{"id":128,"date":"2017-01-06T02:20:24","date_gmt":"2017-01-06T02:20:24","guid":{"rendered":"https:\/\/css.or.kr\/csc2020\/?page_id=128"},"modified":"2020-01-28T15:40:26","modified_gmt":"2020-01-28T06:40:26","slug":"%ed%8f%ac%ec%8a%a4%ed%84%b0-%ec%84%b8%ec%85%98","status":"publish","type":"page","link":"https:\/\/css.or.kr\/csc2020\/?page_id=128","title":{"rendered":"\ud3ec\uc2a4\ud130 \uc138\uc158"},"content":{"rendered":"\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n
\uc18c\uc18d<\/strong><\/td>\n\ubc1c\ud45c\uc790<\/strong><\/td>\n\uc9c0\ub3c4\uad50\uc218<\/strong><\/td>\n\uc81c\ubaa9<\/strong><\/td>\n<\/tr>\n
KAIST<\/td>\n\ubc15\uc911\uc5b8<\/td>\n\ud5c8\uc7ac\ud601<\/td>\nShieldStore: Shielded In-memory Key-value Storage with SGX<\/td>\n<\/tr>\n
KAIST<\/td>\n\ub098\uc120\uc9c4<\/td>\n\ud5c8\uc7ac\ud601<\/td>\nHeterogeneous Isolated Execution for Commodity GPUs<\/td>\n<\/tr>\n
KAIST<\/td>\n\uad8c\uc601\uc740<\/td>\n\uc720\ubbfc\uc218<\/td>\nTensorDIMM: A Practical Near-Memory Processing Architecture for Embeddings and Tensor Operations in Deep Learning<\/td>\n<\/tr>\n
KAIST<\/td>\n\ucd5c\uc720\uc815<\/td>\n\uc720\ubbfc\uc218<\/td>\nPREMA: A Predictive Multi-task Scheduling Algorithm For Preemptible NPUs<\/td>\n<\/tr>\n
KAIST<\/td>\n\ud604\ubd09\uc900<\/td>\n\uc720\ubbfc\uc218<\/td>\nNeuMMU: Architectural Support for Efficient Address Translations in Neural Processing Units<\/td>\n<\/tr>\n
POSTECH<\/td>\n\ud5c8\uc120\uc601<\/td>\n\uae40\ud55c\uc900<\/td>\nReal-time Object Detection System with Multi-path Neural Networks<\/td>\n<\/tr>\n
POSTECH, \uc5f0\uc138\ub300\ud559\uad50<\/td>\n\uc870\uc131\uc900<\/td>\n\uae40\ud55c\uc900<\/td>\nCR2 : A Community-aware Compressed Regular Representation for High-Performance Graph Processing on GPUs<\/td>\n<\/tr>\n
\ud55c\uc591\ub300\ud559\uad50<\/td>\n\uac15\uc11d\uc6d0<\/td>\n\ubc15\uc601\uc900<\/td>\nGATE: A Generalized Dataflow-level Approximation Tuning Engine For Data Parallel Architectures<\/td>\n<\/tr>\n
\uc11c\uc6b8\ub300\ud559\uad50<\/td>\n\uc2e0\uc7ac\ubbfc<\/td>\n\uae40\uc9c0\ud64d<\/td>\n\uc5f0\uc18d\uc131\uc744 \uace0\ub824\ud55c \ud574\uc2dc \uae30\ubc18\uc758 \ud50c\ub798\uc2dc \ubcc0\ud658 \uacc4\uce35<\/td>\n<\/tr>\n
\uc11c\uc6b8\ub300\ud559\uad50<\/td>\n\uae40\uc724\uc544<\/td>\n\uae40\uc9c0\ud64d<\/td>\n\uc2a4\ub9c8\ud2b8SSD\ub97c \uc704\ud55c \uc131\ub2a5\ud3c9\uac00 \ubaa8\ub378 \ubc0f \ud65c\uc6a9<\/td>\n<\/tr>\n
\uc11c\uc6b8\ub300\ud559\uad50<\/td>\n\ubc15\uc900\ubaa8<\/td>\n\uc804\ub3d9\uc11d<\/td>\nMicroarchitecture-Aware Code Generation for Deep Learning on Single-ISA Heterogeneous Multi-Core Mobile Processors<\/td>\n<\/tr>\n
\uc11c\uc6b8\ub300\ud559\uad50<\/td>\n\uc774\uc6d0\uc2dd<\/td>\n\uae40\uc7a5\uc6b0<\/td>\nFIDR: A Scalable Storage System for Fine-Grain Inline Data Reduction with Efficient Memory Handling<\/td>\n<\/tr>\n
\uc11c\uc6b8\ub300\ud559\uad50<\/td>\n\uae40\ucc2c\uba85<\/td>\n\uae40\uc7a5\uc6b0<\/td>\nFlexLearn: Fast and Highly Efficient Brain Simulations Using Flexible On-Chip Learning<\/td>\n<\/tr>\n
\uc11c\uc6b8\ub300\ud559\uad50<\/td>\n\ud5c8\uc218\uc5f0<\/td>\n\uae40\uc7a5\uc6b0<\/td>\nMnnFast: A Fast and Scalable System Architecture for Memory-Augmented Neural Networks<\/td>\n<\/tr>\n
\uc22d\uc2e4\ub300\ud559\uad50<\/td>\n\ud64d\ucc44\uc6d0,\ud55c\ud0dc\ud638,\uc591\uac74\ubaa8,\uae40\uc7ac\uc740,\ucd5c\uc724\uc544<\/td>\n\uae40\uac15\ud76c<\/td>\nNVIDIA AGX Xavier \ubcf4\ub4dc \uc0c1\uc5d0\uc11c\uc758 \uc790\uc728\uc8fc\ud589 \uc18c\ud504\ud2b8\uc6e8\uc5b4 Autoware \uc774\uc2dd<\/td>\n<\/tr>\n
\uc744\uc9c0\ub300\ud559\uad50<\/td>\n\ubc31\uc2b9\uc6b0<\/td>\n\uae40\uaddc\ud638<\/td>\nCNN\uc744 \uc774\uc6a9\ud55c \uac1c \uac10\uc815\uc778\uc2dd \ubc29\ubc95\uc5d0 \uad00\ud55c \uc5f0\uad6c<\/td>\n<\/tr>\n
\uc911\uc559\ub300\ud559\uad50<\/td>\n\uc774\uc900\ud611<\/td>\n\uae40\ud6a8\uc218<\/td>\n\ud130\uce58 \uc18c\ub9ac \uae30\ubc18 \ubaa8\ubc14\uc77c \ub514\ubc14\uc774\uc2a4 \uc0ac\uc6a9\uc790 \uc778\uc2dd \uc5f0\uad6c<\/td>\n<\/tr>\n
\ub2e8\uad6d\ub300\ud559\uad50<\/td>\n\uc131\ud638\uc900<\/td>\n\uc870\uc131\uc81c<\/td>\n\uc784\ubca0\ub514\ub4dc \uc6a9 \uad50\ucc28 \ucef4\ud30c\uc77c\ub7ec \ud658\uacbd\uc5d0\uc11c\uc758 DLL Injection \uacf5\uaca9\uacfc \ubc29\uc5b4 \uae30\ubc95<\/td>\n<\/tr>\n
\uc11c\uc6b8\ub300\ud559\uad50<\/td>\n\ud5c8\uc900<\/td>\n\uc774\uc7ac\uc6b1<\/td>\nCharon: Specialized Near-Memory Processing Architecture for Clearing Dead Objects in Memory<\/td>\n<\/tr>\n
\uc11c\uc6b8\ub300\ud559\uad50<\/td>\n\ubc30\uc885\ud604<\/td>\n\uc774\uc7ac\uc6b1<\/td>\nSSDStreamer: Specializing I\/O Stack for Large-Scale Machine Learning<\/td>\n<\/tr>\n
\uc11c\uc6b8\ub300\ud559\uad50<\/td>\n\uc815\uc131\uc900<\/td>\n\uc774\uc7ac\uc6b1<\/td>\nA^3: Accelerating Neural Network Attention Mechanismwith Approximation<\/td>\n<\/tr>\n
\ud55c\uad6d\ud56d\uacf5\ub300\ud559\uad50<\/td>\n\uc774\uccad\uc900<\/td>\n\uc774\uc7ac\ud658<\/td>\nAnalysis on the I\/O characteristics of intel KNL processor<\/td>\n<\/tr>\n
\uc911\uc559\ub300\ud559\uad50<\/td>\n\uc774\uc7ac\ud658<\/td>\n\uae40\uc131\uc870, \ubc15\uc0c1\uc624<\/td>\n\ubaa8\ubc14\uc77c \uc2dc\uc2a4\ud15c\uc758 \uc804\ub825 \uc18c\ubaa8 \ucd5c\uc801\ud654\ub97c \uc704\ud55c LSTM \uae30\ubc18\uc758 \ud504\ub85c\uc138\uc11c \uc81c\uc5b4 \uc5f0\uad6c<\/td>\n<\/tr>\n
\uc11c\uc6b8\ub300\ud559\uad50<\/td>\n\ubc15\ud558\uc5f0, \ubc15\ub300\ucca0<\/td>\n\uc774\ucc3d\uac74<\/td>\nOSCAR (Open-Source Self-Driving Car) Remote Lab: An Open Remote Self-Driving Laboratory for Students and Developers<\/td>\n<\/tr>\n
\uc544\uc8fc\ub300\ud559\uad50<\/td>\n\ucd5c\uc6d0\uad50, \uae40\uc885\ud604<\/td>\n\uc548\uc815\uc12d<\/td>\nAnalysis of Linux Memory Management to Tiered Memory Systems<\/td>\n<\/tr>\n<\/tbody>\n<\/table>\n","protected":false},"excerpt":{"rendered":"

\uc18c\uc18d \ubc1c\ud45c\uc790 \uc9c0\ub3c4\uad50\uc218 \uc81c\ubaa9 KAIST \ubc15\uc911\uc5b8 \ud5c8\uc7ac\ud601 ShieldStore: Shielded In-memory Key-value Storage with SGX KAIST \ub098\uc120\uc9c4 \ud5c8\uc7ac\ud601 Heterogeneous Isolated Execution for Commodity GPUs KAIST \uad8c\uc601\uc740 \uc720\ubbfc\uc218 TensorDIMM: A Practical Near-Memory Processing Architecture for Embeddings and Tensor Operations in Deep Learning KAIST \ucd5c\uc720\uc815 \uc720\ubbfc\uc218 PREMA: A Predictive Multi-task Scheduling Algorithm For Preemptible NPUs KAIST \ud604\ubd09\uc900 \uc720\ubbfc\uc218 Read More …<\/a><\/p>\n","protected":false},"author":1,"featured_media":0,"parent":0,"menu_order":0,"comment_status":"closed","ping_status":"closed","template":"","meta":[],"_links":{"self":[{"href":"https:\/\/css.or.kr\/csc2020\/index.php?rest_route=\/wp\/v2\/pages\/128"}],"collection":[{"href":"https:\/\/css.or.kr\/csc2020\/index.php?rest_route=\/wp\/v2\/pages"}],"about":[{"href":"https:\/\/css.or.kr\/csc2020\/index.php?rest_route=\/wp\/v2\/types\/page"}],"author":[{"embeddable":true,"href":"https:\/\/css.or.kr\/csc2020\/index.php?rest_route=\/wp\/v2\/users\/1"}],"replies":[{"embeddable":true,"href":"https:\/\/css.or.kr\/csc2020\/index.php?rest_route=%2Fwp%2Fv2%2Fcomments&post=128"}],"version-history":[{"count":15,"href":"https:\/\/css.or.kr\/csc2020\/index.php?rest_route=\/wp\/v2\/pages\/128\/revisions"}],"predecessor-version":[{"id":392,"href":"https:\/\/css.or.kr\/csc2020\/index.php?rest_route=\/wp\/v2\/pages\/128\/revisions\/392"}],"wp:attachment":[{"href":"https:\/\/css.or.kr\/csc2020\/index.php?rest_route=%2Fwp%2Fv2%2Fmedia&parent=128"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}