Transcript
1:52 ladies and Gentlemen please welcome Nvidia co-founder president and CEO Jensen Wong good morning welcome to GTC 2012 this is a conference dedicated to the advancement of GPU technology and this is a conference dedicated to the science that you do on gpus you are the Stars this is a
2:25 celebration of your science of your work and your Innovations today I want to do three things I want to give you an update on the progress that we've made with GPU Computing and Cuda over the last several years since the last time we've met the second thing I want to do is to reveal for the first time some technologies that we've been working on back at the labs and the third thing I want to do is
2:55 to introduce you to some Partners some companies we've been working with that you wouldn't have regularly seen us with we got a big day for you a great show so let's get started Cuda has made enormous progress over the last couple years in fact if you looked at when we first started 2008 was about two years into the introduction of Cuda we had about 150,000 downloads already we had one whopping
3:26 supercomputer 60 universities were already starting to teach Cuda in their coursework and there were about 4,000 papers if you searched that on Google Scholar just a few years later just a few years later there's a million and a half downloads of Cuda we believe a mil one download of Cuda per second about 150,000 active developers of Cuda around the
3:57 world from one super computer to now 35 the top supercomputers in China Japan Russia and this year hopefully the top supercomputer in the world the top supercomputer here in in America from 60 universities we're now all the way up to over 500 48 countries every continent in Chinese and Japanese and Russian the Cuda coursework has been
4:29 translated to universities and students all over the world 4,000 papers now has nearly five-fold to 22,500 the number of papers is really a reflection of the Innovation that people are doing around Cuda the the paper is a reflection of the type of in Innovation and the power of GPU Computing how it's able to advance your research one way to look at
4:59 that is to compare it to other technologies that are very important in the industry Hadoop a distributed file system a powerful high performance parallel data processing algorithm called map map reduce openmp and API directives extensions to Fortran C C++ making that that language capable of understanding shared memory highly parallel systems both of
5:30 those standards are very important to the industry and notice how they've grown over the years here's Cuda it really reflects the power that GPU accelerate Computing has on your work it also reflects the accessibility of Cuda one of the best things that we've ever done was made sure that Cuda was available not only on Tesla but available on every single GPU that we created as a result of that that 100% of
6:01 every computer company in the world offer Cuda today it's available in every single country on the planet it's available in workstations desktop computers and laptops I don't actually know of one country or one computer company where you couldn't buy a computer that has Cuda inside you can literally access Cuda wherever you are whoever you are we've democratized high high performance Computing as we know it now the response from all of you have been fantastic out of
6:33 150,000 3,000 in room today the F the feedback from all of you and the encouragement has been completely fantastic I'm going to embarrass a few people today and read out the read their quotes and read the feedback they've given us out loud laye cousin GPU Computing has utterly transformed the science we can do Gary junan from the University autom Barcelona when I first programmed in Cuda I was so excited to see the progress of the algorithm literally
7:02 within seconds I was unable to sleep well the reason for that was he was expecting to be up all night anyways Stanley seabert I want to say thanks to the engineers at Nvidia Cuda has saved my thesis I want to thank the engineers at Nvidia as well the work that they have done as a result of Cuda has been really really fabulous and it's very rewarding to all of us in the final analysis the the ability to make an
7:30 impact in your work is what drives us the adoption of Cuda is also industrywide in 2007 we went to supercomputing for the very first time that little Green Dot well this is this is the show floor the exhibit floor of supercomputing 2007 it's was hosted in Reno and that little Green Dot is the Nvidia booth and that little Green Dot is the only
8:01 place that you could find Cuda and that was the only place you could find a GPU well the reports that I received from the team that went to supercomputing was glowing they thought we were everywhere they told me the adoption of Cuda was complete if not if I hadn't seen this if not because I didn't see this image I wouldn't have probably
8:33 enthusiastically continued to invest in Cuda thank goodness their reports were glowing we double down and only a few years later this is what supercomputing looked like every single computer company just about every software company Tool Company high performance computing speci specialist vars included Tesla or did work around
9:05 Cuda that year it is very very clear that the democratization of high performance Computing is happening right in front of us now it's very clear the result of the pervasiveness of Cuda and the power that it's able to provide to your work but ultimately it's about the science it's beyond papers it's beyond oems it's beyond industry it's really about your work and that's what GTC is
9:36 about it's about your work and these are some of the things that we're going to be talking about during this conference it's really amazing adaptive radio therapy the ability for CT scans to be interpreted right away so that they could recompute how radiation therapy ought to be directed into the living body dendrite simulations understanding the fact that when metal cools the rate at which it
10:06 cools and the manner at which it cools has a lot to do with its rigidity and its robustness understanding the collect behavior of swarming living things combustion efficiency turbulent flow simulations understanding how to sim how to design engines that are more efficient one of my favorites air traffic modeling there are planes flying all over the country if there was a disaster of some
10:37 kind a weather change in some way the ability to redirect all of those planes instantaneously would be wonderful to Route them to a better path today we utilize literally the heroics of a lot of people in Towers to try to figure out exactly what to do in the event of a disaster seismic to understanding the nature of Earth and understanding how earthquakes could affect ourselves there's so much going
11:09 on here there are so much going on here from image processing the Machine Vision Medical Imaging molecular Dynamics the list just goes on and on even computer Graphics even computer Graphics well that is all just what we've achieved that's our achievement so far and that's nothing compared to the work that we're going to do going forward we've decided to double down yet again today we're going to talk to you
11:39 about some of our most important inventions some of our greatest efforts and in my opinion some of our finest work