— AMD unleas­hes the power of spe­cia­li­zed com­pu­te for the data cen­ter with new AMD EPYC pro­ces­sors for cloud nati­ve and tech­ni­cal computing —

—AMD reve­als details on next-gene­ra­ti­on AMD Instinct pro­ducts for gene­ra­ti­ve AI and high­lights AI soft­ware eco­sys­tem col­la­bo­ra­ti­ons with Hug­ging Face and PyTorch —

SANTA CLARA, Calif., June 13, 2023 (GLOBE NEWSWIRE) — Today, at the “Data Cen­ter and AI Tech­no­lo­gy Pre­mie­re,” AMD (NASDAQ: AMD) announ­ced the pro­ducts, stra­tegy and eco­sys­tem part­ners that will shape the future of com­pu­ting, high­light­ing the next pha­se of data cen­ter inno­va­ti­on. AMD was joi­n­ed on stage with exe­cu­ti­ves from Ama­zon Web Ser­vices (AWS), Cita­del, Hug­ging Face, Meta, Micro­soft Azu­re and PyTorch to show­ca­se the tech­no­lo­gi­cal part­ner­ships with indus­try lea­ders to bring the next gene­ra­ti­on of high per­for­mance CPU and AI acce­le­ra­tor solu­ti­ons to market.

“Today, we took ano­ther signi­fi­cant step for­ward in our data cen­ter stra­tegy as we expan­ded our 4^th Gen EPYC™ pro­ces­sor fami­ly with new lea­der­ship solu­ti­ons for cloud and tech­ni­cal com­pu­ting workloads and announ­ced new public ins­tances and inter­nal deploy­ments with the lar­gest cloud pro­vi­ders,” said AMD Chair and CEO Dr. Lisa Su. “AI is the defi­ning tech­no­lo­gy sha­ping the next gene­ra­ti­on of com­pu­ting and the lar­gest stra­te­gic growth oppor­tu­ni­ty for AMD. We are laser focu­sed on acce­le­ra­ting the deploy­ment of AMD AI plat­forms at sca­le in the data cen­ter, led by the launch of our Instinct MI300 acce­le­ra­tors plan­ned for later this year and the gro­wing eco­sys­tem of enter­pri­se-rea­dy AI soft­ware opti­mi­zed for our hardware.”

Com­pu­te Infra­struc­tu­re Opti­mi­zed for The Modern Data Center
AMD unvei­led a series of updates to its 4^th Gen EPYC fami­ly, desi­gned to offer cus­to­mers the workload spe­cia­liza­ti­on nee­ded to address busi­nesses’ uni­que needs.

• Advan­cing the World’s Best Data Cen­ter CPU. AMD high­ligh­ted how the 4^th Gen AMD EPYC pro­ces­sor con­ti­nues to dri­ve lea­der­ship per­for­mance and ener­gy effi­ci­en­cy. AMD was joi­n­ed by AWS to high­light a pre­view of the next gene­ra­ti­on Ama­zon Ela­s­tic Com­pu­te Cloud (Ama­zon EC2) M7a ins­tances, powered by 4th Gen AMD EPYC pro­ces­sors (“Gen­oa”). Out­side of the event, Ora­cle announ­ced plans to make available new Ora­cle Com­pu­ting Infra­struc­tu­re (OCI) E5 ins­tances with 4th Gen AMD EPYC processors.
• No Com­pro­mi­se Cloud Nati­ve Com­pu­ting. AMD intro­du­ced the 4^th Gen AMD EPYC 97X4 pro­ces­sors, form­er­ly code­na­med “Ber­ga­mo.” With 128 “Zen 4c” cores per socket, the­se pro­ces­sors pro­vi­de the grea­test vCPU den­si­ty¹ and indus­try lea­ding² per­for­mance for appli­ca­ti­ons that run in the cloud, and lea­der­ship ener­gy effi­ci­en­cy. AMD was joi­n­ed by Meta who dis­cus­sed how the­se pro­ces­sors are well sui­ted for their main­stay appli­ca­ti­ons such as Insta­gram, Whats­App and more; how Meta is see­ing impres­si­ve per­for­mance gains with 4^th Gen AMD EPYC 97x4 pro­ces­sors com­pared to 3^rd Gen AMD EPYC across various workloads, while offe­ring sub­stan­ti­al TCO impro­ve­ments as well, and how AMD and Meta opti­mi­zed the EPYC CPUs for Meta’s power-effi­ci­en­cy and com­pu­te-den­si­ty requirements.
• Enab­ling Bet­ter Pro­ducts With Tech­ni­cal Com­pu­ting. AMD intro­du­ced the 4th Gen AMD EPYC pro­ces­sors with AMD 3D V‑Cache™ tech­no­lo­gy, the world’s hig­hest per­for­mance x86 ser­ver CPU for tech­ni­cal com­pu­ting³. Micro­soft announ­ced the gene­ral avai­la­bi­li­ty of Azu­re HBv4 and HX ins­tances, powered by 4^th Gen AMD EPYC pro­ces­sors with AMD 3D V‑Cache technology.

Click here to learn more about the latest 4^th Gen AMD EPYC pro­ces­sors and read about what AMD cus­to­mers have to say, here.

AMD AI Plat­form – The Per­va­si­ve AI Vision
Today, AMD unvei­led a series of announce­ments show­ca­sing its AI Plat­form stra­tegy, giving cus­to­mers a cloud, to edge, to end­point port­fo­lio of hard­ware pro­ducts, with deep indus­try soft­ware col­la­bo­ra­ti­on, to deve­lop sca­lable and per­va­si­ve AI solutions.

• Intro­du­cing the World’s Most Advan­ced Acce­le­ra­tor for Gene­ra­ti­ve AI⁴. AMD reve­a­led new details of the AMD Instinct™ MI300 Series acce­le­ra­tor fami­ly, inclu­ding the intro­duc­tion of the AMD Instinct MI300X acce­le­ra­tor, the world’s most advan­ced acce­le­ra­tor for gene­ra­ti­ve AI. The MI300X is based on the next-gen AMD CDNA™ 3 acce­le­ra­tor archi­tec­tu­re and sup­ports up to 192 GB of HBM3 memo­ry to pro­vi­de the com­pu­te and memo­ry effi­ci­en­cy nee­ded for lar­ge lan­guage model trai­ning and infe­rence for gene­ra­ti­ve AI workloads. With the lar­ge memo­ry of AMD Instinct MI300X, cus­to­mers can now fit lar­ge lan­guage models such as Fal­con-40, a 40B para­me­ter model on a sin­gle, MI300X acce­le­ra­tor⁵. AMD also intro­du­ced the AMD Instinct™ Plat­form, which brings tog­e­ther eight MI300X acce­le­ra­tors into an indus­try-stan­dard design for the ulti­ma­te solu­ti­on for AI infe­rence and trai­ning. The MI300X is sam­pling to key cus­to­mers start­ing in Q3. AMD also announ­ced that the AMD Instinct MI300A, the world’s first APU Acce­le­ra­tor for HPC and AI workloads, is now sam­pling to customers.
• Brin­ging an Open, Pro­ven and Rea­dy AI Soft­ware Plat­form to Mar­ket. AMD show­ca­sed the ROCm™ soft­ware eco­sys­tem for data cen­ter acce­le­ra­tors, high­light­ing the rea­di­ness and col­la­bo­ra­ti­ons with indus­try lea­ders to bring tog­e­ther an open AI soft­ware eco­sys­tem. PyTorch dis­cus­sed the work bet­ween AMD and the PyTorch Foun­da­ti­on to ful­ly upstream the ROCm soft­ware stack, pro­vi­ding imme­dia­te “day zero” sup­port for PyTorch 2.0 with ROCm release 5.4.2 on all AMD Instinct acce­le­ra­tors. This inte­gra­ti­on empowers deve­lo­pers with an exten­si­ve array of AI models powered by PyTorch that are com­pa­ti­ble and rea­dy to use “out of the box” on AMD acce­le­ra­tors. Hug­ging Face, the lea­ding open plat­form for AI buil­ders, announ­ced that it will opti­mi­ze thou­sands of Hug­ging Face models on AMD plat­forms, from AMD Instinct acce­le­ra­tors to AMD Ryzen™ and AMD EPYC pro­ces­sors, AMD Rade­on™ GPUs and Ver­sal™ and Alveo™ adap­ti­ve processors.

A Robust Net­wor­king Port­fo­lio for the Cloud and Enterprise
AMD show­ca­sed a robust net­wor­king port­fo­lio inclu­ding the AMD Pen­san­do™ DPU, AMD Ultra Low Laten­cy NICs and AMD Adap­ti­ve NICs. Addi­tio­nal­ly, AMD Pen­san­do DPUs com­bi­ne a robust soft­ware stack with “zero trust secu­ri­ty” and lea­der­ship pro­gramma­ble packet pro­ces­sor to crea­te the world’s most intel­li­gent and per­for­mant DPU. The AMD Pen­san­do DPU is deploy­ed at sca­le across cloud part­ners such as IBM Cloud, Micro­soft Azu­re and Ora­cle Com­pu­te Infra­struc­tu­re. In the enter­pri­se it is deploy­ed in the HPE Aru­ba Net­wor­king CX 10000 series switch, and with cus­to­mers such as lea­ding IT ser­vices com­pa­ny DXC, and as part of VMware vSphe­re^® Dis­tri­bu­ted Ser­vices Engi­ne™, acce­le­ra­ting appli­ca­ti­on per­for­mance for customers.

AMD high­ligh­ted the next gene­ra­ti­on of its DPU road­map, code­na­med “Giglio,” which aims to bring enhan­ced per­for­mance and power effi­ci­en­cy to cus­to­mers, com­pared to cur­rent gene­ra­ti­on pro­ducts, when it’s expec­ted to be available by the end of 2023.

AMD also announ­ced the AMD Pen­san­do Soft­ware-in-Sili­con Deve­lo­per Kit (SSDK), giving cus­to­mers the abili­ty to rapidly deve­lop or migra­te ser­vices to deploy on the AMD Pen­san­do P4 pro­gramma­ble DPU in coor­di­na­ti­on with the exis­ting rich set of fea­tures alre­a­dy imple­men­ted on the AMD Pen­san­do plat­form. The AMD Pen­san­do SSDK enables cus­to­mers to put the power of the lea­der­ship AMD Pen­san­do DPU to work and tail­or net­work vir­tua­liza­ti­on and secu­ri­ty fea­tures within their infra­struc­tu­re, in coor­di­na­ti­on with the exis­ting rich set of fea­tures alre­a­dy imple­men­ted on the Pen­san­do platform.

Sup­port­ing Resources

• Watch the “Data Cen­ter and AI Tech­no­lo­gy Pre­mie­re” Key­note Video
• Read more about 4^th Gen AMD EPYC Processors
• Read more about AMD Instinct Accelerators
• Read more about AMD Net­wor­king Solutions
• Fol­low AMD on Twit­ter
• Con­nect with AMD on Lin­ke­dIn

About AMD
For more than 50 years AMD has dri­ven inno­va­ti­on in high-per­for­mance com­pu­ting, gra­phics and visua­liza­ti­on tech­no­lo­gies. Bil­li­ons of peo­p­le, lea­ding For­tu­ne 500 busi­nesses and cut­ting-edge sci­en­ti­fic rese­arch insti­tu­ti­ons around the world rely on AMD tech­no­lo­gy dai­ly to impro­ve how they live, work and play. AMD employees are focu­sed on buil­ding lea­der­ship high-per­for­mance and adap­ti­ve pro­ducts that push the boun­da­ries of what is pos­si­ble. For more infor­ma­ti­on about how AMD is enab­ling today and inspi­ring tomor­row, visit the AMD (NASDAQ: AMD) web­site, blog, Lin­ke­dIn and Twit­ter pages.

AMD, the AMD Arrow logo, EPYC, AMD Instinct, ROCm, Ryzen, Rade­on and com­bi­na­ti­ons the­reof are trade­marks of Advan­ced Micro Devices, Inc. Other names are for infor­ma­tio­nal pur­po­ses only and may be trade­marks of their respec­ti­ve owners.

CAUTIONARY STATEMENT

This press release con­ta­ins for­ward-loo­king state­ments con­cer­ning Advan­ced Micro Devices, Inc. (AMD) such as the fea­tures, func­tion­a­li­ty, per­for­mance, avai­la­bi­li­ty, timing and expec­ted bene­fits of AMD pro­ducts inclu­ding, the AMD 4^th Gen EPYC™ pro­ces­sor fami­ly, the AMD Instinct™ MI300 Series acce­le­ra­tor fami­ly, inclu­ding AMD Instinct™ MI300X and AMD Instinct™ MI300A, and the AMD Pen­san­do DPU code­na­med “Giglio”, which are made pur­su­ant to the Safe Har­bor pro­vi­si­ons of the Pri­va­te Secu­ri­ties Liti­ga­ti­on Reform Act of 1995. For­ward-loo­king state­ments are com­mon­ly iden­ti­fied by words such as “would,” “may,” “expects,” “belie­ves,” “plans,” “intends,” “pro­jects” and other terms with simi­lar mea­ning. Inves­tors are cau­tio­ned that the for­ward-loo­king state­ments in this press release are based on cur­rent beliefs, assump­ti­ons and expec­ta­ti­ons, speak only as of the date of this press release and invol­ve risks and uncer­tain­ties that could cau­se actu­al results to dif­fer mate­ri­al­ly from cur­rent expec­ta­ti­ons. Such state­ments are sub­ject to cer­tain known and unknown risks and uncer­tain­ties, many of which are dif­fi­cult to pre­dict and gene­ral­ly bey­ond AMD’s con­trol, that could cau­se actu­al results and other future events to dif­fer mate­ri­al­ly from tho­se expres­sed in, or impli­ed or pro­jec­ted by, the for­ward-loo­king infor­ma­ti­on and state­ments. Mate­ri­al fac­tors that could cau­se actu­al results to dif­fer mate­ri­al­ly from cur­rent expec­ta­ti­ons include, wit­hout limi­ta­ti­on, the fol­lo­wing: Intel Corporation’s domi­nan­ce of the micro­pro­ces­sor mar­ket and its aggres­si­ve busi­ness prac­ti­ces; glo­bal eco­no­mic uncer­tain­ty; cycli­cal natu­re of the semi­con­duc­tor indus­try; mar­ket con­di­ti­ons of the indus­tries in which AMD pro­ducts are sold; loss of a signi­fi­cant cus­to­mer; impact of the COVID-19 pan­de­mic on AMD’s busi­ness, finan­cial con­di­ti­on and results of ope­ra­ti­ons; com­pe­ti­ti­ve mar­kets in which AMD’s pro­ducts are sold; quar­ter­ly and sea­so­nal sales pat­terns; AMD’s abili­ty to ade­qua­te­ly pro­tect its tech­no­lo­gy or other intellec­tu­al pro­per­ty; unfa­vorable cur­ren­cy exch­an­ge rate fluc­tua­tions; abili­ty of third par­ty manu­fac­tu­r­ers to manu­fac­tu­re AMD’s pro­ducts on a time­ly basis in suf­fi­ci­ent quan­ti­ties and using com­pe­ti­ti­ve tech­no­lo­gies; avai­la­bi­li­ty of essen­ti­al equip­ment, mate­ri­als, sub­stra­tes or manu­fac­tu­ring pro­ces­ses; abili­ty to achie­ve expec­ted manu­fac­tu­ring yields for AMD’s pro­ducts; AMD’s abili­ty to intro­du­ce pro­ducts on a time­ly basis with expec­ted fea­tures and per­for­mance levels; AMD’s abili­ty to gene­ra­te reve­nue from its semi-cus­tom SoC pro­ducts; poten­ti­al secu­ri­ty vul­nerabi­li­ties; poten­ti­al secu­ri­ty inci­dents inclu­ding IT outa­ges, data loss, data brea­ches and cyber-attacks; poten­ti­al dif­fi­cul­ties in upgrading and ope­ra­ting AMD’s new enter­pri­se resour­ce plan­ning sys­tem; uncer­tain­ties invol­ving the orde­ring and ship­ment of AMD’s pro­ducts; AMD’s reli­ance on third-par­ty intellec­tu­al pro­per­ty to design and intro­du­ce new pro­ducts in a time­ly man­ner; AMD’s reli­ance on third-par­ty com­pa­nies for design, manu­fac­tu­re and sup­p­ly of mother­boards, soft­ware and other com­pu­ter plat­form com­pon­ents; AMD’s reli­ance on Micro­soft and other soft­ware ven­dors’ sup­port to design and deve­lop soft­ware to run on AMD’s pro­ducts; AMD’s reli­ance on third-par­ty dis­tri­bu­tors and add-in-board part­ners; impact of modi­fi­ca­ti­on or inter­rup­ti­on of AMD’s inter­nal busi­ness pro­ces­ses and infor­ma­ti­on sys­tems; com­pa­ti­bi­li­ty of AMD’s pro­ducts with some or all indus­try-stan­dard soft­ware and hard­ware; cos­ts rela­ted to defec­ti­ve pro­ducts; effi­ci­en­cy of AMD’s sup­p­ly chain; AMD’s abili­ty to rely on third par­ty sup­p­ly-chain logi­stics func­tions; AMD’s abili­ty to effec­tively con­trol sales of its pro­ducts on the gray mar­ket; impact of govern­ment actions and regu­la­ti­ons such as export admi­nis­tra­ti­on regu­la­ti­ons, tariffs and trade pro­tec­tion mea­su­res; AMD’s abili­ty to rea­li­ze its defer­red tax assets; poten­ti­al tax lia­bi­li­ties; cur­rent and future claims and liti­ga­ti­on; impact of envi­ron­men­tal laws, con­flict mine­rals-rela­ted pro­vi­si­ons and other laws or regu­la­ti­ons; impact of acqui­si­ti­ons, joint ven­tures and/or invest­ments on AMD’s busi­ness and AMD’s abili­ty to inte­gra­te acqui­red busi­nesses;  impact of any impair­ment of the com­bi­ned company’s assets on the com­bi­ned company’s finan­cial posi­ti­on and results of ope­ra­ti­on; rest­ric­tions impo­sed by agree­ments gover­ning AMD’s notes, the gua­ran­tees of Xilinx’s notes and the revol­ving cre­dit faci­li­ty; AMD’s indeb­ted­ness; AMD’s abili­ty to gene­ra­te suf­fi­ci­ent cash to meet its working capi­tal requi­re­ments or gene­ra­te suf­fi­ci­ent reve­nue and ope­ra­ting cash flow to make all of its plan­ned R&D or stra­te­gic invest­ments; poli­ti­cal, legal, eco­no­mic risks and natu­ral dis­as­ters; future impairm­ents of good­will and tech­no­lo­gy licen­se purcha­ses; AMD’s abili­ty to attract and retain qua­li­fied per­son­nel; AMD’s stock pri­ce vola­ti­li­ty; and world­wi­de poli­ti­cal con­di­ti­ons. Inves­tors are urged to review in detail the risks and uncer­tain­ties in AMD’s Secu­ri­ties and Exch­an­ge Com­mis­si­on filings, inclu­ding but not limi­t­ed to AMD’s most recent reports on Forms 10‑K and 10‑Q.

____________________
¹ EPYC-049: AMD EPYC 9754 is a 128 core dual threa­ded CPU and in a 2 socket ser­ver with 1 thread per vCPU deli­vers 512 vCPUs per EPYC powered ser­ver which is more than any Ampere or 4 socket Intel CPU based ser­ver as of 05/23/2023.
² SP5-143A: SPECrate®2017_int_base com­pa­ri­son based on per­forming sys­tem published scores from www.spec.org as of 6/13/2013. 2P AMD EPYC 9754 scores 1950 SPECrate®2017_int_base http://www.spec.org/cpu2017/results/res2023q2/cpu2017-20230522–36617.html is hig­her than all other 2P ser­vers. 1P AMD EPYC 9754 scores 981 SPECrate®2017_int_base score (981.4 score/socket) http://www.spec.org/cpu2017/results/res2023q2/cpu2017-20230522–36613.html is hig­her per socket than all other ser­vers. SPEC®, SPEC CPU®, and SPE­Cra­te® are regis­tered trade­marks of the Stan­dard Per­for­mance Eva­lua­ti­on Cor­po­ra­ti­on. See www.spec.org for more information.
³ SP5-165: The EPYC 9684X CPU is the world’s hig­hest per­for­mance x86 ser­ver CPU for tech­ni­cal com­pu­ting, com­pa­ri­son based on SPEC.org publi­ca­ti­ons as of 6/13/2023 mea­su­ring the score, rating or jobs/day for each of SPECrate®2017_fp_base (SP5-009E), Alta­ir Acu­Sol­ve (https://www.amd.com/en/processors/server-tech-docs/amd-epyc-9004x-pb-altair-acusolve.pdf), Ansys Flu­ent (https://www.amd.com/en/processors/server-tech-docs/amd-epyc-9004x-pb-ansys-fluent.pdf), Open­FOAM (https://www.amd.com/en/processors/server-tech-docs/amd-epyc-9004x-pb-openfoam.pdf), Ansys LS-Dyna (https://www.amd.com/en/processors/server-tech-docs/amd-epyc-9004x-pb-ansys-ls-dyna.pdf), and Alta­ir Radioss (https://www.amd.com/en/processors/server-tech-docs/amd-epyc-9004x-pb-altair-radioss.pdf) appli­ca­ti­on test case simu­la­ti­ons avera­ge spee­dup on 2P ser­vers run­ning 96-core EPYC 9684X vs top 2P per­for­mance gene­ral-pur­po­se 56-core Intel Xeon Pla­ti­num 8480+ or top-of-stack 60-core Xeon 8490H based ser­ver for tech­ni­cal com­pu­ting per­for­mance lea­der­ship. “Tech­ni­cal Com­pu­ting” or “Tech­ni­cal Com­pu­ting Workloads” as defi­ned by AMD can include: elec­tro­nic design auto­ma­ti­on, com­pu­ta­tio­nal flu­id dyna­mics, fini­te ele­ment ana­ly­sis, seis­mic tomo­gra­phy, wea­ther fore­cas­ting, quan­tum mecha­nics, cli­ma­te rese­arch, mole­cu­lar mode­ling, or simi­lar workloads. Results may vary based on fac­tors inclu­ding sili­con ver­si­on, hard­ware and soft­ware con­fi­gu­ra­ti­on and dri­ver ver­si­ons. SPEC®, SPE­Cra­te® and SPEC CPU® are regis­tered trade­marks of the Stan­dard Per­for­mance Eva­lua­ti­on Cor­po­ra­ti­on. See www.spec.org for more information.
⁴ MI300-09 — The AMD Instinct™ MI300X acce­le­ra­tor is based on AMD CDNA™ 3 5nm Fin­Fet pro­cess tech­no­lo­gy with 3D chip­let stack­ing, uti­li­zes high speed AMD Infi­ni­ty Fabric tech­no­lo­gy, has 192 GB HBM3 memo­ry capa­ci­ty (vs. 80GB for Nvi­dia Hop­per H100) with 5.218 TFLOPS of sus­tained peak memo­ry band­width per­for­mance, hig­her than the hig­hest band­width Nvi­dia Hop­per H100 GPU.
⁵ MI300-07K: Mea­su­re­ments by inter­nal AMD Per­for­mance Labs as of June 2, 2023 on cur­rent spe­ci­fi­ca­ti­ons and/or inter­nal engi­nee­ring cal­cu­la­ti­ons. Lar­ge Lan­guage Model (LLM) run or cal­cu­la­ted with FP16 pre­cis­i­on to deter­mi­ne the mini­mum num­ber of GPUs nee­ded to run the Fal­con (40B para­me­ter) models. Tes­ted result con­fi­gu­ra­ti­ons: AMD Lab sys­tem con­sis­ting of 1x EPYC 9654 (96-core) CPU with 1x AMD Instinct™ MI300X (192GB HBM3, OAM Modu­le) 750W acce­le­ra­tor tes­ted at FP16 pre­cis­i­on. Ser­ver manu­fac­tu­r­ers may vary con­fi­gu­ra­ti­on offe­rings yiel­ding dif­fe­rent results.