• Àüü
  • ÀüÀÚ/Àü±â
  • Åë½Å
  • ÄÄÇ»ÅÍ
´Ý±â

»çÀÌÆ®¸Ê

Loading..

Please wait....

±¹³» ³í¹®Áö

Ȩ Ȩ > ¿¬±¸¹®Çå > ±¹³» ³í¹®Áö > Çѱ¹Á¤º¸°úÇÐȸ ³í¹®Áö > Á¤º¸°úÇÐȸ³í¹®Áö (Journal of KIISE)

Á¤º¸°úÇÐȸ³í¹®Áö (Journal of KIISE)

Current Result Document :

ÇѱÛÁ¦¸ñ(Korean Title) ¿¡³ÊÁö È¿À²ÀûÀÎ ³ëµå ³» Á¡´ëÁ¡ Åë½ÅÀ» À§ÇÑ MPI ÇÁ·Î±×·¹½º ¿£Áø
¿µ¹®Á¦¸ñ(English Title) MPI Progress Engine for Energy Efficient Intra-Node Point-to-Point Communication
ÀúÀÚ(Author) ±è°Ç¿ì   ÁøÇö¿í   Keon-Woo Kim   Hyun-Wook Jin  
¿ø¹®¼ö·Ïó(Citation) VOL 48 NO. 10 PP. 1069 ~ 1076 (2021. 10)
Çѱ۳»¿ë
(Korean Abstract)
MPI(Message Passing Interface) ¶óÀ̺귯¸®ÀÇ Åë½Å ¿Ï·á °èÃþÀÎ ÇÁ·Î±×·¹½º ¿£ÁøÀº Æú¸µÀ» ÅëÇØ ¸Þ½ÃÁö µµÂø µîÀÇ Åë½Å »óÅ º¯È­¸¦ ÀνÄÇÑ´Ù. ÀÌ·¯ÇÑ Æú¸µÀº ³·Àº Åë½Å Áö¿¬½Ã°£À» Á¦°øÇÏÁö¸¸, Æú¸µÇÏ´Â µ¿¾È¿¡µµ ÇÁ·Î±×·¹½º ¿£ÁøÀº CPU ÀÚ¿øÀ» Á¡À¯Çϱ⠶§¹®¿¡ ³·Àº ¿¡³ÊÁö È¿À²¼ºÀ» ÃÊ·¡ÇÑ´Ù. ¿¢»ç½ºÄÉÀÏ ½Ã½ºÅÛÀÌ µîÀåÇϸ鼭 Áö¿¬¿Ö°î(skew)ÀÌ Ä¿Áü¿¡ µû¶ó MPI ÇÁ·Î±×·¹½º ¿£ÁøÀÇ Æú¸µ¿¡ ÀÇÇÑ ¿¡³ÊÁö È¿À²¼º ÀúÇÏ´Â ´õ¿í ½É°¢ÇØÁö°í ÀÖ´Ù. º» ³í¹®¿¡¼­´Â Æú¸µ°ú ½Ã±×³Î ¹æ½ÄÀ» È¥¿ëÇÏ´Â ÇÁ·Î±×·¹½º ¿£ÁøÀ» Á¦¾ÈÇÏ¿© ÀÛÀº ¸Þ½ÃÁö 󸮸¦ À§ÇÑ Eager ÇÁ·ÎÅäÄÝ°ú Å« ¸Þ½ÃÁö 󸮸¦ À§ÇÑ Rendezvous ÇÁ·ÎÅäÄÝÀÌ ¿¡³ÊÁö È¿À²ÀûÀÎ ³ëµå ³» Åë½ÅÀ» ¼öÇàÇÒ ¼ö ÀÖµµ·Ï ÇÑ´Ù. OSU ¸¶ÀÌÅ©·Îº¥Ä¡¸¶Å©¿Í NAS ¼º´É ÃøÁ¤ °á°ú´Â Áö¿¬¿Ö°îÀÌ Ä¿Áú¼ö·Ï Á¦¾ÈµÈ ½Ã±×³Î ±â¹Ý ÇÁ·Î±×·¹½º ¿£ÁøÀÌ ¿¡³ÊÁö È¿À²À» Å©°Ô Çâ»ó½ÃÅ°¸ç, CPU ÀÚ¿øÀÌ ¿©·¯ ÇÁ·Î¼¼½º¿¡ ÀÇÇؼ­ °øÀ¯µÇ´Â ȯ°æ¿¡¼­´Â ÀÀ¿ëÀÇ ½ÇÇà½Ã°£µµ Çâ»ó½Ãų ¼ö ÀÖÀ½À» º¸ÀδÙ.
¿µ¹®³»¿ë
(English Abstract)
The communication completion layer called progress engine in the Message Passing Interface (MPI) library recognizes changes in communication states, such as message arrivals, by polling. Although polling provides a low communication latency, its use results in low energy efficiency because the progress engine occupies CPU resources while performing polling. The decrease in energy efficiency induced by polling has become more severe as the skew has increased with the advent of exascale systems. In this paper, we suggest a progress engine that uses both polling and signal, with the Eager protocol for small message processing and the Rendezvous protocol for large message processing to perform energy-efficient intra-node communication. The OSU microbenchmark and NAS performance measurements show that the suggested signal-based progress engine improves energy efficiency as the skew increases and reduces the execution time of applications when CPU resources are shared between multiple processes.
Å°¿öµå(Keyword) MPI   ³ëµå ³» Åë½Å   ¿¡³ÊÁö È¿À²   ÇÁ·Î±×·¹½º ¿£Áø   MPI   intra-node communication   energy efficiency   progress engine  
ÆÄÀÏ÷ºÎ PDF ´Ù¿î·Îµå