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

»çÀÌÆ®¸Ê

Loading..

Please wait....

±¹³» ³í¹®Áö

Ȩ Ȩ > ¿¬±¸¹®Çå > ±¹³» ³í¹®Áö > ÀüÀÚ°øÇÐȸ ³í¹®Áö (Journal of The Institute of Electronics and Information Engineers)

ÀüÀÚ°øÇÐȸ ³í¹®Áö (Journal of The Institute of Electronics and Information Engineers)

Current Result Document : 5 / 237 ÀÌÀü°Ç ÀÌÀü°Ç   ´ÙÀ½°Ç ´ÙÀ½°Ç

ÇѱÛÁ¦¸ñ(Korean Title) ÀÚÀ²ÁÖÇà ÀÚµ¿Â÷¿¡¼­ DNN ±â¹Ý ¾ÈÀü Çʼö ÀÀ¿ëÀÇ ¼º´É ÀúÇÏ ¹æÁö
¿µ¹®Á¦¸ñ(English Title) Avoiding Performance Degradation of DNN based Safety-Critical Applications in Autonomous Vehicles
ÀúÀÚ(Author) ±èµµ¿µ   ±è¸í¼±   Doyoung Kim   Myungsun Kim  
¿ø¹®¼ö·Ïó(Citation) VOL 57 NO. 09 PP. 0056 ~ 0064 (2020. 09)
Çѱ۳»¿ë
(Korean Abstract)
DNN(deep neural network) ±â¼úÀÇ ºü¸¥ ¹ßÀüÀº ÀÚÀ²ÁÖÇà ÀÚµ¿Â÷¿¡¼­ ¶Ù¾î³­ ¼º´É Çâ»óÀ¸·Î À̾îÁö°í ÀÖ´Ù. ÀÌ·¯ÇÑ ½Ã½ºÅÛ¿¡¼­´Â °´Ã¼ ŽÁö, Â÷¼± ÀÌÅ» ¹æÁö, Á¹À½¿îÀü ¹æÁö, ÁÖº¯ À̹ÌÁö ºÐ¼® µî ¿©·¯ °¡Áö DNN ¸ðµ¨ ±â¹ÝÀÇ ÀÀ¿ëµéÀÌ ¼öÇàµÈ´Ù. À̶§ ¼öÇàµÇ´Â DNN ¸ðµ¨µéÀ» È£½ºÆÃÇÏ´Â ÀÀ¿ëµéÀº ³ôÀº ¿ì¼±¼øÀ§¸¦ °®´Â ¿îÇà¿¡ °ü·ÃµÈ ¾ÈÀü ÇʼöÀûÀÎ °Íµé°ú Áß°£ ȤÀº ³·Àº ¿ì¼±¼øÀ§¸¦ °®°í ¿îÇà¿¡´Â Á÷Á¢ÀûÀ¸·Î °ü·ÃµÇÁö ¾Ê´Â ÀÀ¿ëµé·Î ÀÌ·ç¾îÁø´Ù. µû¶ó¼­ GPU¿Í °°Àº DNN °¡¼Ó ÀåÄ¡¸¦ ¼­·Î ´Ù¸¥ ¿ì¼±¼øÀ§¸¦ °®´Â DNN ¸ðµ¨µéÀÌ °øÀ¯Çؼ­ »ç¿ëÇÑ´Ù. ¶ÇÇÑ ÇϳªÀÇ DNN ¸ðµ¨ ´ç ÇϳªÀÇ ¹èÄ¡ ÇÁ·Î¼¼½º ÇüÅ·Π¼öÇàµÇ¾î, ºñÁÖ±âÀûÀ¸·Î µµÂøÇÏ´Â ¾ÈÀü Çʼö ÀÀ¿ëµéÀÌ »ç¿ëÇÏ´Â DNN ¸ðµ¨µéÀ» ÃÖ¿ì¼±ÀûÀ¸·Î GPU¿¡ ÇÒ´çÇϱâ´Â ¾î·Æ´Ù. º» ³í¹®¿¡¼­´Â ÀÌ·¯ÇÑ ¹®Á¦¸¦ ÇØ°áÇϱâ À§ÇÏ¿© ¿ì¼±¼øÀ§¸¦ ±â¹ÝÀ¸·Î DNN ¸ðµ¨ÀÇ °¢ ·¹ÀÌ¾î º°·Î GPU¿¡ ¿¬»êÀ» ÀÇ·ÚÇÏ´Â ÇÁ·¹ÀÓ¿öÅ©¸¦ Á¦¾ÈÇÑ´Ù. Á¦¾ÈµÈ ±â¹ýÀ» ½ÇÁ¦ »ó¿ë º¸µå¿¡ žÀçÇÑ ÈÄ ½ÇÇèÇÑ °á°ú ³ôÀº ¿ì¼±¼øÀ§ÀÇ DNN ¸ðµ¨µéÀÇ ¼º´ÉÀÌ Àû¿ë Àü ´ëºñ ÃÖ´ë 43.9% Çâ»óµÇ¾ú´Ù.
¿µ¹®³»¿ë
(English Abstract)
The rapid development of DNN technology is leading to outstanding performance improvement in autonomous vehicles. In this system, various DNN model-based applications such as object detection, lane departure prevention, drowsy driving prevention, and surrounding image analysis are performed. The applications hosting the DNN models are classified into safety-critical ones with high priority and others with medium or low priority not directly related to driving. Therefore, DNN accelerators such as GPUs are shared and used by DNN models with different priorities. In addition, since it is performed in the form of one batch process per DNN model, it is difficult to firstly allocate the DNN models used by safety-critical applications that arrive sporadically to the GPU. In order to solve this problem, this paper proposes a framework for requesting computation to the GPU for each layer of the DNN model based on priority. As a result of experiments after running the proposed technique on an actual off-the-shelf board, the performance of high-priority DNN models improved upto 43.9% compared to that of before.
Å°¿öµå(Keyword) Deep neural network   Safety-critical   Multi-DNN   Priority scheduling   Latency  
ÆÄÀÏ÷ºÎ PDF ´Ù¿î·Îµå