{"id":2283,"date":"2019-09-10T10:13:02","date_gmt":"2019-09-10T10:13:02","guid":{"rendered":"https:\/\/www.thegioimaychu.vn\/blog\/?p=2283"},"modified":"2019-09-11T00:33:54","modified_gmt":"2019-09-11T00:33:54","slug":"deeplearning11-he-thong-deep-learning-tham-khao-voi-10x-nvidia-gtx-1080-ti","status":"publish","type":"post","link":"https:\/\/thegioimaychu.vn\/blog\/ai-hpc\/deeplearning11-he-thong-deep-learning-tham-khao-voi-10x-nvidia-gtx-1080-ti-p2283\/","title":{"rendered":"DeepLearning11: H\u1ec7 th\u1ed1ng Deep Learning tham kh\u1ea3o v\u1edbi 10x NVIDIA GTX 1080 Ti"},"content":{"rendered":"<style>.iblink{color:#975!important;}.iblink:hover{color:orange!important;}<\/style><p><span class=\"\">STH &#8211;  H\u00f4m nay ch\u00fang t\u00f4i \u0111ang tr\u00ecnh b\u00e0y m\u1ed9t b\u1ea3n d\u1ef1ng c\u00f3 l\u1ebd l\u00e0 c\u1ea5u h\u00ecnh h\u1ecdc s\u00e2u \u0111\u01b0\u1ee3c t\u00ecm ki\u1ebfm nhi\u1ec1u nh\u1ea5t hi\u1ec7n nay. <\/span>DeepLearning11 c\u00f3 10 GPU NVIDIA GeForce GTX 1080 Ti 11GB, Mellanox Infiniband v\u00e0 ph\u00f9 h\u1ee3p v\u1edbi m\u1ed9t form factor 4.5U nh\u1ecf g\u1ecdn. Ngo\u00e0i ra c\u00f2n c\u00f3 m\u1ed9t s\u1ef1 kh\u00e1c bi\u1ec7t quan tr\u1ecdng gi\u1eefa h\u1ec7 th\u1ed1ng n\u00e0y v\u00e0 DeepLearning10, b\u1ea3n d\u1ef1ng 8x GTX 1080 Ti c\u1ee7a ch\u00fang t\u00f4i. DeepLearning11 l\u00e0 m\u1ed9t thi\u1ebft k\u1ebf \u0111\u01a1n g\u1ed1c \u0111\u00e3 tr\u1edf n\u00ean ph\u1ed5 bi\u1ebfn trong c\u1ed9ng \u0111\u1ed3ng h\u1ecdc t\u1eadp s\u00e2u.<\/p>\n<p>T\u1ea1i STH, ch\u00fang t\u00f4i \u0111ang t\u1ea1o ra c\u00e1c b\u1ea3n d\u1ef1ng tham kh\u1ea3o h\u1ecdc t\u1eadp s\u00e2u h\u01a1n kh\u00f4ng ch\u1ec9 l\u00e0 c\u00e1c b\u00e0i t\u1eadp l\u00fd thuy\u1ebft. Nh\u1eefng m\u00e1y n\u00e0y \u0111ang \u0111\u01b0\u1ee3c s\u1eed d\u1ee5ng b\u1edfi nh\u00f3m c\u1ee7a ch\u00fang t\u00f4i ho\u1eb7c kh\u00e1ch h\u00e0ng c\u1ee7a ch\u00fang t\u00f4i. Ch\u00fang t\u00f4i \u0111\u00e3 th\u1ef1c hi\u1ec7n m\u1ed9t s\u1ed1 b\u1ea3n d\u1ef1ng nh\u1ecf h\u01a1n bao g\u1ed3m DeepLearning01 v\u00e0 DeepLearning02 m\u00e0 ch\u00fang t\u00f4i \u0111\u00e3 \u0111\u0103ng. Trong khi c\u00e1c b\u1ea3n d\u1ef1ng \u0111\u00f3 t\u1eadp trung v\u00e0o ph\u1ea7n gi\u1edbi thi\u1ec7u, DeepLearning11 l\u00e0 m\u1ed9t b\u1ea3n ho\u00e0n thi\u1ec7n ho\u00e0n to\u00e0n kh\u00e1c c\u1ee7a series. Ch\u00fang t\u00f4i bi\u1ebft c\u1ea5u h\u00ecnh n\u00e0y \u0111\u00e3 \u0111\u01b0\u1ee3c s\u1eed d\u1ee5ng b\u1edfi m\u1ed9t c\u00f4ng ty chuy\u00ean v\u1ec1 nghi\u00ean c\u1ee9u h\u1ecdc t\u1eadp s\u00e2u h\u00e0ng \u0111\u1ea7u tr\u00ean th\u1ebf gi\u1edbi.<\/p>\n<h2>DeepLearning11: C\u00e1c th\u00e0nh ph\u1ea7n<\/h2>\n<p>N\u1ebfu ch\u00fang t\u00f4i h\u1ecfi NVIDIA c\u00f3 l\u1ebd \u0111\u00e3 \u0111\u01b0\u1ee3c y\u00eau c\u1ea7u mua card Tesla ho\u1eb7c Quadro. NVIDIA \u0111\u1eb7c bi\u1ec7t y\u00eau c\u1ea7u c\u00e1c m\u00e1y ch\u1ee7 OEM kh\u00f4ng s\u1eed d\u1ee5ng card GTX c\u1ee7a h\u1ecd trong c\u00e1c m\u00e1y ch\u1ee7. T\u1ea5t nhi\u00ean, \u0111i\u1ec1u n\u00e0y \u0111\u01a1n gi\u1ea3n c\u00f3 ngh\u0129a l\u00e0 c\u00e1c \u0111\u1ea1i l\u00fd l\u1eafp \u0111\u1eb7t card tr\u01b0\u1edbc khi giao ch\u00fang cho kh\u00e1ch h\u00e0ng. L\u00e0 m\u1ed9t trang web \u0111\u00e1nh gi\u00e1 bi\u00ean t\u1eadp, ch\u00fang t\u00f4i c\u00f3 nh\u1eefng h\u1ea1n ch\u1ebf v\u1ec1 ng\u00e2n s\u00e1ch ch\u1eb7t ch\u1ebd n\u00ean ch\u00fang t\u00f4i \u0111\u00e3 mua 10 card NVIDIA GTX 1080 Ti. M\u1ed7i NVIDIA GTX 1080 Ti c\u00f3 b\u1ed9 nh\u1edb 11GB (t\u0103ng t\u1eeb 8GB tr\u00ean GTX 1080) v\u00e0 3584 nh\u00e2n CUDA (t\u0103ng t\u1eeb 2560 tr\u00ean GTX 1080.) S\u1ef1 kh\u00e1c bi\u1ec7t v\u1ec1 gi\u00e1 \u0111\u1ec3 ch\u00fang t\u00f4i n\u00e2ng c\u1ea5p t\u1eeb GTX 1080 l\u00e0 kho\u1ea3ng 1.500 USD. Ch\u00fang t\u00f4i \u0111\u00e3 mua card t\u1eeb nhi\u1ec1u nh\u00e0 cung c\u1ea5p.<\/p>\n<p><img decoding=\"async\" src=\"https:\/\/thegioimaychu.vn\/blog\/wp-content\/uploads\/2019\/09\/10x-NVIDIA-GTX-1080-TI-FE-Plus-Mellanox-Top.jpg\" sizes=\"(max-width: 790px) 100vw, 790px\" srcset=\"https:\/\/thegioimaychu.vn\/blog\/wp-content\/uploads\/2019\/09\/10x-NVIDIA-GTX-1080-TI-FE-Plus-Mellanox-Top.jpg 790w, https:\/\/www.servethehome.com\/wp-content\/uploads\/2017\/07\/10x-NVIDIA-GTX-1080-TI-FE-Plus-Mellanox-Top-400x266.jpg 400w, https:\/\/www.servethehome.com\/wp-content\/uploads\/2017\/07\/10x-NVIDIA-GTX-1080-TI-FE-Plus-Mellanox-Top-696x463.jpg 696w, https:\/\/www.servethehome.com\/wp-content\/uploads\/2017\/07\/10x-NVIDIA-GTX-1080-TI-FE-Plus-Mellanox-Top-632x420.jpg 632w\" alt=\"10 chi\u1ebfc NVIDIA GTX 1080 TI FE Plus Mellanox h\u00e0ng \u0111\u1ea7u\"><\/p>\r\n\t\t\t\t<div class=\"tgmcad hor\" style=\"text-align:center;margin:30px 0; padding: 10px 0; border-radius:5px; border-top:solid 1px #ccc; border-bottom:solid 1px #ccc; \">\r\n\t\t\t\t<a href=\"https:\/\/thegioimaychu.vn\/giai-phap\/deep-learning\/\"><img decoding=\"async\" alt=\"Gi\u1ea3i ph\u00e1p h\u1ea1 t\u1ea7ng Deep Learning, Tr\u00ed tu\u1ec7 Nh\u00e2n t\u1ea1o - AI\" src=\"https:\/\/thegioimaychu.vn\/blog\/wp-content\/uploads\/2019\/11\/tgmc-deep-learning.jpg\" \/><\/a>\r\n\t\t\t\t<\/div>\n<p>10 chi\u1ebfc NVIDIA GTX 1080 TI FE Plus Mellanox h\u00e0ng \u0111\u1ea7u<\/p>\n<p>H\u1ec7 th\u1ed1ng c\u1ee7a ch\u00fang t\u00f4i l\u00e0 <a class=\"iblink\" href=\"\/supermicro\/\">Supermicro<\/a> SYS-4028GR-TR2 (<a href=\"https:\/\/thegioimaychu.vn\/search\/?search=SYS-4028GR-TRT2&qt=sku&ref=tmsol\" class=\"sku-link\">SYS-4028GR-TRT2<\/a>), m\u1ed9t trong nh\u1eefng h\u1ec7 th\u1ed1ng m\u1eadt \u0111\u1ed9 GPU cao ch\u1ee7 \u0111\u1ea1o tr\u00ean th\u1ecb tr\u01b0\u1eddng. C\u00e1c <strong>-TR2<\/strong> r\u1ea5t c\u00f3 \u00fd ngh\u0129a v\u00ec n\u00f3 l\u00e0 phi\u00ean b\u1ea3n g\u1ed1c duy nh\u1ea5t c\u1ee7a chassis v\u00e0 kh\u00e1c bi\u1ec7t so v\u1edbi h\u1ec7 th\u1ed1ng r\u1ec5 k\u00e9p -TR DeepLearning10.<\/p>\n<p><img decoding=\"async\" src=\"https:\/\/thegioimaychu.vn\/blog\/wp-content\/uploads\/2019\/09\/DeepLearning11-GTX-1080-Ti-Same-CPU-Opposite-Ends.jpg\" sizes=\"(max-width: 800px) 100vw, 800px\" srcset=\"https:\/\/thegioimaychu.vn\/blog\/wp-content\/uploads\/2019\/09\/DeepLearning11-GTX-1080-Ti-Same-CPU-Opposite-Ends.jpg 800w, https:\/\/www.servethehome.com\/wp-content\/uploads\/2017\/07\/DeepLearning11-GTX-1080-Ti-Same-CPU-Opposite-Ends-400x300.jpg 400w, https:\/\/www.servethehome.com\/wp-content\/uploads\/2017\/07\/DeepLearning11-GTX-1080-Ti-Same-CPU-Opposite-Ends-80x60.jpg 80w, https:\/\/www.servethehome.com\/wp-content\/uploads\/2017\/07\/DeepLearning11-GTX-1080-Ti-Same-CPU-Opposite-Ends-265x198.jpg 265w, https:\/\/www.servethehome.com\/wp-content\/uploads\/2017\/07\/DeepLearning11-GTX-1080-Ti-Same-CPU-Opposite-Ends-696x522.jpg 696w, https:\/\/www.servethehome.com\/wp-content\/uploads\/2017\/07\/DeepLearning11-GTX-1080-Ti-Same-CPU-Opposite-Ends-560x420.jpg 560w\" alt=\"DeepLearning11 GTX 1080 Ti C\u00f9ng CPU K\u1ebft th\u00fac \u0111\u1ed1i di\u1ec7n\"><\/p>\n<p>DeepLearning11 GTX 1080 Ti C\u00f9ng CPU K\u1ebft th\u00fac \u0111\u1ed1i di\u1ec7n<\/p>\n<p>Gi\u1ed1ng nh\u01b0 b\u1ea3n d\u1ef1ng DeepLearning10, DeepLearning11 c\u00f3 ph\u1ea7n nh\u00f4 l\u00ean, n\u00e2ng t\u1ed5ng k\u00edch th\u01b0\u1edbc h\u1ec7 th\u1ed1ng l\u00ean t\u1edbi 4,5U. B\u1ea1n c\u00f3 th\u1ec3 \u0111\u1ecdc th\u00eam v\u1ec1 xu h\u01b0\u1edbng n\u00e0y trong Avert Your Eyes c\u1ee7a ch\u00fang t\u00f4i  t\u1eeb M\u00e1y ch\u1ee7 Xu h\u01b0\u1edbng Humping \u0111\u1ec9nh trong ph\u1ea7n Trung t\u00e2m d\u1eef li\u1ec7u  .<\/p>\n<p><img decoding=\"async\" src=\"https:\/\/thegioimaychu.vn\/blog\/wp-content\/uploads\/2019\/09\/Supermicro-4028GR-TR-Bump-for-GTX-Power-Cables.jpg\" sizes=\"(max-width: 800px) 100vw, 800px\" srcset=\"https:\/\/thegioimaychu.vn\/blog\/wp-content\/uploads\/2019\/09\/Supermicro-4028GR-TR-Bump-for-GTX-Power-Cables.jpg 800w, https:\/\/www.servethehome.com\/wp-content\/uploads\/2017\/05\/Supermicro-4028GR-TR-Bump-for-GTX-Power-Cables-400x226.jpg 400w, https:\/\/www.servethehome.com\/wp-content\/uploads\/2017\/05\/Supermicro-4028GR-TR-Bump-for-GTX-Power-Cables-696x392.jpg 696w, https:\/\/www.servethehome.com\/wp-content\/uploads\/2017\/05\/Supermicro-4028GR-TR-Bump-for-GTX-Power-Cables-745x420.jpg 745w\" alt=\"Supermicro 4028GR TR Bump cho c\u00e1p \u0111i\u1ec7n GTX\"><\/p>\n<p><a class=\"iblink\" href=\"\/supermicro\/\">Supermicro<\/a> 4028GR-TR \/ -TR2 Bump cho c\u00e1p \u0111i\u1ec7n GTX<\/p>\n<p>Ph\u1ea7n nh\u00f4 l\u00ean n\u00e0y cho ph\u00e9p ch\u00fang t\u00f4i s\u1eed d\u1ee5ng card NVIDIA GeForce GTX trong h\u1ec7 th\u1ed1ng c\u1ee7a ch\u00fang t\u00f4i v\u1edbi c\u00e1c c\u1ed5ng ngu\u1ed3n \u0111\u1ed1i x\u1ee9ng h\u00e0ng \u0111\u1ea7u c\u1ee7a ch\u00fang.<\/p>\n<p>Ch\u00fang t\u00f4i \u0111ang s\u1eed d\u1ee5ng b\u1ed9 chuy\u1ec3n \u0111\u1ed5i VPI Mellanox ConnectX-3 Pro h\u1ed7 tr\u1ee3 c\u1ea3 40GbE (m\u1ea1ng ph\u00f2ng th\u00ed nghi\u1ec7m ch\u00ednh) c\u0169ng nh\u01b0 Infiniband 56Gbps (m\u1ea1ng h\u1ecdc s\u00e2u) Ch\u00fang t\u00f4i \u0111\u00e3 c\u00f3 card tr\u00ean tay nh\u01b0ng s\u1eed d\u1ee5ng FDR Infiniband v\u1edbi RDMA r\u1ea5t ph\u1ed5 bi\u1ebfn v\u1edbi c\u00e1c m\u00e1y n\u00e0y. M\u1ea1ng 1GbE \/ 10GbE \u0111\u01a1n gi\u1ea3n l\u00e0 kh\u00f4ng th\u1ec3 cung c\u1ea5p cho c\u00e1c m\u00e1y n\u00e0y \u0111\u1ee7 nhanh. Ch\u00fang t\u00f4i \u0111\u00e3 c\u00e0i \u0111\u1eb7t b\u1ed9 chuy\u1ec3n \u0111\u1ed5i Intel Omni-Path trong ph\u00f2ng th\u00ed nghi\u1ec7m, \u0111\u00e2y s\u1ebd l\u00e0 k\u1ebft c\u1ea5u 100Gbps \u0111\u1ea7u ti\u00ean c\u1ee7a ch\u00fang t\u00f4i trong ph\u00f2ng th\u00ed nghi\u1ec7m.<\/p>\n<p><img decoding=\"async\" src=\"https:\/\/thegioimaychu.vn\/blog\/wp-content\/uploads\/2019\/09\/Mellanox-ConnectX-3-MCX354A-FCBT.jpg\" sizes=\"(max-width: 638px) 100vw, 638px\" srcset=\"https:\/\/thegioimaychu.vn\/blog\/wp-content\/uploads\/2019\/09\/Mellanox-ConnectX-3-MCX354A-FCBT.jpg 638w, https:\/\/www.servethehome.com\/wp-content\/uploads\/2013\/09\/Mellanox-ConnectX-3-MCX354A-FCBT-300x224.jpg 300w, https:\/\/www.servethehome.com\/wp-content\/uploads\/2013\/09\/Mellanox-ConnectX-3-MCX354A-FCBT-600x449.jpg 600w, https:\/\/www.servethehome.com\/wp-content\/uploads\/2013\/09\/Mellanox-ConnectX-3-MCX354A-FCBT-16x12.jpg 16w, https:\/\/www.servethehome.com\/wp-content\/uploads\/2013\/09\/Mellanox-ConnectX-3-MCX354A-FCBT-32x24.jpg 32w, https:\/\/www.servethehome.com\/wp-content\/uploads\/2013\/09\/Mellanox-ConnectX-3-MCX354A-FCBT-28x21.jpg 28w, https:\/\/www.servethehome.com\/wp-content\/uploads\/2013\/09\/Mellanox-ConnectX-3-MCX354A-FCBT-56x42.jpg 56w, https:\/\/www.servethehome.com\/wp-content\/uploads\/2013\/09\/Mellanox-ConnectX-3-MCX354A-FCBT-64x48.jpg 64w\" alt=\"Mellanox ConnectX-3 MCX354A-FCBT\"><\/p>\n<p>Mellanox ConnectX-3 Pro<\/p>\n<p>V\u1ec1 CPU v\u00e0 RAM, ch\u00fang t\u00f4i s\u1eed d\u1ee5ng 2 CPU Intel Xeon E5-2628L V4 v\u00e0 RAM DDR4 256GB ECC. Ch\u00fang t\u00f4i s\u1ebd l\u01b0u \u00fd r\u1eb1ng Intel Xeon E5-2650 V4 k\u00e9p   l\u00e0 chip ph\u1ed5 bi\u1ebfn cho c\u00e1c h\u1ec7 th\u1ed1ng n\u00e0y. Ch\u00fang l\u00e0 b\u1ed9 x\u1eed l\u00fd ch\u00ednh c\u1ea5p th\u1ea5p nh\u1ea5t h\u1ed7 tr\u1ee3 t\u1ed1c \u0111\u1ed9 QPI 9.6GT \/ s. Ch\u00fang t\u00f4i \u0111ang s\u1eed d\u1ee5ng CPU Intel Xeon E5-2628L V4 do thi\u1ebft k\u1ebf root \u0111\u01a1n mang l\u1ea1i l\u1ee3i \u00edch quan tr\u1ecdng kh\u00e1c, kh\u00f4ng c\u00f2n l\u01b0u l\u01b0\u1ee3ng QPI li\u00ean GPU. M\u1eb7c d\u00f9 ch\u00fang t\u00f4i \u0111\u00e3 nghe n\u00f3i ng\u01b0\u1eddi ta c\u00f3 th\u1ec3 s\u1eed d\u1ee5ng m\u1ed9t GPU duy nh\u1ea5t \u0111\u1ec3 cung c\u1ea5p n\u0103ng l\u01b0\u1ee3ng cho h\u1ec7 th\u1ed1ng, nh\u01b0ng ch\u00fang t\u00f4i v\u1eabn \u0111ang s\u1eed d\u1ee5ng hai GPU \u0111\u1ec3 c\u00f3 th\u00eam dung l\u01b0\u1ee3ng RAM b\u1eb1ng c\u00e1ch s\u1eed d\u1ee5ng RDIMM 16GB r\u1ebb ti\u1ec1n c\u1ee7a ch\u00fang t\u00f4i. C\u00e1c h\u1ec7 th\u1ed1ng n\u00e0y c\u00f3 th\u1ec3 m\u1ea5t t\u1ed1i \u0111a 24x DDR4 LRDIMM cho dung l\u01b0\u1ee3ng b\u1ed9 nh\u1edb l\u1edbn.<\/p>\n<p>Ch\u00fang t\u00f4i s\u1ebd s\u1edbm th\u1ef1c hi\u1ec7n m\u1ed9t ph\u1ea7n g\u1ed1c, nh\u01b0ng \u0111\u1ed1i v\u1edbi nh\u1eefng ng\u01b0\u1eddi h\u1ecdc s\u00e2u s\u1eed d\u1ee5ng c\u00e1c kh\u1ed1i x\u00e2y d\u1ef1ng nh\u01b0 NVIDIA nccl , m\u1ed9t g\u1ed1c PCIe ph\u1ed5 bi\u1ebfn l\u00e0 r\u1ea5t quan tr\u1ecdng. \u0110\u00f3 c\u0169ng l\u00e0 m\u1ed9t l\u00fd do m\u00e0 nhi\u1ec1u c\u00f4ng c\u1ee5 x\u00e2y d\u1ef1ng h\u1ecdc t\u1eadp s\u00e2u s\u1ebd kh\u00f4ng chuy\u1ec3n sang s\u1ed1 l\u01b0\u1ee3ng PCIe cao h\u01a1n nh\u01b0ng \u0111\u1ed9 tr\u1ec5 cao h\u01a1n \/ thi\u1ebft k\u1ebf h\u1ea1n ch\u1ebf h\u01a1n nh\u01b0 AMD EPYC v\u1edbi Infinity Fabric.<\/p>\n<h2>Chi ph\u00ed h\u1ec7 th\u1ed1ng<\/h2>\n<p>X\u00e9t v\u1ec1 s\u1ef1 c\u1ed1 chi ph\u00ed, \u0111\u00e2y l\u00e0 nh\u1eefng g\u00ec b\u1ea1n c\u00f3 th\u1ec3 th\u1ea5y n\u1ebfu b\u1ea1n \u0111ang s\u1eed d\u1ee5ng chip Intel E5-2650 V4:<\/p>\n<p><img decoding=\"async\" src=\"https:\/\/thegioimaychu.vn\/blog\/wp-content\/uploads\/2019\/09\/DeepLearning11-Cost-by-Component-and-Vendor-Area-Chart-1.jpg\" sizes=\"(max-width: 756px) 100vw, 756px\" srcset=\"https:\/\/thegioimaychu.vn\/blog\/wp-content\/uploads\/2019\/09\/DeepLearning11-Cost-by-Component-and-Vendor-Area-Chart-1.jpg 756w, https:\/\/www.servethehome.com\/wp-content\/uploads\/2017\/07\/DeepLearning11-Cost-by-Component-and-Vendor-Area-Chart-1-400x184.jpg 400w, https:\/\/www.servethehome.com\/wp-content\/uploads\/2017\/07\/DeepLearning11-Cost-by-Component-and-Vendor-Area-Chart-1-696x320.jpg 696w\" alt=\"DeepLearning11 Chi ph\u00ed theo Th\u00e0nh ph\u1ea7n v\u00e0 Khu v\u1ef1c Nh\u00e0 cung c\u1ea5p Bi\u1ec3u \u0111\u1ed3 1\"><\/p>\n<p>DeepLearning11 Chi ph\u00ed g\u1ea7n \u0111\u00fang theo Th\u00e0nh ph\u1ea7n v\u00e0 Khu v\u1ef1c nh\u00e0 cung c\u1ea5p Bi\u1ec3u \u0111\u1ed3 1<\/p>\n<p>Ph\u1ea7n n\u1ed5i b\u1eadt \u1edf \u0111\u00e2y l\u00e0 t\u1ed5ng chi ph\u00ed kho\u1ea3ng 16.500 \u0111\u00f4 la c\u00f3 th\u1eddi gian ho\u00e0n v\u1ed1n d\u01b0\u1edbi 90 ng\u00e0y so v\u1edbi c\u00e1c lo\u1ea1i \u0111\u1ed1i t\u01b0\u1ee3ng AWS g2.16xlarge. Ch\u00fang t\u00f4i s\u1ebd bao g\u1ed3m chi ph\u00ed l\u01b0u tr\u1eef b\u00ean d\u01b0\u1edbi \u0111\u1ec3 hi\u1ec3n th\u1ecb c\u00e1ch so s\u00e1nh tr\u00ean c\u01a1 s\u1edf TCO.<\/p>\n<p>So s\u00e1nh v\u00ed d\u1ee5 v\u1ec1 GPU DeepLearning11 10x v\u1edbi DeepLearning10 v\u1edbi GPU 8 x c\u1ee7a n\u00f3, b\u1ea1n c\u00f3 th\u1ec3 th\u1ea5y r\u1eb1ng vi\u1ec7c gi\u1ea3m hi\u1ec7u su\u1ea5t ~ 25% c\u00f3 chi ph\u00ed t\u01b0\u01a1ng \u0111\u1ed1i \u00edt v\u1ec1 chi ph\u00ed h\u1ec7 th\u1ed1ng:<\/p>\n<p><img decoding=\"async\" src=\"https:\/\/thegioimaychu.vn\/blog\/wp-content\/uploads\/2019\/09\/DeepLearning10-Approxmiate-Cost-By-Vendor-and-Component-Area-Chart.jpg\" sizes=\"(max-width: 753px) 100vw, 753px\" srcset=\"https:\/\/thegioimaychu.vn\/blog\/wp-content\/uploads\/2019\/09\/DeepLearning10-Approxmiate-Cost-By-Vendor-and-Component-Area-Chart.jpg 753w, https:\/\/www.servethehome.com\/wp-content\/uploads\/2017\/06\/DeepLearning10-Approxmiate-Cost-By-Vendor-and-Component-Area-Chart-400x172.jpg 400w, https:\/\/www.servethehome.com\/wp-content\/uploads\/2017\/06\/DeepLearning10-Approxmiate-Cost-By-Vendor-and-Component-Area-Chart-696x299.jpg 696w\" alt=\"DeepLearning10 Chi ph\u00ed x\u1ea5p x\u1ec9 theo nh\u00e0 cung c\u1ea5p v\u00e0 bi\u1ec3u \u0111\u1ed3 khu v\u1ef1c th\u00e0nh ph\u1ea7n\"><\/p>\n<p>DeepLearning10 Chi ph\u00ed g\u1ea7n \u0111\u00fang theo nh\u00e0 cung c\u1ea5p v\u00e0 bi\u1ec3u \u0111\u1ed3 khu v\u1ef1c th\u00e0nh ph\u1ea7n<\/p>\n<p>Nh\u01b0 m\u1ecdi ng\u01b0\u1eddi c\u00f3 th\u1ec3 t\u01b0\u1edfng t\u01b0\u1ee3ng, vi\u1ec7c th\u00eam nhi\u1ec1u GPU c\u00f3 ngh\u0129a l\u00e0 ph\u1ea7n tr\u00ean c\u1ee7a ph\u1ea7n c\u00f2n l\u1ea1i c\u1ee7a h\u1ec7 th\u1ed1ng \u0111\u01b0\u1ee3c kh\u1ea5u hao tr\u00ean nhi\u1ec1u GPU h\u01a1n. K\u1ebft qu\u1ea3 l\u00e0, n\u1ebfu \u1ee9ng d\u1ee5ng c\u1ee7a b\u1ea1n c\u00f3 quy m\u00f4 t\u1ed1t, h\u00e3y nh\u1eadn GPU 10 l\u1ea7n cho m\u1ed7i h\u1ec7 th\u1ed1ng.<\/p>\n<h2>DeepLearning11: C\u00e2n nh\u1eafc v\u1ec1 m\u00f4i tr\u01b0\u1eddng<\/h2>\n<p>H\u1ec7 th\u1ed1ng c\u1ee7a ch\u00fang t\u00f4i c\u00f3 b\u1ed1n PSU, c\u1ea7n thi\u1ebft cho c\u1ea5u h\u00ecnh GPU 10 x. \u0110\u1ec3 ki\u1ec3m tra \u0111i\u1ec1u n\u00e0y, ch\u00fang t\u00f4i cho ph\u00e9p h\u1ec7 th\u1ed1ng ch\u1ea1y v\u1edbi m\u1ed9t m\u00f4 h\u00ecnh kh\u1ed5ng l\u1ed3 (\u0111\u1ed1i v\u1edbi ch\u00fang t\u00f4i) trong v\u00e0i ng\u00e0y ch\u1ec9 \u0111\u1ec3 xem c\u00f3 bao nhi\u00eau n\u0103ng l\u01b0\u1ee3ng \u0111ang \u0111\u01b0\u1ee3c s\u1eed d\u1ee5ng. D\u01b0\u1edbi \u0111\u00e2y l\u00e0 m\u1ee9c ti\u00eau th\u1ee5 n\u0103ng l\u01b0\u1ee3ng c\u1ee7a <a class=\"iblink\" href=\"\/server\/gpu-system\/\">m\u00e1y ch\u1ee7 GPU<\/a> 10 x tr\u00f4ng nh\u01b0 \u0111\u01b0\u1ee3c \u0111o b\u1eb1ng PDU ch\u1ea1y kh\u1ed1i l\u01b0\u1ee3ng c\u00f4ng vi\u1ec7c GAN h\u00e0ng ch\u1ee5c c\u1ee7a ch\u00fang t\u00f4i:<\/p>\n<p><img decoding=\"async\" src=\"https:\/\/thegioimaychu.vn\/blog\/wp-content\/uploads\/2019\/09\/DeepLearning11-Power-Consumption-TF.jpg\" sizes=\"(max-width: 1010px) 100vw, 1010px\" srcset=\"https:\/\/thegioimaychu.vn\/blog\/wp-content\/uploads\/2019\/09\/DeepLearning11-Power-Consumption-TF.jpg 1010w, https:\/\/www.servethehome.com\/wp-content\/uploads\/2017\/07\/DeepLearning11-Power-Consumption-TF-400x155.jpg 400w, https:\/\/www.servethehome.com\/wp-content\/uploads\/2017\/07\/DeepLearning11-Power-Consumption-TF-800x310.jpg 800w, https:\/\/www.servethehome.com\/wp-content\/uploads\/2017\/07\/DeepLearning11-Power-Consumption-TF-696x270.jpg 696w\" alt=\"DeepLearning11 Ti\u00eau th\u1ee5 n\u0103ng l\u01b0\u1ee3ng TF\"><\/p>\n<p>DeepLearning11 Ti\u00eau th\u1ee5 n\u0103ng l\u01b0\u1ee3ng TF<\/p>\n<p>Kho\u1ea3ng 2600W ch\u1eafc ch\u1eafn kh\u00f4ng t\u1ec7. T\u00f9y thu\u1ed9c v\u00e0o n\u01a1i m\u00f4 h\u00ecnh \u0111\u01b0\u1ee3c \u0111\u00e0o t\u1ea1o, ch\u00fang t\u00f4i \u0111\u00e3 th\u1ea5y m\u1ee9c ti\u00eau th\u1ee5 n\u0103ng l\u01b0\u1ee3ng \u0111\u01b0\u1ee3c duy tr\u00ec cao h\u01a1n trong ph\u1ea1m vi 3.0-3.2kW tr\u00ean m\u00e1y n\u00e0y m\u00e0 kh\u00f4ng ch\u1ea1m v\u00e0o gi\u1edbi h\u1ea1n n\u0103ng l\u01b0\u1ee3ng tr\u00ean GPU.<\/p>\n<p>B\u1ea1n s\u1ebd nh\u1eadn th\u1ea5y c\u00f3 m\u1ed9t \u0111\u1ec9nh c\u1ef1c \u0111\u1ea1i 5278W, trong m\u1ed9t s\u1ed1 l\u1ea7n b\u1ebb kh\u00f3a m\u1eadt kh\u1ea9u, nhi\u1ec1u h\u01a1n v\u1ec1 \u0111i\u1ec1u n\u00e0y trong m\u1ed9t \u0111o\u1ea1n STH trong t\u01b0\u01a1ng lai. \u0110\u1ec9nh cao trong m\u1ed9t v\u00e0i tu\u1ea7n s\u1eed d\u1ee5ng c\u00e1c v\u1ea5n \u0111\u1ec1 v\u00e0 khu\u00f4n kh\u1ed5 kh\u00e1c nhau trong l\u0129nh v\u1ef1c h\u1ecdc t\u1eadp s\u00e2u ch\u1ec9 d\u01b0\u1edbi 4kW. S\u1eed d\u1ee5ng 4kW l\u00e0m c\u01a1 s\u1edf c\u1ee7a ch\u00fang t\u00f4i, ch\u00fang t\u00f4i c\u00f3 th\u1ec3 t\u00ednh to\u00e1n chi ph\u00ed colocation cho m\u1ed9t m\u00e1y nh\u01b0 v\u1eady m\u1ed9t c\u00e1ch d\u1ec5 d\u00e0ng.<\/p>\n<p><img decoding=\"async\" src=\"https:\/\/thegioimaychu.vn\/blog\/wp-content\/uploads\/2019\/09\/DeepLearning11-Cost-by-Component-and-Vendor-Area-Chart-with-Colocation.jpg\" sizes=\"(max-width: 756px) 100vw, 756px\" srcset=\"https:\/\/thegioimaychu.vn\/blog\/wp-content\/uploads\/2019\/09\/DeepLearning11-Cost-by-Component-and-Vendor-Area-Chart-with-Colocation.jpg 756w, https:\/\/www.servethehome.com\/wp-content\/uploads\/2017\/07\/DeepLearning11-Cost-by-Component-and-Vendor-Area-Chart-with-Colocation-400x184.jpg 400w, https:\/\/www.servethehome.com\/wp-content\/uploads\/2017\/07\/DeepLearning11-Cost-by-Component-and-Vendor-Area-Chart-with-Colocation-696x319.jpg 696w\" alt=\"DeepLearning11 Chi ph\u00ed theo Bi\u1ec3u \u0111\u1ed3 khu v\u1ef1c th\u00e0nh ph\u1ea7n v\u00e0 nh\u00e0 cung c\u1ea5p v\u1edbi Colocation\"><\/p>\n<p>DeepLearning11 Chi ph\u00ed theo Bi\u1ec3u \u0111\u1ed3 khu v\u1ef1c th\u00e0nh ph\u1ea7n v\u00e0 nh\u00e0 cung c\u1ea5p v\u1edbi Colocation<\/p>\n<p>Nh\u01b0 b\u1ea1n c\u00f3 th\u1ec3 th\u1ea5y, sau 12 th\u00e1ng, chi ph\u00ed colocation b\u1eaft \u0111\u1ea7u gi\u1ea3m b\u1edbt chi ph\u00ed ph\u1ea7n c\u1ee9ng. \u0110\u1ed1i v\u1edbi nh\u1eefng \u0111i\u1ec1u n\u00e0y, ch\u00fang t\u00f4i \u0111ang s\u1eed d\u1ee5ng chi ph\u00ed colocation ph\u00f2ng th\u00ed nghi\u1ec7m trung t\u00e2m d\u1eef li\u1ec7u th\u1ef1c t\u1ebf c\u1ee7a ch\u00fang t\u00f4i. N\u1ebfu b\u1ea1n mu\u1ed1n x\u00e2y d\u1ef1ng m\u1ed9t m\u00f4 h\u00ecnh chi ph\u00ed t\u01b0\u01a1ng t\u1ef1, ch\u00fang t\u00f4i s\u1eb5n l\u00f2ng cung c\u1ea5p th\u00f4ng tin li\u00ean h\u1ec7 v\u1edbi nh\u1eefng ng\u01b0\u1eddi ch\u00fang t\u00f4i s\u1eed d\u1ee5ng \u0111\u1ec3 b\u1ea1n c\u00f3 th\u1ec3 sao ch\u00e9p nh\u1eefng \u0111i\u1ec1u tr\u00ean. \u0110\u00e2y kh\u00f4ng ph\u1ea3i l\u00e0 nh\u1eefng con s\u1ed1 l\u00fd thuy\u1ebft, ch\u00fang t\u00f4i th\u1ef1c s\u1ef1 chi kho\u1ea3ng $ 1k \/ th\u00e1ng \u0111\u1ec3 ch\u1ea1y h\u1ec7 th\u1ed1ng n\u00e0y trong trung t\u00e2m d\u1eef li\u1ec7u.<\/p>\n<p>So s\u00e1nh \u1edf tr\u00ean v\u1edbi DeepLearning10 v\u1edbi GPU 8 x v\u00e0 b\u1ea1n c\u00f3 th\u1ec3 th\u1ea5y t\u00e1c \u0111\u1ed9ng c\u1ee7a vi\u1ec7c th\u00eam ~ 500W t\u00ednh to\u00e1n b\u1ed5 sung:<\/p>\n<p><img decoding=\"async\" src=\"https:\/\/thegioimaychu.vn\/blog\/wp-content\/uploads\/2019\/09\/DeepLearning10-Approxmiate-Cost-By-Vendor-and-Component-Area-Chart-12-Months-Ongoing.jpg\" sizes=\"(max-width: 752px) 100vw, 752px\" srcset=\"https:\/\/thegioimaychu.vn\/blog\/wp-content\/uploads\/2019\/09\/DeepLearning10-Approxmiate-Cost-By-Vendor-and-Component-Area-Chart-12-Months-Ongoing.jpg 752w, https:\/\/www.servethehome.com\/wp-content\/uploads\/2017\/06\/DeepLearning10-Approxmiate-Cost-By-Vendor-and-Component-Area-Chart-12-Months-Ongoing-400x182.jpg 400w, https:\/\/www.servethehome.com\/wp-content\/uploads\/2017\/06\/DeepLearning10-Approxmiate-Cost-By-Vendor-and-Component-Area-Chart-12-Months-Ongoing-696x317.jpg 696w\" alt=\"DeepLearning10 Chi ph\u00ed x\u1ea5p x\u1ec9 theo nh\u00e0 cung c\u1ea5p v\u00e0 khu v\u1ef1c th\u00e0nh ph\u1ea7n Bi\u1ec3u \u0111\u1ed3 12 th\u00e1ng \u0111ang di\u1ec5n ra\"><\/p>\n<p>DeepLearning10 Chi ph\u00ed g\u1ea7n \u0111\u00fang theo nh\u00e0 cung c\u1ea5p v\u00e0 khu v\u1ef1c th\u00e0nh ph\u1ea7n Bi\u1ec3u \u0111\u1ed3 12 th\u00e1ng \u0111ang di\u1ec5n ra<\/p>\n<p>Th\u00eam GPU b\u1ed5 sung th\u00eam chi ph\u00ed v\u1eadn h\u00e0nh ph\u00f9 h\u1ee3p v\u1edbi chi ph\u00ed h\u1ec7 th\u1ed1ng so v\u1edbi DeepLearning10. B\u01b0\u1edbc sang nh\u1eefng n\u0103m ti\u1ebfp theo, chi ph\u00ed colocation s\u1ebd v\u01b0\u1ee3t xa chi ph\u00ed ph\u1ea7n c\u1ee9ng.<\/p>\n<h2>DeepLearning11: T\u00e1c \u0111\u1ed9ng hi\u1ec7u su\u1ea5t<\/h2>\n<p>Ch\u00fang t\u00f4i mu\u1ed1n ch\u1ec9 cho b\u1ea1n th\u1ea5y m\u1ed9t ch\u00fat v\u1ec1 hi\u1ec7u su\u1ea5t m\u00e0 ch\u00fang t\u00f4i \u0111\u1ea1t \u0111\u01b0\u1ee3c t\u1eeb h\u1ec7 th\u1ed1ng m\u1edbi n\u00e0y. C\u00f3 m\u1ed9t s\u1ef1 kh\u00e1c bi\u1ec7t l\u1edbn gi\u1eefa h\u1ec7 th\u1ed1ng $ 1600 v\u00e0 h\u1ec7 th\u1ed1ng $ 16.000 + v\u00ec v\u1eady ch\u00fang t\u00f4i hy v\u1ecdng t\u00e1c \u0111\u1ed9ng s\u1ebd l\u1edbn t\u01b0\u01a1ng t\u1ef1. Ch\u00fang t\u00f4i \u0111\u00e3 l\u1ea5y tr\u01b0\u1eddng h\u1ee3p th\u1eed nghi\u1ec7m \u0111\u00e0o t\u1ea1o h\u00ecnh \u1ea3nh Tensorflow Generative Adversarial Network (GAN) m\u1eabu c\u1ee7a ch\u00fang t\u00f4i v\u00e0 ch\u1ea1y n\u00f3 tr\u00ean c\u00e1c th\u1ebb \u0111\u01a1n sau \u0111\u00f3 b\u01b0\u1edbc l\u00ean h\u1ec7 th\u1ed1ng GPU 10 x. Ch\u00fang t\u00f4i b\u00e0y t\u1ecf k\u1ebft qu\u1ea3 c\u1ee7a ch\u00fang t\u00f4i v\u1ec1 c\u00e1c chu k\u1ef3 \u0111\u00e0o t\u1ea1o m\u1ed7i ng\u00e0y.<\/p>\n<p><img decoding=\"async\" src=\"https:\/\/thegioimaychu.vn\/blog\/wp-content\/uploads\/2019\/09\/DeepLearning11-Tensorflow-GAN-Training-Example.jpg\" sizes=\"(max-width: 858px) 100vw, 858px\" srcset=\"https:\/\/thegioimaychu.vn\/blog\/wp-content\/uploads\/2019\/09\/DeepLearning11-Tensorflow-GAN-Training-Example.jpg 858w, https:\/\/www.servethehome.com\/wp-content\/uploads\/2017\/07\/DeepLearning11-Tensorflow-GAN-Training-Example-400x259.jpg 400w, https:\/\/www.servethehome.com\/wp-content\/uploads\/2017\/07\/DeepLearning11-Tensorflow-GAN-Training-Example-800x518.jpg 800w, https:\/\/www.servethehome.com\/wp-content\/uploads\/2017\/07\/DeepLearning11-Tensorflow-GAN-Training-Example-696x451.jpg 696w, https:\/\/www.servethehome.com\/wp-content\/uploads\/2017\/07\/DeepLearning11-Tensorflow-GAN-Training-Example-648x420.jpg 648w\" alt=\"V\u00ed d\u1ee5 \u0111\u00e0o t\u1ea1o GAN DeepLearning11 Tensorflow\"><\/p>\n<p>V\u00ed d\u1ee5 \u0111\u00e0o t\u1ea1o GAN DeepLearning11 Tensorflow<\/p>\n<p>\u0110\u00e2y l\u00e0 m\u1ed9t v\u00ed d\u1ee5 tuy\u1ec7t v\u1eddi v\u1ec1 c\u00e1ch th\u00eam $ 1400 tr\u1edf l\u00ean v\u00e0o gi\u00e1 mua c\u1ee7a h\u1ec7 th\u1ed1ng mang l\u1ea1i k\u1ebft qu\u1ea3 r\u00f5 r\u00e0ng. Trong khi m\u1ed9t NVIDIA GeForce GTX 1080 Ti duy nh\u1ea5t cho ph\u00e9p ch\u00fang t\u00f4i \u0111\u00e0o t\u1ea1o m\u00f4 h\u00ecnh m\u1ed9t l\u1ea7n trong t\u00e1m gi\u1edd, m\u1ed9t h\u1ed9p GPU 10 x cho ph\u00e9p ch\u00fang t\u00f4i luy\u1ec7n t\u1eadp tr\u00ean nh\u1ecbp l\u1edbn h\u01a1n h\u00e0ng gi\u1edd. N\u1ebfu b\u1ea1n mu\u1ed1n \u0111\u1ea1t \u0111\u01b0\u1ee3c ti\u1ebfn b\u1ed9 trong m\u1ed9t ng\u00e0y l\u00e0m vi\u1ec7c, m\u1ed9t h\u1ed9p l\u1edbn ho\u1eb7c m\u1ed9t c\u1ee5m c\u00e1c h\u1ed9p l\u1edbn s\u1ebd gi\u00fap \u00edch.<\/p>\n<h2>L\u1eddi cu\u1ed1i c\u00f9ng<\/h2>\n<p>Nh\u01b0 m\u1ecdi ng\u01b0\u1eddi c\u00f3 th\u1ec3 t\u01b0\u1edfng t\u01b0\u1ee3ng, DeepLearning10 v\u00e0 DeepLearning11 s\u1eed d\u1ee5ng r\u1ea5t nhi\u1ec1u s\u1ee9c m\u1ea1nh. Ch\u1ec9 ri\u00eang hai m\u00e1y ch\u1ee7 \u0111\u00f3 l\u00e0 trung b\u00ecnh tr\u00ean 5kW c\u00f4ng su\u1ea5t ph\u00f9 h\u1ee3p v\u1edbi m\u1ee9c t\u0103ng \u0111\u1ed9t bi\u1ebfn cao h\u01a1n nhi\u1ec1u. \u0110i\u1ec1u \u0111\u00f3 c\u00f3 \u00fd ngh\u0129a l\u1edbn \u0111\u1ed1i v\u1edbi vi\u1ec7c l\u01b0u tr\u1eef khi &#8220;kh\u1ed1i u&#8221; c\u1ed9ng th\u00eam 0,5 RU kh\u00f4ng \u0111\u00e1ng k\u1ec3 trong nhi\u1ec1u rack. H\u1ea7u h\u1ebft c\u00e1c rack colocation kh\u00f4ng th\u1ec3 cung c\u1ea5p 25kW + n\u0103ng l\u01b0\u1ee3ng v\u00e0 l\u00e0m m\u00e1t tr\u00ean m\u1ed7i rack \u0111\u1ec3 l\u1ea5p \u0111\u1ea7y ch\u00fang v\u1edbi c\u00e1c <a class=\"iblink\" href=\"\/server\/gpu-system\/\">m\u00e1y ch\u1ee7 GPU<\/a>. Ch\u00fang ta th\u01b0\u1eddng th\u1ea5y nh\u1eefng m\u00e1y ch\u1ee7 n\u00e0y \u0111\u01b0\u1ee3c l\u01b0u tr\u1eef v\u1edbi 2 m\u00e1y t\u00ednh GPU tr\u00ean m\u1ed7i rack 30A 208V, v\u00ec v\u1eady v\u1ea5n \u0111\u1ec1 l\u1eafp \u0111\u1eb7t v\u00e0 ch\u1ed7 tr\u1ed1ng tr\u1edf n\u00ean quan tr\u1ecdng.<\/p>\n<p>Cu\u1ed1i c\u00f9ng, ch\u00fang t\u00f4i mu\u1ed1n c\u00f3 m\u1ed9t h\u1ec7 th\u1ed1ng g\u1ed1c \u0111\u00e1ng c\u00f3 trong ph\u00f2ng lab v\u00e0 ch\u00fang t\u00f4i \u0111\u1ea1t \u0111i\u1ec1u \u0111\u00f3 v\u1edbi DeepLearning11 c\u00f9ng v\u1edbi GPU NVIDIA GTX 1080 Ti 11GB c\u1ee7a n\u00f3. V\u00ec ch\u00fang t\u00f4i ch\u1ee7 tr\u01b0\u01a1ng t\u0103ng k\u00edch th\u01b0\u1edbc GPU tr\u01b0\u1edbc, sau \u0111\u00f3 l\u00e0 s\u1ed1 GPU tr\u00ean m\u1ed7i m\u00e1y, sau \u0111\u00f3 \u0111\u1ebfn nhi\u1ec1u m\u00e1y, DeepLearning11 v\u1eeba l\u00e0 m\u1ed9t m\u00e1y \u0111\u01a1n h\u00e0ng \u0111\u1ea7u tuy\u1ec7t v\u1eddi nh\u01b0ng c\u0169ng l\u00e0 m\u1ed9t n\u1ec1n t\u1ea3ng \u0111\u1ec3 nh\u00e2n r\u1ed9ng ra nhi\u1ec1u m\u00e1y d\u1ef1a tr\u00ean thi\u1ebft k\u1ebf. C\u00f3 m\u1ed9t s\u1ed1 t\u00ednh n\u0103ng nh\u01b0 GPUDirect s\u1eed d\u1ee5ng RDMA r\u1ea5t t\u1ed1t tr\u00ean n\u1ec1n t\u1ea3ng n\u00e0y v\u1edbi gi\u1ea3 \u0111\u1ecbnh ph\u1ea7n m\u1ec1m v\u00e0 ph\u1ea7n c\u1ee9ng c\u1ee7a b\u1ea1n c\u00f3 th\u1ec3 h\u1ed7 tr\u1ee3 ch\u00fang. Ch\u00fang t\u00f4i th\u1ef1c t\u1ebf b\u1ecb gi\u1edbi h\u1ea1n b\u1edfi ng\u00e2n s\u00e1ch v\u00ec v\u1eady ch\u00fang t\u00f4i \u0111\u00e3 nh\u1eadn \u0111\u01b0\u1ee3c nh\u1eefng card t\u1ed1t nh\u1ea5t c\u00f3 th\u1ec3, GTX 1080 Ti.<\/p>\n<div>____<br \/><b>B\u00e0i vi\u1ebft li\u00ean quan<\/b><\/div>\n<ul>\n<li><a href=\"https:\/\/thegioimaychu.vn\/blog\/?post_type=post&#038;p=23322\">NVIDIA RTX PRO 4500 Blackwell Server Edition: B\u01b0\u1edbc nh\u1ea3y v\u1ecdt t\u1eeb th\u1ebf h\u1ec7 L4 cho h\u1ea1 t\u1ea7ng AI suy lu\u1eadn hi\u1ec7u n\u0103ng cao<\/a><\/li>\n<li><a href=\"https:\/\/thegioimaychu.vn\/blog\/?post_type=post&#038;p=22728\">Hu\u1ea5n luy\u1ec7n m\u00f4 h\u00ecnh h\u00e0ng tr\u0103m t\u1ef7 tham s\u1ed1 ngay t\u1ea1i b\u00e0n v\u1edbi MSI EdgeXpert<\/a><\/li>\n<li><a href=\"https:\/\/thegioimaychu.vn\/blog\/?post_type=post&#038;p=22612\">NVIDIA Merlin: T\u1ed5ng quan v\u1ec1 toolkit cho h\u1ec7 th\u1ed1ng g\u1ee3i \u00fd quy m\u00f4 l\u1edbn<\/a><\/li>\n<li><a href=\"https:\/\/thegioimaychu.vn\/blog\/?post_type=post&#038;p=22403\">Th\u1eed so s\u00e1nh m\u00e1y t\u00ednh DGX Spark v\u00e0 m\u1ed9t PC c\u1ea5u h\u00ecnh cao v\u1edbi GPU RTX 5080<\/a><\/li>\n<li><a href=\"https:\/\/thegioimaychu.vn\/blog\/?post_type=post&#038;p=22376\">OpenAI l\u1ea7n \u0111\u1ea7u ph\u00e1t h\u00e0nh mi\u1ec5n ph\u00ed m\u00f4 h\u00ecnh ng\u00f4n ng\u1eef m\u1edbi v\u1edbi t\u00ean g\u1ecdi l\u00e0 GPT-OSS<\/a><\/li>\n<li><a href=\"https:\/\/thegioimaychu.vn\/blog\/?post_type=post&#038;p=22258\">NVIDIA hi\u1ec7n \u0111ang cung c\u1ea5p nh\u1eefng d\u00f2ng GPU n\u00e0o?<\/a><\/li>\n<\/ul>\n<\/p>","protected":false},"excerpt":{"rendered":"<p>STH &ndash; H&ocirc;m nay ch&uacute;ng t&ocirc;i &#273;ang tr&igrave;nh b&agrave;y m&#7897;t b&#7843;n d&#7921;ng c&oacute; l&#7869; l&agrave; c&#7845;u h&igrave;nh h&#7885;c s&acirc;u &#273;&#432;&#7907;c t&igrave;m ki&#7871;m nhi&#7873;u nh&#7845;t hi&#7879;n nay. DeepLearning11 c&oacute; 10 GPU NVIDIA GeForce GTX 1080 Ti 11GB, Mellanox Infiniband v&agrave; ph&ugrave; h&#7907;p v&#7899;i m&#7897;t form factor 4.5U nh&#7887; g&#7885;n. Ngo&agrave;i ra c&ograve;n c&oacute; m&#7897;t s&#7921; kh&aacute;c&#8230;<\/p>\n","protected":false},"author":2,"featured_media":2284,"comment_status":"closed","ping_status":"closed","sticky":false,"template":"","format":"standard","meta":{"inline_featured_image":false,"footnotes":""},"categories":[3,44],"tags":[7,12,304,671,1316],"class_list":["post-2283","post","type-post","status-publish","format-standard","has-post-thumbnail","hentry","category-ai-hpc","category-product","tag-ai","tag-deep-learning","tag-geforce","tag-gpu","tag-supermicro"],"yoast_head":"<!-- This site is optimized with the Yoast SEO plugin v28.3 - https:\/\/yoast.com\/product\/yoast-seo-wordpress\/ -->\n<title>DeepLearning11: H\u1ec7 th\u1ed1ng Deep Learning tham kh\u1ea3o v\u1edbi 10x NVIDIA GTX 1080 Ti - Blog | TheGioiMayChu<\/title>\n<meta name=\"robots\" content=\"index, follow, max-snippet:-1, max-image-preview:large, max-video-preview:-1\" \/>\n<link rel=\"canonical\" href=\"https:\/\/thegioimaychu.vn\/blog\/ai-hpc\/deeplearning11-he-thong-deep-learning-tham-khao-voi-10x-nvidia-gtx-1080-ti-p2283\/\" \/>\n<meta name=\"twitter:card\" content=\"summary_large_image\" \/>\n<meta name=\"twitter:title\" content=\"DeepLearning11: H\u1ec7 th\u1ed1ng Deep Learning tham kh\u1ea3o v\u1edbi 10x NVIDIA GTX 1080 Ti - Blog | TheGioiMayChu\" \/>\n<meta name=\"twitter:description\" content=\"STH &ndash; H&ocirc;m nay ch&uacute;ng t&ocirc;i &#273;ang tr&igrave;nh b&agrave;y m&#7897;t b&#7843;n d&#7921;ng c&oacute; l&#7869; l&agrave; c&#7845;u h&igrave;nh h&#7885;c s&acirc;u &#273;&#432;&#7907;c t&igrave;m ki&#7871;m nhi&#7873;u nh&#7845;t hi&#7879;n nay. DeepLearning11 c&oacute; 10 GPU NVIDIA GeForce GTX 1080 Ti 11GB, Mellanox Infiniband v&agrave; ph&ugrave; h&#7907;p v&#7899;i m&#7897;t form factor 4.5U nh&#7887; g&#7885;n. Ngo&agrave;i ra c&ograve;n c&oacute; m&#7897;t s&#7921; kh&aacute;c...\" \/>\n<meta name=\"twitter:image\" content=\"https:\/\/thegioimaychu.vn\/blog\/wp-content\/uploads\/2019\/09\/10x-NVIDIA-GTX-1080-TI-FE-Plus-Mellanox-Top1.jpg\" \/>\n<meta name=\"twitter:label1\" content=\"Written by\" \/>\n\t<meta name=\"twitter:data1\" content=\"Vanito Hoang\" \/>\n\t<meta name=\"twitter:label2\" content=\"Est. reading time\" \/>\n\t<meta name=\"twitter:data2\" content=\"14 minutes\" \/>\n<script type=\"application\/ld+json\" class=\"yoast-schema-graph\">{\"@context\":\"https:\\\/\\\/schema.org\",\"@graph\":[{\"@type\":\"Article\",\"@id\":\"https:\\\/\\\/thegioimaychu.vn\\\/blog\\\/ai-hpc\\\/deeplearning11-he-thong-deep-learning-tham-khao-voi-10x-nvidia-gtx-1080-ti-p2283\\\/#article\",\"isPartOf\":{\"@id\":\"https:\\\/\\\/thegioimaychu.vn\\\/blog\\\/ai-hpc\\\/deeplearning11-he-thong-deep-learning-tham-khao-voi-10x-nvidia-gtx-1080-ti-p2283\\\/\"},\"author\":{\"name\":\"Vanito Hoang\",\"@id\":\"https:\\\/\\\/thegioimaychu.vn\\\/blog\\\/#\\\/schema\\\/person\\\/feccec179330cfebfaa805639a357d18\"},\"headline\":\"DeepLearning11: H\u1ec7 th\u1ed1ng Deep Learning tham kh\u1ea3o v\u1edbi 10x NVIDIA GTX 1080 Ti\",\"datePublished\":\"2019-09-10T10:13:02+00:00\",\"dateModified\":\"2019-09-11T00:33:54+00:00\",\"mainEntityOfPage\":{\"@id\":\"https:\\\/\\\/thegioimaychu.vn\\\/blog\\\/ai-hpc\\\/deeplearning11-he-thong-deep-learning-tham-khao-voi-10x-nvidia-gtx-1080-ti-p2283\\\/\"},\"wordCount\":2862,\"publisher\":{\"@id\":\"https:\\\/\\\/thegioimaychu.vn\\\/blog\\\/#organization\"},\"image\":{\"@id\":\"https:\\\/\\\/thegioimaychu.vn\\\/blog\\\/ai-hpc\\\/deeplearning11-he-thong-deep-learning-tham-khao-voi-10x-nvidia-gtx-1080-ti-p2283\\\/#primaryimage\"},\"thumbnailUrl\":\"https:\\\/\\\/thegioimaychu.vn\\\/blog\\\/wp-content\\\/uploads\\\/2019\\\/09\\\/10x-NVIDIA-GTX-1080-TI-FE-Plus-Mellanox-Top1.jpg\",\"keywords\":[\"ai\",\"deep learning\",\"geforce\",\"gpu\",\"Supermicro\"],\"articleSection\":[\"AI - HPC\",\"S\u1ea3n ph\u1ea9m\"],\"inLanguage\":\"en-US\"},{\"@type\":\"WebPage\",\"@id\":\"https:\\\/\\\/thegioimaychu.vn\\\/blog\\\/ai-hpc\\\/deeplearning11-he-thong-deep-learning-tham-khao-voi-10x-nvidia-gtx-1080-ti-p2283\\\/\",\"url\":\"https:\\\/\\\/thegioimaychu.vn\\\/blog\\\/ai-hpc\\\/deeplearning11-he-thong-deep-learning-tham-khao-voi-10x-nvidia-gtx-1080-ti-p2283\\\/\",\"name\":\"DeepLearning11: H\u1ec7 th\u1ed1ng Deep Learning tham kh\u1ea3o v\u1edbi 10x NVIDIA GTX 1080 Ti - Blog | TheGioiMayChu\",\"isPartOf\":{\"@id\":\"https:\\\/\\\/thegioimaychu.vn\\\/blog\\\/#website\"},\"primaryImageOfPage\":{\"@id\":\"https:\\\/\\\/thegioimaychu.vn\\\/blog\\\/ai-hpc\\\/deeplearning11-he-thong-deep-learning-tham-khao-voi-10x-nvidia-gtx-1080-ti-p2283\\\/#primaryimage\"},\"image\":{\"@id\":\"https:\\\/\\\/thegioimaychu.vn\\\/blog\\\/ai-hpc\\\/deeplearning11-he-thong-deep-learning-tham-khao-voi-10x-nvidia-gtx-1080-ti-p2283\\\/#primaryimage\"},\"thumbnailUrl\":\"https:\\\/\\\/thegioimaychu.vn\\\/blog\\\/wp-content\\\/uploads\\\/2019\\\/09\\\/10x-NVIDIA-GTX-1080-TI-FE-Plus-Mellanox-Top1.jpg\",\"datePublished\":\"2019-09-10T10:13:02+00:00\",\"dateModified\":\"2019-09-11T00:33:54+00:00\",\"breadcrumb\":{\"@id\":\"https:\\\/\\\/thegioimaychu.vn\\\/blog\\\/ai-hpc\\\/deeplearning11-he-thong-deep-learning-tham-khao-voi-10x-nvidia-gtx-1080-ti-p2283\\\/#breadcrumb\"},\"inLanguage\":\"en-US\",\"potentialAction\":[{\"@type\":\"ReadAction\",\"target\":[\"https:\\\/\\\/thegioimaychu.vn\\\/blog\\\/ai-hpc\\\/deeplearning11-he-thong-deep-learning-tham-khao-voi-10x-nvidia-gtx-1080-ti-p2283\\\/\"]}]},{\"@type\":\"ImageObject\",\"inLanguage\":\"en-US\",\"@id\":\"https:\\\/\\\/thegioimaychu.vn\\\/blog\\\/ai-hpc\\\/deeplearning11-he-thong-deep-learning-tham-khao-voi-10x-nvidia-gtx-1080-ti-p2283\\\/#primaryimage\",\"url\":\"https:\\\/\\\/thegioimaychu.vn\\\/blog\\\/wp-content\\\/uploads\\\/2019\\\/09\\\/10x-NVIDIA-GTX-1080-TI-FE-Plus-Mellanox-Top1.jpg\",\"contentUrl\":\"https:\\\/\\\/thegioimaychu.vn\\\/blog\\\/wp-content\\\/uploads\\\/2019\\\/09\\\/10x-NVIDIA-GTX-1080-TI-FE-Plus-Mellanox-Top1.jpg\",\"width\":790,\"height\":525},{\"@type\":\"BreadcrumbList\",\"@id\":\"https:\\\/\\\/thegioimaychu.vn\\\/blog\\\/ai-hpc\\\/deeplearning11-he-thong-deep-learning-tham-khao-voi-10x-nvidia-gtx-1080-ti-p2283\\\/#breadcrumb\",\"itemListElement\":[{\"@type\":\"ListItem\",\"position\":1,\"name\":\"Trang ch\u1ee7\",\"item\":\"https:\\\/\\\/thegioimaychu.vn\\\/blog\\\/\"},{\"@type\":\"ListItem\",\"position\":2,\"name\":\"AI - HPC\",\"item\":\"https:\\\/\\\/thegioimaychu.vn\\\/blog\\\/.\\\/ai-hpc\\\/\"},{\"@type\":\"ListItem\",\"position\":3,\"name\":\"DeepLearning11: H\u1ec7 th\u1ed1ng Deep Learning tham kh\u1ea3o v\u1edbi 10x NVIDIA GTX 1080 Ti\"}]},{\"@type\":\"WebSite\",\"@id\":\"https:\\\/\\\/thegioimaychu.vn\\\/blog\\\/#website\",\"url\":\"https:\\\/\\\/thegioimaychu.vn\\\/blog\\\/\",\"name\":\"Blog | TheGioiMayChu\",\"description\":\"C\u1eadp nh\u1eadt xu h\u01b0\u1edbng c\u00f4ng ngh\u1ec7 m\u00e1y ch\u1ee7\",\"publisher\":{\"@id\":\"https:\\\/\\\/thegioimaychu.vn\\\/blog\\\/#organization\"},\"potentialAction\":[{\"@type\":\"SearchAction\",\"target\":{\"@type\":\"EntryPoint\",\"urlTemplate\":\"https:\\\/\\\/thegioimaychu.vn\\\/blog\\\/?s={search_term_string}\"},\"query-input\":{\"@type\":\"PropertyValueSpecification\",\"valueRequired\":true,\"valueName\":\"search_term_string\"}}],\"inLanguage\":\"en-US\"},{\"@type\":\"Organization\",\"@id\":\"https:\\\/\\\/thegioimaychu.vn\\\/blog\\\/#organization\",\"name\":\"Blog | TheGioiMayChu\",\"url\":\"https:\\\/\\\/thegioimaychu.vn\\\/blog\\\/\",\"logo\":{\"@type\":\"ImageObject\",\"inLanguage\":\"en-US\",\"@id\":\"https:\\\/\\\/thegioimaychu.vn\\\/blog\\\/#\\\/schema\\\/logo\\\/image\\\/\",\"url\":\"https:\\\/\\\/thegioimaychu.vn\\\/blog\\\/wp-content\\\/uploads\\\/2018\\\/08\\\/cropped-ntcnlog-logo-2.png\",\"contentUrl\":\"https:\\\/\\\/thegioimaychu.vn\\\/blog\\\/wp-content\\\/uploads\\\/2018\\\/08\\\/cropped-ntcnlog-logo-2.png\",\"width\":500,\"height\":73,\"caption\":\"Blog | TheGioiMayChu\"},\"image\":{\"@id\":\"https:\\\/\\\/thegioimaychu.vn\\\/blog\\\/#\\\/schema\\\/logo\\\/image\\\/\"},\"sameAs\":[\"https:\\\/\\\/www.facebook.com\\\/thegioimaychu\\\/\"]},{\"@type\":\"Person\",\"@id\":\"https:\\\/\\\/thegioimaychu.vn\\\/blog\\\/#\\\/schema\\\/person\\\/feccec179330cfebfaa805639a357d18\",\"name\":\"Vanito Hoang\",\"image\":{\"@type\":\"ImageObject\",\"inLanguage\":\"en-US\",\"@id\":\"https:\\\/\\\/secure.gravatar.com\\\/avatar\\\/e8187f37f72d74f9380a1367c8c530debee44f226a3301f0765c68f7cc1f3fa4?s=96&d=mm&r=g\",\"url\":\"https:\\\/\\\/secure.gravatar.com\\\/avatar\\\/e8187f37f72d74f9380a1367c8c530debee44f226a3301f0765c68f7cc1f3fa4?s=96&d=mm&r=g\",\"contentUrl\":\"https:\\\/\\\/secure.gravatar.com\\\/avatar\\\/e8187f37f72d74f9380a1367c8c530debee44f226a3301f0765c68f7cc1f3fa4?s=96&d=mm&r=g\",\"caption\":\"Vanito Hoang\"},\"url\":\"https:\\\/\\\/thegioimaychu.vn\\\/blog\\\/author\\\/vanito\\\/\"}]}<\/script>\n<!-- \/ Yoast SEO plugin. -->","yoast_head_json":{"title":"DeepLearning11: H\u1ec7 th\u1ed1ng Deep Learning tham kh\u1ea3o v\u1edbi 10x NVIDIA GTX 1080 Ti - Blog | TheGioiMayChu","robots":{"index":"index","follow":"follow","max-snippet":"max-snippet:-1","max-image-preview":"max-image-preview:large","max-video-preview":"max-video-preview:-1"},"canonical":"https:\/\/thegioimaychu.vn\/blog\/ai-hpc\/deeplearning11-he-thong-deep-learning-tham-khao-voi-10x-nvidia-gtx-1080-ti-p2283\/","twitter_card":"summary_large_image","twitter_title":"DeepLearning11: H\u1ec7 th\u1ed1ng Deep Learning tham kh\u1ea3o v\u1edbi 10x NVIDIA GTX 1080 Ti - Blog | TheGioiMayChu","twitter_description":"STH &ndash; H&ocirc;m nay ch&uacute;ng t&ocirc;i &#273;ang tr&igrave;nh b&agrave;y m&#7897;t b&#7843;n d&#7921;ng c&oacute; l&#7869; l&agrave; c&#7845;u h&igrave;nh h&#7885;c s&acirc;u &#273;&#432;&#7907;c t&igrave;m ki&#7871;m nhi&#7873;u nh&#7845;t hi&#7879;n nay. DeepLearning11 c&oacute; 10 GPU NVIDIA GeForce GTX 1080 Ti 11GB, Mellanox Infiniband v&agrave; ph&ugrave; h&#7907;p v&#7899;i m&#7897;t form factor 4.5U nh&#7887; g&#7885;n. Ngo&agrave;i ra c&ograve;n c&oacute; m&#7897;t s&#7921; kh&aacute;c...","twitter_image":"https:\/\/thegioimaychu.vn\/blog\/wp-content\/uploads\/2019\/09\/10x-NVIDIA-GTX-1080-TI-FE-Plus-Mellanox-Top1.jpg","twitter_misc":{"Written by":"Vanito Hoang","Est. reading time":"14 minutes"},"schema":{"@context":"https:\/\/schema.org","@graph":[{"@type":"Article","@id":"https:\/\/thegioimaychu.vn\/blog\/ai-hpc\/deeplearning11-he-thong-deep-learning-tham-khao-voi-10x-nvidia-gtx-1080-ti-p2283\/#article","isPartOf":{"@id":"https:\/\/thegioimaychu.vn\/blog\/ai-hpc\/deeplearning11-he-thong-deep-learning-tham-khao-voi-10x-nvidia-gtx-1080-ti-p2283\/"},"author":{"name":"Vanito Hoang","@id":"https:\/\/thegioimaychu.vn\/blog\/#\/schema\/person\/feccec179330cfebfaa805639a357d18"},"headline":"DeepLearning11: H\u1ec7 th\u1ed1ng Deep Learning tham kh\u1ea3o v\u1edbi 10x NVIDIA GTX 1080 Ti","datePublished":"2019-09-10T10:13:02+00:00","dateModified":"2019-09-11T00:33:54+00:00","mainEntityOfPage":{"@id":"https:\/\/thegioimaychu.vn\/blog\/ai-hpc\/deeplearning11-he-thong-deep-learning-tham-khao-voi-10x-nvidia-gtx-1080-ti-p2283\/"},"wordCount":2862,"publisher":{"@id":"https:\/\/thegioimaychu.vn\/blog\/#organization"},"image":{"@id":"https:\/\/thegioimaychu.vn\/blog\/ai-hpc\/deeplearning11-he-thong-deep-learning-tham-khao-voi-10x-nvidia-gtx-1080-ti-p2283\/#primaryimage"},"thumbnailUrl":"https:\/\/thegioimaychu.vn\/blog\/wp-content\/uploads\/2019\/09\/10x-NVIDIA-GTX-1080-TI-FE-Plus-Mellanox-Top1.jpg","keywords":["ai","deep learning","geforce","gpu","Supermicro"],"articleSection":["AI - HPC","S\u1ea3n ph\u1ea9m"],"inLanguage":"en-US"},{"@type":"WebPage","@id":"https:\/\/thegioimaychu.vn\/blog\/ai-hpc\/deeplearning11-he-thong-deep-learning-tham-khao-voi-10x-nvidia-gtx-1080-ti-p2283\/","url":"https:\/\/thegioimaychu.vn\/blog\/ai-hpc\/deeplearning11-he-thong-deep-learning-tham-khao-voi-10x-nvidia-gtx-1080-ti-p2283\/","name":"DeepLearning11: H\u1ec7 th\u1ed1ng Deep Learning tham kh\u1ea3o v\u1edbi 10x NVIDIA GTX 1080 Ti - Blog | TheGioiMayChu","isPartOf":{"@id":"https:\/\/thegioimaychu.vn\/blog\/#website"},"primaryImageOfPage":{"@id":"https:\/\/thegioimaychu.vn\/blog\/ai-hpc\/deeplearning11-he-thong-deep-learning-tham-khao-voi-10x-nvidia-gtx-1080-ti-p2283\/#primaryimage"},"image":{"@id":"https:\/\/thegioimaychu.vn\/blog\/ai-hpc\/deeplearning11-he-thong-deep-learning-tham-khao-voi-10x-nvidia-gtx-1080-ti-p2283\/#primaryimage"},"thumbnailUrl":"https:\/\/thegioimaychu.vn\/blog\/wp-content\/uploads\/2019\/09\/10x-NVIDIA-GTX-1080-TI-FE-Plus-Mellanox-Top1.jpg","datePublished":"2019-09-10T10:13:02+00:00","dateModified":"2019-09-11T00:33:54+00:00","breadcrumb":{"@id":"https:\/\/thegioimaychu.vn\/blog\/ai-hpc\/deeplearning11-he-thong-deep-learning-tham-khao-voi-10x-nvidia-gtx-1080-ti-p2283\/#breadcrumb"},"inLanguage":"en-US","potentialAction":[{"@type":"ReadAction","target":["https:\/\/thegioimaychu.vn\/blog\/ai-hpc\/deeplearning11-he-thong-deep-learning-tham-khao-voi-10x-nvidia-gtx-1080-ti-p2283\/"]}]},{"@type":"ImageObject","inLanguage":"en-US","@id":"https:\/\/thegioimaychu.vn\/blog\/ai-hpc\/deeplearning11-he-thong-deep-learning-tham-khao-voi-10x-nvidia-gtx-1080-ti-p2283\/#primaryimage","url":"https:\/\/thegioimaychu.vn\/blog\/wp-content\/uploads\/2019\/09\/10x-NVIDIA-GTX-1080-TI-FE-Plus-Mellanox-Top1.jpg","contentUrl":"https:\/\/thegioimaychu.vn\/blog\/wp-content\/uploads\/2019\/09\/10x-NVIDIA-GTX-1080-TI-FE-Plus-Mellanox-Top1.jpg","width":790,"height":525},{"@type":"BreadcrumbList","@id":"https:\/\/thegioimaychu.vn\/blog\/ai-hpc\/deeplearning11-he-thong-deep-learning-tham-khao-voi-10x-nvidia-gtx-1080-ti-p2283\/#breadcrumb","itemListElement":[{"@type":"ListItem","position":1,"name":"Trang ch\u1ee7","item":"https:\/\/thegioimaychu.vn\/blog\/"},{"@type":"ListItem","position":2,"name":"AI - HPC","item":"https:\/\/thegioimaychu.vn\/blog\/.\/ai-hpc\/"},{"@type":"ListItem","position":3,"name":"DeepLearning11: H\u1ec7 th\u1ed1ng Deep Learning tham kh\u1ea3o v\u1edbi 10x NVIDIA GTX 1080 Ti"}]},{"@type":"WebSite","@id":"https:\/\/thegioimaychu.vn\/blog\/#website","url":"https:\/\/thegioimaychu.vn\/blog\/","name":"Blog | TheGioiMayChu","description":"C\u1eadp nh\u1eadt xu h\u01b0\u1edbng c\u00f4ng ngh\u1ec7 m\u00e1y ch\u1ee7","publisher":{"@id":"https:\/\/thegioimaychu.vn\/blog\/#organization"},"potentialAction":[{"@type":"SearchAction","target":{"@type":"EntryPoint","urlTemplate":"https:\/\/thegioimaychu.vn\/blog\/?s={search_term_string}"},"query-input":{"@type":"PropertyValueSpecification","valueRequired":true,"valueName":"search_term_string"}}],"inLanguage":"en-US"},{"@type":"Organization","@id":"https:\/\/thegioimaychu.vn\/blog\/#organization","name":"Blog | TheGioiMayChu","url":"https:\/\/thegioimaychu.vn\/blog\/","logo":{"@type":"ImageObject","inLanguage":"en-US","@id":"https:\/\/thegioimaychu.vn\/blog\/#\/schema\/logo\/image\/","url":"https:\/\/thegioimaychu.vn\/blog\/wp-content\/uploads\/2018\/08\/cropped-ntcnlog-logo-2.png","contentUrl":"https:\/\/thegioimaychu.vn\/blog\/wp-content\/uploads\/2018\/08\/cropped-ntcnlog-logo-2.png","width":500,"height":73,"caption":"Blog | TheGioiMayChu"},"image":{"@id":"https:\/\/thegioimaychu.vn\/blog\/#\/schema\/logo\/image\/"},"sameAs":["https:\/\/www.facebook.com\/thegioimaychu\/"]},{"@type":"Person","@id":"https:\/\/thegioimaychu.vn\/blog\/#\/schema\/person\/feccec179330cfebfaa805639a357d18","name":"Vanito Hoang","image":{"@type":"ImageObject","inLanguage":"en-US","@id":"https:\/\/secure.gravatar.com\/avatar\/e8187f37f72d74f9380a1367c8c530debee44f226a3301f0765c68f7cc1f3fa4?s=96&d=mm&r=g","url":"https:\/\/secure.gravatar.com\/avatar\/e8187f37f72d74f9380a1367c8c530debee44f226a3301f0765c68f7cc1f3fa4?s=96&d=mm&r=g","contentUrl":"https:\/\/secure.gravatar.com\/avatar\/e8187f37f72d74f9380a1367c8c530debee44f226a3301f0765c68f7cc1f3fa4?s=96&d=mm&r=g","caption":"Vanito Hoang"},"url":"https:\/\/thegioimaychu.vn\/blog\/author\/vanito\/"}]}},"_links":{"self":[{"href":"https:\/\/thegioimaychu.vn\/blog\/wp-json\/wp\/v2\/posts\/2283","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/thegioimaychu.vn\/blog\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/thegioimaychu.vn\/blog\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/thegioimaychu.vn\/blog\/wp-json\/wp\/v2\/users\/2"}],"replies":[{"embeddable":true,"href":"https:\/\/thegioimaychu.vn\/blog\/wp-json\/wp\/v2\/comments?post=2283"}],"version-history":[{"count":3,"href":"https:\/\/thegioimaychu.vn\/blog\/wp-json\/wp\/v2\/posts\/2283\/revisions"}],"predecessor-version":[{"id":2297,"href":"https:\/\/thegioimaychu.vn\/blog\/wp-json\/wp\/v2\/posts\/2283\/revisions\/2297"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/thegioimaychu.vn\/blog\/wp-json\/wp\/v2\/media\/2284"}],"wp:attachment":[{"href":"https:\/\/thegioimaychu.vn\/blog\/wp-json\/wp\/v2\/media?parent=2283"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/thegioimaychu.vn\/blog\/wp-json\/wp\/v2\/categories?post=2283"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/thegioimaychu.vn\/blog\/wp-json\/wp\/v2\/tags?post=2283"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}