[B! fpga][CPU] y_uukiã®ãƒ–ãƒƒã‚¯ãƒžãƒ¼ã‚¯

y_uuki id:y_uuki

fpgaã¨CPUã«é–¢ã™ã‚‹y_uukiã®ãƒ–ãƒƒã‚¯ãƒžãƒ¼ã‚¯ (3)

${{author_name}}$

{{{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}}

Linux OSã‹ã‚‰FPGAã‚’é€éŽçš„ã«åˆ©ç”¨ã™ã‚‹æ§‹æƒ³ã€‚æ–‡å—åˆ—å‡¦ç†ã‚’CPUã‹ã‚‰FPGAã¸ã‚ªãƒ•ãƒãƒ¼ãƒ‰ã§10å€é€Ÿã«ãªã£ãŸç ”ç©¶çµæžœã‚’ãƒŸãƒ©ã‚¯ãƒ«ãƒ»ãƒªãƒŠãƒƒã‚¯ã‚¹ãŒç™ºè¡¨
Linux OSã‹ã‚‰FPGAã‚’é€éŽçš„ã«åˆ©ç”¨ã™ã‚‹æ§‹æƒ³ã€‚æ–‡å—åˆ—å‡¦ç†ã‚’CPUã‹ã‚‰FPGAã¸ã‚ªãƒ•ãƒãƒ¼ãƒ‰ã§10å€é€Ÿã«ãªã£ãŸç ”ç©¶çµæžœã‚’ãƒŸãƒ©ã‚¯ãƒ«ãƒ»ãƒªãƒŠãƒƒã‚¯ã‚¹ãŒç™ºè¡¨ ãƒ—ãƒã‚»ãƒƒã‚µå†…éƒ¨ã®ãƒã‚¸ãƒƒã‚¯ã‚’ã‚½ãƒ•ãƒˆã‚¦ã‚§ã‚¢ã§å‹•çš„ã«æ›¸ãæ›ãˆã‚‹ã“ã¨ãŒã§ãã‚‹FPGAã¯ã€ã‚¢ãƒ—ãƒªã‚±ãƒ¼ã‚·ãƒ§ãƒ³ã”ã¨ã«ãƒã‚¸ãƒƒã‚¯ã‚’æœ€é©åŒ–ã§ãã‚‹ãŸã‚ã€x86ãªã©ã®æ±Žç”¨ãƒ—ãƒã‚»ãƒƒã‚µã‚ˆã‚Šã‚‚é«˜é€Ÿã‹ã¤åŠ¹çŽ‡çš„ãªã‚¢ãƒ—ãƒªã‚±ãƒ¼ã‚·ãƒ§ãƒ³ã®å®Ÿè¡ŒãŒå¯èƒ½ã«ãªã‚‹ã¨æ³¨ç›®ã•ã‚Œã¦ã„ã¾ã™ã€‚ ç‰¹ã«ã€ä»Šå¹´ã®1æœˆã«ã‚¤ãƒ³ãƒ†ãƒ«ãŒFPGAå¤§æ‰‹ã®ã‚¢ãƒ«ãƒ†ãƒ©ã‚’è²·åŽã—ã€Xeonã«FPGAã‚’çµ±åˆã™ã‚‹ã¨ã„ã†å°†æ¥è¨ˆç”»ã‚’ç™ºè¡¨ã—ã¦ã‹ã‚‰ã€FPGAã¸ã®æ³¨ç›®åº¦ã¯ã•ã‚‰ã«é«˜ã¾ã£ã¦ã„ã¾ã™ã€‚ ã—ã‹ã—ã„ã¾ã®ã¨ã“ã‚ã€FPGAã‚’ã‚¢ãƒ—ãƒªã‚±ãƒ¼ã‚·ãƒ§ãƒ³ã‹ã‚‰åˆ©ç”¨ã™ã‚‹ã«ã¯ã€ãƒã‚¸ãƒƒã‚¯ã‚’é–‹ç™ºãƒ„ãƒ¼ãƒ«ãªã©ã‚’ç”¨ã„ã¦FPGAã«æ›¸ãè¾¼ã¿ã€ãã‚Œã‚’ãƒ‡ãƒã‚¤ã‚¹ãƒ‰ãƒ©ã‚¤ãƒã‚’çµŒç”±ã—ã¦ãƒ•ãƒ¬ãƒ¼ãƒ ãƒ¯ãƒ¼ã‚¯ãªã©ã‚’é€šã˜ã¦åˆ©ç”¨ã™ã‚‹ã¨ã„ã†ã‚‚ã®ã«ãªã‚Šã¾ã™ã€‚ ã“ã‚Œã‚’ã‚‚ã£ã¨ç°¡å˜ã«ã—ã‚ˆã†ã¨ã„ã†æ§‹æƒ³ã‚’æ˜Žã‚‰ã‹ã«ã—ãŸ
y_uuki 2016/12/05
fpga

CPU
ãƒªãƒ³ã‚¯
ä»Šã©ãCPUã ã‘ã§å¤§ä¸ˆå¤«ï¼Ÿãƒ“ãƒƒã‚°ãƒ‡ãƒ¼ã‚¿ã‚„äººå·¥çŸ¥èƒ½ã§GPU/FPGAã‚’ä½¿ã†å‰ã«çŸ¥ã£ã¦ãŠããŸã„â€œãƒãƒ¼ãƒ‰å±‹â€ã¨â€œã‚½ãƒ•ãƒˆå±‹â€ã®é•ã„
é–¢é€£è¨˜äº‹ ã€ŒBig Surã€ã¯ãƒ‹ãƒ¥ãƒ¼ãƒ©ãƒ«ãƒãƒƒãƒˆãƒ¯ãƒ¼ã‚¯ã‚·ã‚¹ãƒ†ãƒ ã®ãŸã‚ã«ãƒ•ã‚§ã‚¤ã‚¹ãƒ–ãƒƒã‚¯ãŒè¨è¨ˆã—ãŸãƒãƒ¼ãƒ‰ã‚¦ã‚¨ã‚¢ ç±³ãƒ•ã‚§ã‚¤ã‚¹ãƒ–ãƒƒã‚¯ã¯ã€Œãƒ‹ãƒ¥ãƒ¼ãƒ©ãƒ«ãƒãƒƒãƒˆãƒ¯ãƒ¼ã‚¯ã‚·ã‚¹ãƒ†ãƒ å‘ã‘ã€ã¨ã„ã†ãƒãƒ¼ãƒ‰ã‚¦ã‚¨ã‚¢ã‚¢ãƒ¼ã‚ãƒ†ã‚¯ãƒãƒ£ã€ŒBig Surã€ã‚’ã‚ªãƒ¼ãƒ—ãƒ³ã‚½ãƒ¼ã‚¹ã§å…¬é–‹ã—ãŸã€‚ AWSï¼Azureä¸Šã§TensorFlowã€CNTKã‚’â€•â€•æ©Ÿæ¢°å¦ç¿’ç’°å¢ƒã®ãƒžãƒãƒ¼ã‚¸ãƒ‰ã‚µãƒ¼ãƒ“ã‚¹ãŒç™»å ´ ã‚¯ãƒ©ã‚¹ã‚ãƒ£ãƒƒãƒˆãŒã€æ©Ÿæ¢°å¦ç¿’ç’°å¢ƒã®ãƒžãƒãƒ¼ã‚¸ãƒ‰ã‚µãƒ¼ãƒ“ã‚¹ã€ŒClassCat Deep Learning Service v2ã€ã®æä¾›ã‚’ç™ºè¡¨ã—ãŸã€‚Amazon Web Servicesã¨Microsoft Azureã«å¯¾å¿œã™ã‚‹ã€‚ ã‚¶ã‚¤ãƒªãƒ³ã‚¯ã‚¹ã€OpenCLã€Cï¼C++ã§FPGAã‚¢ã‚¯ã‚»ãƒ©ãƒ¬ãƒ¼ã‚·ãƒ§ãƒ³ã‚’è¡Œã†é–‹ç™ºç’°å¢ƒæœ€æ–°ç‰ˆã®ãƒªãƒªãƒ¼ã‚¹ã‚’ç™ºè¡¨ ã‚¶ã‚¤ãƒªãƒ³ã‚¯ã‚¹ã¯5æœˆ26æ—¥ï¼ˆç±³å›½æ™‚é–“ï¼‰ã€FPGAã‚’ä½¿ç”¨ã—ãŸã‚¢ã‚¯ã‚»ãƒ©ãƒ¬ãƒ¼ã‚¿ãƒ¼é–‹ç™ºç’°å¢ƒã®æœ€æ–°ç‰ˆã€ŒSDAccelé–‹ç™ºç’°å¢ƒ2015.1ã€ã®ãƒªãƒªãƒ¼
y_uuki 2016/05/10
CPU

GPU

fpga
ãƒªãƒ³ã‚¯
ãƒãƒ¼ãƒ‰ç´ äººãŒ32bit CPUã‚’FPGAã§è‡ªä½œã—ã¦å‹•ã‹ã™ã¾ã§èªã‚“ã æœ¬ã®ã¾ã¨ã‚ - ã‚¹ãƒ†ã‚£ãƒ«ãƒã‚¦ã‚¹ã®æ›¸åº«ã®æ›¸åº«
ç”·åãŸã‚‹ã‚‚ã®ä¸€åº¦ã¯è‡ªåˆ†ã§CPUã‚’ä½œã£ã¦ã¿ãŸã„ã‚‚ã®ã ã‘ã©ã€ICã§LEDã‚’ãƒ”ã‚«ãƒ”ã‚«ã•ã›ãŸç¨‹åº¦ã®çµŒé¨“ã—ã‹ãªã„ãƒãƒ¼ãƒ‰ç´ äººãªä¿ºã ã£ãŸã®ã§ã€CPUã‚’è‡ªä½œã—ã¦ã‚‹æ±å¤§ç”Ÿãªã©ã‚’é ãã‹ã‚‰è¦‹ã¦æ†§ã‚Œã¦ã‚‹ã ã‘ã ã£ãŸã€‚ã—ã‹ã—ãŠã‚ˆãä¸€å¹´å‰ã®ã“ã¨ã€ã€ŒMIPSãªã‚“ã¦ç°¡å˜ã«ä½œã‚Œã¾ã™ã‚ˆï¼ã€ã¨Kã•ã‚“ï¼ˆâ†FPGAã§Lispãƒžã‚·ãƒ³ã‚’è‡ªä½œã™ã‚‹ã‚ˆã†ãªäººï¼‰ã«è¨€ã‚ã‚Œã¦ã€ãŠã€ãŠã…ã€‚ã€‚ãã‚Šã‚ƒKã•ã‚“ã¯ãã†ã§ã—ã‚‡ã†ã‚ˆã€‚ã€‚ã‚ã‚Œã€ã‚‚ã—ã‹ã—ã¦ä¿ºã«ã‚‚ã§ãã‚‹ã‹ãªã€‚ã€‚ï¼Ÿã€€ã¨æ€ã£ãŸã€‚ã“ã®ä¸€è¨€ãŒãªã‘ã‚Œã°32bitã®CPUã‚’è‡ªä½œã—ã‚ˆã†ãªã‚“ã¦è€ƒãˆãªã‹ã£ãŸã ã‚ã†ã€‚ ãã‚Œã‹ã‚‰ä¸€å¹´ã¡ã‚‡ã„ã€ã¨ãã«ä»Šå¹´ã®æ£æœˆä¼‘ã¿ã‚„FPGA æ¸©æ³‰ã§ãŒã£ã¤ã‚ŠãŒã‚“ã°ã£ã¦ã€ãªã‚“ã¨ã‹MIPS Iã‚µãƒ–ã‚»ãƒƒãƒˆã®è‡ªä½œCPUãŒå‹•ã„ãŸã€‚ã“ã‚Œã¯ãƒ•ã‚£ãƒœãƒŠãƒƒãƒã‚’è¨ˆç®—ã—ã¦ã‚‹ã¨ã“ã‚ã€‚ ã¡ãªã¿ã«ã€ã“ã‚Œã¯ã“ã‚“ãªæ„Ÿã˜ã®ãƒ•ã‚£ãƒœãƒŠãƒƒãƒã®ã‚³ãƒ¼ãƒ‰ã‚’Cã§æ›¸ã„ã¦ã€ void main() { int i, *r = (int *)0x7f00
y_uuki 2014/02/10
CPU

fpga
ãƒªãƒ³ã‚¯
1