{"id":56296,"date":"2020-06-08T17:08:53","date_gmt":"2020-06-08T15:08:53","guid":{"rendered":"https:\/\/www.planet3dnow.de\/cms\/?p=56296"},"modified":"2020-06-08T17:08:53","modified_gmt":"2020-06-08T15:08:53","slug":"amd-epyc-processors-ecosystem-continues-to-grow-with-integration-into-new-nvidia-dgx-a100","status":"publish","type":"post","link":"https:\/\/www.planet3dnow.de\/cms\/56296-amd-epyc-processors-ecosystem-continues-to-grow-with-integration-into-new-nvidia-dgx-a100\/","title":{"rendered":"AMD<\/span> EPYC<\/span>\u2122 Processors Ecosystem Continues to Grow with Integration into New NVIDIA<\/span> DGX<\/span> A100<\/span>"},"content":{"rendered":"

<\/p>

New NVIDIA<\/span> DGX<\/span> A100<\/span> sys\u00adtem harnes\u00adses PCIe\u00ae 4 sup\u00adport, high-core count, and over\u00adall per\u00adfor\u00admance of 2nd Gen AMD<\/span> EPYC<\/span>\u2122 Pro\u00adces\u00adsors to crea\u00adte the world\u2019s most advan\u00adced AI<\/span> system<\/b><\/center>\n

SANTA<\/span> CLARA<\/span>, Calif., June 08, 2020 (GLOBE<\/span> NEWSWIRE<\/span>) \u2014 AMD<\/span><\/a> (NASDAQ<\/span>: AMD<\/span>) today announ\u00adced the NVIDIA<\/span> DGX<\/span>\u2122 A100<\/span><\/a>, the third gene\u00adra\u00adti\u00adon of the world\u2019s most advan\u00adced AI<\/span> sys\u00adtem, is the latest high-per\u00adfor\u00admance com\u00adpu\u00adting sys\u00adtem fea\u00adturing 2nd Gen AMD<\/span> EPYC<\/span>\u2122 pro\u00adces\u00adsors. Deli\u00adve\u00adring 5 peta\u00adflops of AI<\/span> per\u00adfor\u00admance, the ela\u00ads\u00adtic archi\u00adtec\u00adtu\u00adre of the NVIDIA<\/span> DGX<\/span> A100<\/span> enables enter\u00adpri\u00adses to acce\u00adle\u00adra\u00adte diver\u00adse AI<\/span> workloads such as data ana\u00adly\u00adtics, trai\u00adning, and inference.<\/p>\n

NVIDIA<\/span> DGX<\/span> A100<\/span> lever\u00ada\u00adges the high-per\u00adfor\u00admance capa\u00adbi\u00adli\u00adties, 128 cores, DDR4-3200MHz and PCIe\u00ae 4 sup\u00adport from two AMD<\/span> EPYC<\/span> 7742 pro\u00adces\u00adsors run\u00adning at speeds up to 3.4 GHz\u00b9. The 2nd Gen AMD<\/span> EPYC<\/span> pro\u00adces\u00adsor is the first and only cur\u00adrent x86-archi\u00adtec\u00adtu\u00adre ser\u00adver pro\u00adces\u00adsor that sup\u00adports PCIe\u00ae 4, pro\u00advi\u00adding lea\u00adder\u00adship high-band\u00adwidth I\/O that\u2019s cri\u00adti\u00adcal for high per\u00adfor\u00admance com\u00adpu\u00adting and con\u00adnec\u00adtions bet\u00adween the CPU<\/span> and other devices like GPUs.<\/p>\n

\u201c<\/span>Only 2nd Gen AMD<\/span> EPYC<\/span> pro\u00adces\u00adsors can pro\u00advi\u00adde up to 64 cores and 128 lanes of PCIe 4 inter\u00adcon\u00adnec\u00adti\u00advi\u00adty in a sin\u00adgle x86 data cen\u00adter pro\u00adces\u00adsor, and we\u2019re exci\u00adted to see how the power of the NVIDIA<\/span> DGX<\/span> A100<\/span> sys\u00adtem enables the I\/O band\u00adwidth to be effec\u00adtively dou\u00adbled,\u201d said Rag\u00adhu Nam\u00adbi\u00adar, cor\u00adpo\u00adra\u00adte vice pre\u00adsi\u00addent, data cen\u00adter eco\u00adsys\u00adtems and appli\u00adca\u00adti\u00adon engi\u00adnee\u00adring, AMD<\/span>. \u201cWith 2nd Gen AMD<\/span> EPYC<\/span> pro\u00adces\u00adsors, our part\u00adners and cus\u00adto\u00admers can maxi\u00admi\u00adze per\u00adfor\u00admance and cost effi\u00adci\u00aden\u00adci\u00ades in hete\u00adro\u00adge\u00adneous com\u00adpu\u00adting, vir\u00adtua\u00adli\u00adzed and hyper con\u00adver\u00adged infra\u00adstruc\u00adtu\u00adre workloads, pro\u00advi\u00adding teams with the fle\u00adxi\u00adbi\u00adli\u00adty and capa\u00adbi\u00adli\u00adty to stay at the fore\u00adfront of innovation.\u201d<\/p>\n

\u201c<\/span>The NVIDIA<\/span> DGX<\/span> A100<\/span> deli\u00advers a tre\u00admen\u00addous leap in per\u00adfor\u00admance and capa\u00adbi\u00adli\u00adties,\u201d said Char\u00adlie Boyle, vice pre\u00adsi\u00addent and gene\u00adral mana\u00adger, DGX<\/span> sys\u00adtems at NVIDIA<\/span>. \u201cThe 2nd Gen AMD<\/span> EPYC<\/span> pro\u00adces\u00adsors used in DGX<\/span> A100<\/span> pro\u00advi\u00adde high per\u00adfor\u00admance and sup\u00adport for PCIe Gen4. NVIDIA<\/span> has put tho\u00adse fea\u00adtures to work to crea\u00adte the world\u2019s most powerful AI<\/span> sys\u00adtem while main\u00adtai\u00adning com\u00adpa\u00adti\u00adbi\u00adli\u00adty with the GPU-opti\u00admi\u00adzed soft\u00adware stack used across the enti\u00adre DGX<\/span> family.\u201d<\/p>\n

Sup\u00adport\u00ading Resources<\/strong><\/p>\n