University of Florida, NVIDIA to Build Fastest AI Supercomputer in Academia

Effort will infu­se AI throughout UF’s cur­ri­cu­lum and help address wide ran­ge of chal­len­ges across the state.
 
 
 
 
 

The Uni­ver­si­ty of Flo­ri­da and NVIDIA Tues­day unvei­led a plan to build the world’s fas­test AI super­com­pu­ter in aca­de­mia, deli­vering 700 peta­flops of AI performance.

The effort is ancho­red by a $50 mil­li­on gift: $25 mil­li­on from alum­nus and NVIDIA co-foun­der Chris Mal­a­chow­sky and $25 mil­li­on in hard­ware, soft­ware, trai­ning and ser­vices from NVIDIA.

We’ve crea­ted a repli­ca­ble, power­ful model of public-pri­va­te coope­ra­ti­on for everyone’s bene­fit,” said Mal­a­chow­sky, who ser­ves as an NVIDIA Fel­low, in an online event fea­turing lea­ders from both the UF and NVIDIA.

UF will invest an addi­tio­nal $20 mil­li­on to crea­te an AI-centric super­com­pu­ting and data center.

The $70 mil­li­on public-pri­va­te part­ners­hip pro­mi­ses to make UF one of the lea­ding AI uni­ver­si­ties in the coun­try, advan­ce aca­de­mic rese­arch and help address some of the state’s most com­plex challenges.

This is going to be a tre­men­dous part­ners­hip,” Flo­ri­da Gov. Ron DeS­an­tis said. “As we look to keep our best talent  in sta­te, this will be a signi­fi­cant car­rot, you’ll also see peop­le around the coun­try want to come to Florida.”

  

The Uni­ver­si­ty of Flo­ri­da and NVIDIA Tues­day unvei­led a plan to build the world’s fas­test AI super­com­pu­ter in aca­de­mia, deli­vering 700 peta­flops of AI performance.

The effort is ancho­red by a $50 mil­li­on gift: $25 mil­li­on from alum­nus and NVIDIA co-foun­der Chris Mal­a­chow­sky and $25 mil­li­on in hard­ware, soft­ware, trai­ning and ser­vices from NVIDIA.

We’ve crea­ted a repli­ca­ble, power­ful model of public-pri­va­te coope­ra­ti­on for everyone’s bene­fit,” said Mal­a­chow­sky, who ser­ves as an NVIDIA Fel­low, in an online event fea­turing lea­ders from both the UF and NVIDIA.

UF will invest an addi­tio­nal $20 mil­li­on to crea­te an AI-centric super­com­pu­ting and data center.

The $70 mil­li­on public-pri­va­te part­ners­hip pro­mi­ses to make UF one of the lea­ding AI uni­ver­si­ties in the coun­try, advan­ce aca­de­mic rese­arch and help address some of the state’s most com­plex challenges.

This is going to be a tre­men­dous part­ners­hip,” Flo­ri­da Gov. Ron DeS­an­tis said. “As we look to keep our best talent  in sta­te, this will be a signi­fi­cant car­rot, you’ll also see peop­le around the coun­try want to come to Florida.”

 

Working clo­se­ly with NVIDIA, UF will boost the capa­bi­li­ties of its exis­ting super­com­pu­ter, HiP­er­Ga­tor, with the recent­ly announ­ced NVIDIA DGX Super­POD archi­tec­tu­re. The sys­tem will be up and run­ning by ear­ly 2021, just a few weeks after it’s delivered.

This gives facul­ty and stu­dents wit­hin and bey­ond UF the tools to app­ly AI across a mul­ti­tu­de of are­as to address major chal­len­ges such as rising seas, aging popu­la­ti­ons, data secu­ri­ty, per­so­na­li­zed medi­ci­ne, urban trans­por­ta­ti­on and food inse­cu­ri­ty. UF expects to crea­te 30,000 AI-enab­led gra­dua­tes by 2030.

The part­ners­hip here with the UF, the sta­te of Flo­ri­da, and NVIDIA, ancho­red by Chris’ generous dona­ti­on, goes bey­ond just money,” said NVIDIA CEO Jen­sen Huang, who foun­ded NVIDIA in 1993 along with Mal­a­chow­sky and Cur­tis Priem. “We are exci­ted to con­tri­bu­te NVIDIA’s exper­ti­se to work tog­e­ther to make UF a natio­nal lea­der in AI and help address not only the region’s, but the nation’s challenges.”

UFran­ked seventh among public uni­ver­si­ties in the United Sta­tes by US News & World Report and aims to break into the top five, offers an extra­or­di­na­ri­ly broad ran­ge of disci­pli­nes, Mal­a­chow­sky said.

The regi­on is also a “living labo­ra­to­ry for some of society’s big­gest chal­len­ges,” Mal­a­chow­sky said.

Regio­nal, Natio­nal AI Leadership

The effort aims to help defi­ne a rese­arch land­s­cape to deal with the COVID-19, pan­de­mic, which has seen super­com­pu­ters take a lea­ding role.

Our visi­on is to beco­me the nation’s first AI uni­ver­si­ty,” Uni­ver­si­ty of Flo­ri­da Pre­si­dent Kent Fuchs said. “I am so gra­te­ful again to Mr. Mal­a­chow­sky and NVIDIA CEO Jen­sen Huang.

Sta­te and regio­nal lea­ders alrea­dy look to the uni­ver­si­ty to bring its capa­bi­li­ties to bear on an array of regio­nal and natio­nal issues.

Among them: sup­por­ting agri­cul­tu­re in a time of cli­ma­te chan­ge, addres­sing the needs of an aging popu­la­ti­on, and mana­ging the effects of rising sea levels in a sta­te with more than 1,300 miles of coastline.

And to ensu­re no com­mu­ni­ty is left behind, UF plans to pro­mo­te wide acces­si­bi­li­ty to the­se com­pu­ting capabilities.

As part of this, UF will:

  • Estab­lish UF’s Equi­ta­ble AI pro­gram, to bring facul­ty mem­bers across the uni­ver­si­ty tog­e­ther to crea­te stan­dards and cer­ti­fi­ca­ti­ons for deve­lo­ping tools and solu­ti­ons that are cognizant of bias, unethi­cal prac­ti­ce and legal and moral issues.
  • Part­ner with indus­try and other aca­de­mic groups, such as the Inclu­si­ve Engi­nee­ring Con­sor­ti­um, who­se stu­dents will work with mem­bers to con­duct rese­arch and recruit­ment to UF gra­dua­te programs.

Broad Ran­ge of AI Initiatives

Mal­a­chow­sky has ser­ved in a num­ber of lea­ders­hip roles as NVIDIA has grown from a start­up to the glo­bal lea­der in visu­al and par­al­lel com­pu­ting. A reco­gni­zed aut­ho­ri­ty on inte­gra­ted-cir­cuit design and metho­do­lo­gy, he has autho­red clo­se to 40 patents.

In addi­ti­on to hol­ding a BSEE from the Uni­ver­si­ty of Flo­ri­da, he has an MSCS from San­ta Cla­ra Uni­ver­si­ty. He has been named a dis­tin­guis­hed alum­ni of both uni­ver­si­ties, in addi­ti­on to being induc­ted last year into the Flo­ri­da Inven­tors Hall of Fame.

UF is the first insti­tu­ti­on of hig­her lear­ning in the U.S. to recei­ve NVIDIA DGX A100 sys­tems. The­se sys­tems are based on the modu­lar archi­tec­tu­re of the NVIDIA DGX Super­POD, which enab­les the rapid deploy­ment and sca­ling of mas­si­ve AI infrastructure.

UF’s HiP­er­Ga­tor 3 super­com­pu­ter will inte­gra­te 140 NVIDIA DGX A100 sys­tems powe­red by a com­bi­ned 1,120 NVIDIA A100 Ten­sor Core GPUs. It will inclu­de 4 peta­bytes of high-per­for­mance sto­rage. An NVIDIA Mel­lanox HDR 200Gb/s Infi­ni­Band net­work will pro­vi­de the high through­put and extre­me­ly low-laten­cy net­work connectivity.

DGX A100 sys­tems are built to make the most of the­se capa­bi­li­ties as a sin­gle soft­ware-defi­ned plat­form. NVIDIA DGX sys­tems are alrea­dy used by eight of the ten top US natio­nal universities.

That plat­form inclu­des the most advan­ced suite of AI app­li­ca­ti­on frame­works in the world. It’s a soft­ware suite that covers data ana­ly­tics, AI trai­ning and infe­rence acce­le­ra­ti­on, and recom­men­da­ti­on sys­tems. Its mul­ti-modal capa­bi­li­ties com­bi­ne sound, visi­on, speech and a con­tex­tu­al under­stan­ding of the world around us.

Tog­e­ther, the­se tools have alrea­dy had a signi­fi­cant impact on health­ca­re, trans­por­ta­ti­on, sci­ence, inter­ac­ti­ve app­li­an­ces, the inter­net and other areas.

More Than Just a Machine

Friday’s announ­ce­ment, howe­ver, goes bey­ond any sin­gle, if sin­gu­lar, machine.

NVIDIA will also con­tri­bu­te its AI exper­ti­se to UF through ongo­ing sup­port and col­la­bo­ra­ti­on across the fol­lowing initiatives:

  • The NVIDIA Deep Lear­ning Insti­tu­te will col­la­bo­ra­te with UF on deve­lo­ping new cur­ri­cu­lum and cour­se­work for both stu­dents and the com­mu­ni­ty, inclu­ding pro­gra­ming tun­ed to address the needs of young adults and teens to encou­ra­ge their inte­rest in STEM and AI, bet­ter pre­pa­ring them for future edu­ca­tio­nal and employ­ment opportunities.
  • UF will beco­me the site of the latest NVIDIA AI Tech­no­lo­gy Cen­ter, whe­re UF Gra­dua­te Fel­lows and NVIDIA employees will work tog­e­ther to advan­ce AI.
  • NVIDIA solu­ti­on archi­tects and pro­duct engi­neers will part­ner with UF on the instal­la­ti­on, ope­ra­ti­on and opti­miz­a­ti­on of the NVI­DIA-based super­com­pu­ting resour­ces on cam­pus, inclu­ding the latest AI soft­ware applications.

UF will also make invest­ments all around its new machi­ne, well bey­ond the $20 mil­li­on tar­ge­ted at upgrading their data center.

Collec­tively, all of the data sci­en­ces-rela­ted acti­vi­ties and pro­grams — and UF’s new super­com­pu­ter — will sup­port the university’s broa­der AI-rela­ted aspirations.

To sup­port that effort, the uni­ver­si­ty has com­mit­ted to fill 100 new facul­ty posi­ti­ons in AI and rela­ted fiel­ds, making it one of the top AI uni­ver­si­ties in the country.

That’s in addi­ti­on to the 500 recent­ly hired facul­ty across disci­pli­nes, many of whom will wea­ve AI into their tea­ching and research.

It’s been thril­ling to watch all this,” Mal­a­chow­sky said. “It pro­vi­des a blue­print for how other sta­tes can work with their region’s resour­ces to make simi­lar invest­ments that bring their resi­dents the bene­fits of AI, while bols­te­ring our nation’s com­pe­ti­ti­ve­ness, capa­bi­li­ties, and expertise.”

Durch die weitere Nutzung der Seite stimmst du der Verwendung von Cookies zu. Weitere Informationen

Die Cookie-Einstellungen auf dieser Website sind auf "Cookies zulassen" eingestellt, um das beste Surferlebnis zu ermöglichen. Wenn du diese Website ohne Änderung der Cookie-Einstellungen verwendest oder auf "Akzeptieren" klickst, erklärst du dich damit einverstanden.

Schließen