From May 8-11, I’ll be spending my time at the GPU Technology Conference (GTC). GTC offers valuable training, talks, panels, and more. I’m excited to see the most vital work in the computing industry. I’m expecting to see information on artificial intelligence and deep learning, healthcare, virtual reality, accelerated analytics, and self-driving cars. I can’t wait to see all the remarkable breakthroughs that my colleagues will be talking about.
During the show, I’ll be presenting two different sessions.
S7506: Rolling in the deep: How to debug machine learning call stacks
Python is a popular language for deep learning but debugging calls to existing C/C++ code in shared libraries can be extremely challenging. Untangling the confusing maze of library calls, data translations and linked in CUDA code can be convoluted and time-consuming, as neither Python nor C/C++ debuggers provide a comprehensive view across the languages. In this talk, we’ll look at how Python-C/C++ transformations combined with a multi-threaded, multi-process debugger helps you understand what’s going on within your deep learning code.
S7507: Computer preemption and TotalView have made debugging Pascal much more seamless
With Pascal, NVIDIA released computer preemption built right into the card. Debugging now is much smoother because when we stop a thread on the GPU we no longer stop the whole GPU, enabling interactive debugging on single-GPU systems and debugging multiple processes using the same GPU. Having said that, TotalView, the leading multi-threaded Linux debugger, has invested into improving its architecture in order to support multi-GPU systems at scale resulting in a much more seamless debugging experience. Come listen to this discussion to get a better understanding of the latest technology and how and where we are looking to go next.
In addition to my individual sessions, I’ll also be participating in a roundtable about CUDA development tools.
Having attended and presented at GTC before I can say it’s a great conference.
At this year’s show, we’ll be announcing an update to our TotalView for HPC and CodeDynamics products, with new mixed-language debugging for Python and C/C++. Mixing languages has become an important process in developing applications that take advantage of deep learning.
My two individual sessions are closely tied to the improvements of our products, as well as what I’m seeing in the industry. In my S7506 session, I’ll address how we manage the call stacks and data in these mixed language environments. My S7507 session is focused on NVIDIA’s new Pascal chip architecture and the added computer preemption. This allows TotalView to work better than before. TotalView has a long history of supporting NVIDIA GPUs and this new functionality enhances the user experience.