒ‰ã‚’æ›¸ã‹ãªã‘ã‚Œã°ã„ã‘ã¾ã›ã‚“ãŒã€C++ã§ã¯å˜ä½“ã§ï¼ˆã‚¢ã‚»ãƒ³ãƒ–ãƒ©ã‚’ç”¨ã„ãšã«ï¼‰AVXå¯¾å¿œã®ã‚³ãƒ¼ãƒ‰ã‚’æ›¸ãã“ã¨ãŒã§ãã‚‹ã®ã§ã€ã„ã‚ã„ã‚éŠã‚“ã§ã¿ã¾ã—ãŸã€‚ ã‚³ãƒ¼ãƒ‰ã®ä½œæ³• ãƒ»#include "immintrin.h" ã‚’è¨˜è¿° AVXã‚’ä½¿ã†ãŸã‚ã®é–¢æ•°ãŒã“ã“ã«è¨˜è¿°ã•ã‚Œã¦ã„ã‚‹ã®ã§ã€includeã—ã¾ã™ã€‚ ãƒ»å¤‰æ•°ã®alignment å¤‰æ•°ã®å‰ã«ã€Œ_declspec(align(32))ã€ã¨ã„ã†ãŠã¾ã˜ãªã„ã‚’æ›¸ã„ã¦ãŠã‘ã°ã€è‡ªå‹•ã§å¢ƒç•ŒãŒ32ãƒã‚¤ãƒˆå¢ƒç•Œã«ãªã‚‹ã‚ˆã†ã«èª¿æ•´ã•ã‚Œã¾ã™ã€‚ã“ã®ãŠã¾ã˜ãªã„ãŒã‚ã‚‹ã¨ã€ã‚ã‚Šã¨ä½•ã‚‚è€ƒãˆãªãã¦ã‚‚AVXå‘½ä»¤ã‚’ä½¿ãˆã‚‹
yuroyoro 2015/06/25
SIMD
ãƒªãƒ³ã‚¯
ãƒ¡ãƒ¢ãƒªã‚¢ãƒ©ã‚¤ãƒ¡ãƒ³ãƒˆã‚’æƒãˆãšã« SIMD ã™ã‚‹æ–¹æ³• - kawa0810 ã®ãƒ–ãƒã‚°
ãƒ¡ãƒ¢ãƒªã‚¢ãƒ©ã‚¤ãƒ¡ãƒ³ãƒˆã‚’æƒãˆãšã« SIMD æ¼”ç®—ã™ã‚‹æ–¹æ³• ä»Šã¾ã§ã¯ _mm_malloc() ã‚’ä½¿ç”¨ã—ã¦ãƒ¡ãƒ¢ãƒªã‚¢ãƒ©ã‚¤ãƒ¡ãƒ³ãƒˆã‚’ 32byte å¢ƒç•Œã«æƒãˆã¦ã‹ã‚‰ AVX å‘½ä»¤ã§ SIMD æ¼”ç®—ã‚’ã—ã¦ã„ã¾ã—ãŸãŒï¼Œãƒ¡ãƒ¢ãƒªã‚¢ãƒ©ã‚¤ãƒ¡ãƒ³ãƒˆã‚’æ°—ã«ã›ãšã« AVX ã‚’ç”¨ã„ã¦ SIMD æ¼”ç®—ã™ã‚‹æ–¹æ³•ã§ã™ï¼Ž ãƒ‡ãƒ¼ã‚¿ã‚’ 32byte å¢ƒç•Œã«æƒãˆã¦ã„ã‚‹å ´åˆã¯ _mm_load_ps ã‚„ _mm_load_pd ãŒæš—é»™çš„ã«ä½¿ã‚ã‚Œãƒ‡ãƒ¼ã‚¿ã‚’ãƒãƒ¼ãƒ‰ã—ã¾ã™ï¼Žè¨ˆç®—ã—ãŸã„ãƒ‡ãƒ¼ã‚¿ãŒ 32byte å¢ƒç•Œã«æƒãˆã¦ã„ãªã„å ´åˆã¯ _mm_loadu_ps ã‚„ _mm_storeu_ps ç‰ã‚’æ˜Žç¤ºçš„ã«ä½¿ç”¨ã™ã‚‹ã“ã¨ã§ SIMD æ¼”ç®—ã‚’ä½¿ç”¨ã™ã‚‹ã“ã¨ãŒå¯èƒ½ã«ãªã‚Šã¾ã™ï¼ŽãŸã ã—ï¼Œã‚ªãƒ¼ãƒãƒ¼ãƒ˜ãƒƒãƒ‰ãŒç™ºç”Ÿã™ã‚‹ãŸã‚è¨ˆç®—ãŒé…ããªã‚‹ãã†ã§å¯èƒ½ãªé™ã‚Šãƒ‡ãƒ¼ã‚¿ã¯ 32byte å¢ƒç•Œã«æƒãˆã¾ã—ã‚‡ã†ï¼Ž ç·å’Œè¨ˆç®—ã®ã‚µãƒ³ãƒ—ãƒ«ã‚³ãƒ¼ãƒ‰ ãƒ¡ãƒ¢ãƒªã‚¢ãƒ©ã‚¤ãƒ¡ãƒ³ãƒˆã‚’æ°—ã«ã›ãšã«ç·å’Œè¨ˆç®—ã‚’ã™ã‚‹ã‚µãƒ³ãƒ—
yuroyoro 2015/06/25
SIMD
ãƒªãƒ³ã‚¯
gccã§intrinsicsã§SSEã§ãƒ™ã‚¯ã‚¿ãƒ©ã‚¤ã‚ºã™ã‚‹æ™‚ã®ç°¡æ˜“çš„ãªã¾ã¨ã‚ - ã¬ã†ã±ã‚“ã®å‚™å¿˜éŒ²
ã“ã®è¨˜äº‹ã¯ intrinsicsã‚’ä½¿ã£ã¦SSEã§ãƒ™ã‚¯ã‚¿ãƒ©ã‚¤ã‚ºã™ã‚‹ã®ã«å¿…è¦ãªè¶³ãŒã‹ã‚Šã‚’è‡ªåˆ†ç”¨ã«ã¾ã¨ã‚ãŸã‚‚ã®ã§ã™ã€‚ ãã‚‚ãã‚‚SSEã£ã¦ä½• SIMD(http://ja.wikipedia.org/wiki/Streaming_SIMD_Extensions)ã‚’å®Ÿç¾ã™ã‚‹æ‹¡å¼µå‘½ä»¤ã‚»ãƒƒãƒˆã®åå‰ã€‚ã„ã‚ã„ã‚ç¨®é¡žãŒã‚ã‚‹ã—ãƒãƒ¼ã‚¸ãƒ§ãƒ³ã‚‚ã‚ã‚‹ã€‚åŸºæœ¬çš„ãªè€ƒãˆæ–¹ã‚’ã™ã”ãä¹±æš´ã«è¨€ã†ã¨ã€floatã®æŽ›ã‘ç®—ã‚’ï¼”å›žç¹°ã‚Šæ›¿ãˆã™ã‚ˆã‚Šã‚‚4ã¤ã®floatã®æŽ›ã‘ç®—ã‚’ä¸€æ¯ã«ã‚„ã£ãŸã‚‰æ—©ããï¼Ÿã€€ã¨ã„ã†æ„Ÿã˜ã€‚ intrinsicsã£ã¦ä½• å‰è¿°ã®é€šã‚Šã€SSEã¨ã‹AVXã¯CPUã®å‘½ä»¤ãªã®ã§ç›´æŽ¥ä½¿ãŠã†ã¨æ€ã†ã¨è‡ªåˆ†ã§ã‚¢ã‚»ãƒ³ãƒ–ãƒ©ã‚’æ›¸ã‹ãªãã¦ã¯ã„ã‘ãªã„ã€‚ãŒã€æ—¢å˜ã®C++ã‚³ãƒ¼ãƒ‰ã‚’ä½•ã¨ã‹ã—ãŸã„ã¨ã„ã†ã®ãŒèª¿ã¹ã¦ã‚‹å‹•æ©Ÿãªã®ã§gccã®SIMD intrinsicsã‚’ä½¿ã†ã“ã¨ã«ã™ã‚‹ã€‚ã“ã®SIMD intrinsicsã¯ã‚³ãƒ³ãƒ‘ã‚¤ãƒ©ãƒ¼ã®çµ„ã¿è¾¼ã¿é–¢æ•°ã§ã€ã“ã„ã¤ã‚‰ã‚’ä½¿ã†ã“
yuroyoro 2015/06/24
SIMD

c++
ãƒªãƒ³ã‚¯
çµ„ã¿è¾¼ã¿é–¢æ•°ï¼ˆintrinsicï¼‰ã«ã‚ˆã‚‹SIMDå…¥é–€
6. ç›®æ¨™ç›®æ¨™ OpenCVã«ã‚ˆã‚Šã‚‚é«˜é€Ÿãªãƒ—ãƒã‚°ãƒ©ãƒ ã‚’ç°¡OpenCVã«ã‚ˆã‚Šã‚‚é«˜é€Ÿãªãƒ—ãƒã‚°ãƒ©ãƒ ã‚’ç°¡ å˜ã«ã¤ãã‚Œã‚‹æ›¸ãæ–¹ã‚’è¦šãˆã‚ˆã†å˜ ãã‚Œã‚‹æ›¸ãæ–¹ã‚’è¦šãˆã‚ˆã† å‡ºæ¥ã‚‹ã‚ˆã†ã«ãªã‚‹ã“ã¨ ãƒ»é«˜é€Ÿã‚³ãƒ”ãƒ¼ src.copyTo(dest);é«˜é€Ÿã‚³ãƒ” src.copyTo(dest); ãƒ»é«˜é€ŸãªåŠ æ¸›ä¹—ç®— dest =Â src1Â +Â src2; ãƒ»é«˜é€Ÿåž‹å¤‰æ› src.convert(dest,CV_16U); ãƒ»é«˜é€ŸBGR2RGB cvtColor(src,dest,CV_BGR2RGB); ãƒ»é«˜é€ŸBGR2Gray cvtColor(src,dest,CV_BGR2Gray); â€»å‡ºæ¥ãªã„ã“ã¨ æ•´æ•°æ¼”ç®—ã ã‘ æµ®å‹•å°æ•°ç‚¹ã®å‘½ä»¤ã¯ä½¿ã‚ãªã„ æ¯”è¼ƒæ¼”ç®—ã¯ä½¿æ•´æ•°æ¼”ç®—ã ã‘ï¼Œæµ®å‹•å°æ•°ç‚¹ã®å‘½ä»¤ã¯ä½¿ã‚ãªã„ï¼Œæ¯”è¼ƒæ¼”ç®—ã¯ä½¿ ã‚ãªã„ï¼Œæ°´å¹³æ¼”ç®—ã¯ä½¿ã‚ãªã„ï¼ŒAVXã¯å«ã¾ãªã„(SSE4.1ã¾ã§) 7. ç›®çš„ç›®çš„ ã‚’ä½¿ãˆã° åŒ–ã¯é›£ã—ããªã„ã‚ˆ
yuroyoro 2015/06/24
SIMD

c++
ãƒªãƒ³ã‚¯
introdunction to SIMD programming - primitive: blog
Unite 2015 Tokyo ã®è¬›æ¼”ã§è©³ç´°ã‚’è©±ã›ãªã‹ã£ãŸã®ãŒå¿ƒæ®‹ã‚Šã ã£ãŸã®ã§ã€å¤§é‡ã®ã‚ªãƒ–ã‚¸ã‚§ã‚¯ãƒˆã®æ›´æ–°å‡¦ç†ã«ã¤ã„ã¦ã“ã®å ´ã§æ›¸ã„ã¦ã¿ã¾ã™ã€‚ ä¸»ã« C++ ã§ã€ç°¡å˜ãªãƒ‘ãƒ¼ãƒ†ã‚£ã‚¯ãƒ«ã‚¨ãƒ³ã‚¸ãƒ³ã‚’ä½œã‚Šã€ãã‚Œã‚’ SIMD ã‚’ç”¨ã„ã¦é«˜é€ŸåŒ–ã™ã‚‹æ‰‹é †ã‚’è§£èª¬ã—ã¾ã™ã€‚ è©±ã‚’ç°¡å˜ã«ã™ã‚‹ãŸã‚ã€ä»¥ä¸‹ã®å‰æã‚’è¨ã‘ã¾ã™ã€‚ ãƒ»x86 ç’°å¢ƒã®ã¿è€ƒæ…® ãƒ»ãƒ‘ãƒ¼ãƒ†ã‚£ã‚¯ãƒ«ã¯ä½ç½®ã¨é€Ÿåº¦ã®ã¿ã‚’ä¿æŒ ãƒ»ãƒ‘ãƒ¼ãƒ†ã‚£ã‚¯ãƒ«åŒå£«ã®ç›¸äº’è¡çªã¯ç·å½“ãŸã‚Šã§è¨ˆç®— ç·å½“ãŸã‚Šãªã®ã§è¶…é…ã„ã§ã™ãŒã€å®Ÿè£…ã¯å®¹æ˜“ã§ SIMD ã«ã‚ˆã‚‹æ©æµã‚’å—ã‘ã‚„ã™ãã€é¡Œæã¨ã—ã¦æ‰‹é ƒã§ã™ã€‚ ã“ã®è¨˜äº‹ã®ä¸ã§å¼•ç”¨ã•ã‚Œã¦ã„ã‚‹ã‚½ãƒ¼ã‚¹ã®å…ƒã¯ ã“ã¡ã‚‰ã€ãƒ“ãƒ«ãƒ‰çµæžœ (ä¸Šã®ã‚¹ã‚¯ãƒªãƒ¼ãƒ³ã‚·ãƒ§ãƒƒãƒˆã®ãƒ‡ãƒ¢ãƒ—ãƒã‚°ãƒ©ãƒ ) ã¯ ã“ã¡ã‚‰ ã«ãªã‚Šã¾ã™ã€‚ ç›¸äº’è¡çªã™ã‚‹ãƒ‘ãƒ¼ãƒ†ã‚£ã‚¯ãƒ«ã‚’å®Ÿè£…ã™ã‚‹å ´åˆã€ãŠäº’ã„ã®è·é›¢ã‚’è¨ˆç®—ã—ã€å½“ãŸã£ã¦ã„ãŸã‚‰ã‚ã‚Šè¾¼ã¿å…·åˆã«å¿œã˜ã¦æŠ¼ã—è¿”ã™ã€ã¨ã„ã†ã®ãŒã‚ˆãã‚ã‚‹å®Ÿè£…ã ã¨æ€ã„ã¾ã™ã€‚ã¾ãšã¯ãã‚Œã‚’ã‚¹ãƒˆãƒ¬ãƒ¼ãƒˆã«
yuroyoro 2015/06/24
c++

x86

SIMD
ãƒªãƒ³ã‚¯
1