{"id":3193,"date":"2010-09-23T19:15:02","date_gmt":"2010-09-23T18:15:02","guid":{"rendered":"http:\/\/www.itwriting.com\/blog\/3193-nvidia-ceo-on-the-spot-explains-fermi-delays-cuda-vs-opencl-rise-of-the-tablet.html"},"modified":"2010-09-23T19:15:02","modified_gmt":"2010-09-23T18:15:02","slug":"nvidia-ceo-on-the-spot-explains-fermi-delays-cuda-vs-opencl-rise-of-the-tablet","status":"publish","type":"post","link":"https:\/\/www.itwriting.com\/blog\/3193-nvidia-ceo-on-the-spot-explains-fermi-delays-cuda-vs-opencl-rise-of-the-tablet.html","title":{"rendered":"NVIDIA CEO on the spot: explains Fermi delays, CUDA vs OpenCL, rise of the tablet"},"content":{"rendered":"

NVIDIA CEO Jen-Hsung Huang spoke to the press at the GPU Technology Conference<\/a> and I took the opportunity to ask some questions.<\/p>\n

\"image\"<\/a><\/p>\n

I asked for his views on the cloud as a supercomputer and whether that would impact the need for local supercomputers of the kind GPU computing enables.<\/p>\n

Although we expect more and more to happen in the cloud, in the meantime we\u2019re going to keep buying devices with more and more solid state memory. The way to think about it is, storage is simply a surrogate for bandwidth. If we had infinite bandwidth none of us would need storage. As bandwidth improves the requirement for storage should reduce. But there\u2019s another trend which is that the amount of data we collect is growing incredibly fast \u2026 It\u2019s going to be quite a long time before our need for storage will reduce.<\/p><\/blockquote>\n

But what about local computing power, Gigaflops as opposed to storage?<\/p>\n

Wherever there is storage, there\u2019s GigaFlops. Local storage, local computing.<\/p><\/blockquote>\n

Next, I brought up a subject which has been puzzling me here at GTC. You can do GPU programming with NVIDIA\u2019s CUDA C, which only works on NVIDIA GPUs, or with OpenCL which works with other vendor\u2019s GPUs as well. Why is there more focus here on CUDA, when on the face of it developers would be better off with the cross-GPU approach? (Of course I know part of the answer, that NVIDIA does not mind locking developers to its own products).<\/p>\n

The reason we focus all our evangelism and energy on CUDA is because CUDA requires us to, OpenCL does not. OpenCL has the benefit of IBM, AMD, Intel, and ourselves. Now CUDA is a little difference in that its programming approach is different. Instead of an API it\u2019s a language extension. You program in C, it\u2019s a different model.<\/p>\n

The reason why CUDA is more adopted than OpenCL is because it is simply more advanced. We\u2019ve invested in CUDA much longer. The quality of the compiler is much better. The robustness of the programming environment is better. The tools around it are better, and there are more people programming it. The ecosystem is richer.<\/p>\n

People ask me how do we feel about the fact that it is proprietary. There\u2019s two ways to think about it. There\u2019s CUDA and there\u2019s Tesla. Tesla\u2019s not proprietary at all, Tesla supports OpenCL and CUDA. If you bought a server with Tesla in it, you\u2019re not getting anything less, you\u2019re getting CUDA more. That\u2019s the reason Tesla has been adopted by all the OEMs. If you want a GPU cluster, would you want one that only does OpenCL? Or does OpenCL and CUDA? 80% of GPU computing today is CUDA, 20% is OpenCL. If you want to reach 100% of it, you\u2019re better off using Tesla. Over time, if more people use OpenCL that\u2019s fine with us. The most important thing is GPU computing, the next most important thing to us is NVIDIA\u2019s GPUs, and the next is CUDA. It\u2019s way down the list.<\/p><\/blockquote>\n

Next, a hot topic. Jen-Hsun Huang explained why he announced a roadmap for future graphics chip architectures – Kepler in 2011, Maxwell in 2013 – so that software developers engaged in GPU programming can plan their projects. I asked him why Fermi, the current chip architecture, had been so delayed, and whether there was good reason to have confidence in the newly announced dates.<\/p>\n

He answered by explaining the Fermi delay in both technical and management terms.<\/p>\n

The technical answer is that there\u2019s a piece of functionality that is between the shared symmetric multiprocessors (SMs), 236 processors, that need to communicate with each other, and with memory of all different types. So there\u2019s SMs up here, and underneath the memories. In between there is a very complicated inter-connecting system that is very fast. It\u2019s nearly all wires, dense metal with very little logic \u2026 we call that the fabric.<\/p>\n

When you have wires that are next to each other that closely they couple, they interfere \u2026 it\u2019s a solid mesh of metal. We found a major breakdown between the models, the tools, and reality. We got the first Fermi back. That piece of fabric \u2013 imagine we are all processors. All of us seem to be working. But we can\u2019t talk to each other. We found out it\u2019s because the connection between us is completely broken. We re-engineered the whole thing and made it work.<\/p>\n

Your question was deeper than that. Your question wasn\u2019t just what broke with Fermi \u2013 it was the fabric \u2013 but the question is how would you not let it happen again? It won\u2019t be fabric next time, it will be something else.<\/p>\n

The reason why the fabric failed isn\u2019t because it was hard, but because it sat between the responsibility of two groups. The fabric is complicated because there\u2019s an architectural component, a logic design component, and there\u2019s a physics component. My engineers who know physics and my engineers who know architecture are in two different organisations. We let it sit right in the middle. So the management lesson learned \u2013 there should always be a pilot in charge.<\/p><\/blockquote>\n

Huang spent some time discussing changes in the industry. He identifies mobile computing \u201csuperphones\u201d and tablets as the focus of a major shift happening now. Someone asked \u201cWhat does that mean for your Geforce business?\u201d<\/p>\n

I don\u2019t think like that. The way I think is, \u201cwhat is my personal computer business\u201d. The personal computer business is Geforce plus Tegra. If you start a business, don\u2019t think about the product you make. Think about the customer you\u2019re making it for. I want to give them the best possible personal computing experience.<\/p><\/blockquote>\n

Tegra is NVIDIA\u2019s complete system on a chip, including ARM processor and of course NVIDIA graphics, aimed at mobile devices. NVIDIA\u2019s challenge is that its success with Geforce does not guarantee success with Tegra, for which it is early days.<\/p>\n

The further implication is that the immediate future may not be easy, as traditional PC and laptop sales decline.<\/p>\n

The mainstream business for the personal computer industry will be rocky for some time. The reason is not because of the economy but because of mobile computing. The PC \u2026 will be under disruption from tablets. The difference between a tablet and a PC is going to become very small. Over the next few years we\u2019re going to see that more and more people use their mobile device as their primary computer.<\/p>\n

[Holds up Blackberry] There\u2019s no question right now that this is my primary computer.<\/p><\/blockquote>\n

The rise of mobile devices is a topic Huang has returned to on several occasions here. \u201cARM is the most important CPU architecture, instruction set architecture, of the future\u201d he told the keynote audience.<\/p>\n

Clearly NVIDIA\u2019s business plans are not without risk; but you cannot fault Huang for enthusiasm or awareness of coming changes. It is clear to me that NVIDIA has the attention of the scientific and academic community for GPU computing, and workstation OEMs are scrambling to built Tesla GPU computing cards into their systems, but transitions in the market for its mass-market graphics cards will be tricky for the company.<\/p>\n

Update<\/strong>: Huang\u2019s comments about the reasons for Fermi\u2019s delay raised considerable interest as apparently he had not spoken about this on record before. Journalist Nico Ernst captured the moment on video:<\/p>\n

<\/embed><\/object><\/p>\n
Video: Nvidia-Chef Jen-Hsun Huang \u00fcber Fermis technische Schwierigkeiten<\/a> (4:32)<\/div>\n","protected":false},"excerpt":{"rendered":"

NVIDIA CEO Jen-Hsung Huang spoke to the press at the GPU Technology Conference and I took the opportunity to ask some questions. I asked for his views on the cloud as a supercomputer and whether that would impact the need for local supercomputers of the kind GPU computing enables. Although we expect more and more … Continue reading NVIDIA CEO on the spot: explains Fermi delays, CUDA vs OpenCL, rise of the tablet<\/span> →<\/span><\/a><\/p>\n","protected":false},"author":1,"featured_media":0,"comment_status":"open","ping_status":"closed","sticky":false,"template":"","format":"standard","meta":{"footnotes":""},"categories":[26,56,58,67,80],"tags":[169,383,436,599,617,654,679,902],"class_list":["post-3193","post","type-post","status-publish","format-standard","hentry","category-development","category-mobile","category-multimedia","category-professional","category-software-development","tag-arm","tag-fermi","tag-gpu-computing","tag-mobile","tag-multimedia","tag-nvidia","tag-opencl","tag-tesla"],"_links":{"self":[{"href":"https:\/\/www.itwriting.com\/blog\/wp-json\/wp\/v2\/posts\/3193","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/www.itwriting.com\/blog\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/www.itwriting.com\/blog\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/www.itwriting.com\/blog\/wp-json\/wp\/v2\/users\/1"}],"replies":[{"embeddable":true,"href":"https:\/\/www.itwriting.com\/blog\/wp-json\/wp\/v2\/comments?post=3193"}],"version-history":[{"count":0,"href":"https:\/\/www.itwriting.com\/blog\/wp-json\/wp\/v2\/posts\/3193\/revisions"}],"wp:attachment":[{"href":"https:\/\/www.itwriting.com\/blog\/wp-json\/wp\/v2\/media?parent=3193"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/www.itwriting.com\/blog\/wp-json\/wp\/v2\/categories?post=3193"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/www.itwriting.com\/blog\/wp-json\/wp\/v2\/tags?post=3193"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}