OpenCL Training - Parallel Programming & Heterogeneous Computing

YetiWare's mission is to solve the big hairy parallel programming problem, and to bring back software scaling with minimal programmer effort. Raw processing power in everything from mobile phones, to PCs, to supercomputers, continuously grows every year with increasingly more cores, wider SIMD lanes, and additional software accelerators such as GPUs and FPGAs. Today it is extremely complicated to write applications that scale so that consumers directly benefit by upgrading to the latest computer processors. YetiWare's mission is to provide that revolutionary solution to performance programming so that application developers can build faster and more scalable software.

YetiWare Inc.  |   www.yetiware.com   |  @yetiware

YetiWare announces its OpenCL and parallel computing training to take place July 26th - 28th, 2016 in Sunnyvale/Santa Clara, California, USA.
We will hold the course over three (3) days at a downtown Sunnyvale venue.  All training at this session will be done by AJ Guillon, a subject matter expert and contributor to the Khronos OpenCL specification.  This training is geared toward professional developers who want to learn from another professional, without a lot of review of basics or time spent on trivial concepts.
An optional (extra fee) <A HREF="https://www.eventbrite.ca/e/opencl-training-review-day-tickets-25959863651" TARGET="_blank">Review Day</A> is available in advance of this training to prepare you.
<H2>What Is Parallel Programming and OpenCL?</H2>
Intel has a great video that explains why OpenCL is important to you, <A HREF="https://software.intel.com/en-us/videos/opencl-and-intel-why-it-matters-to-you" TARGET="_self" DATA-MSYS-CLICKTRACK="0" REL="nofollow noopener noreferrer">watch it here</A>.  Quite simply, OpenCL is the future of parallel data processing, and is supported by all the major processor companies.  AJ has some OpenCL YouTube training videos that you can <A HREF="http://goo.gl/6HY0bS" TARGET="_blank" DATA-MSYS-CLICKTRACK="0" REL="nofollow noopener noreferrer">watch here</A>.
Computer processor design has changed drastically in the past 20 years, yet most developers continue to write software according to a simplified model that results in poor application performance and disappointing user experiences.  Today, all major computer processors are parallel, and these parallel processors are in everything from smart phones, to desktop PCs, to supercomputers.  Despite the ubiquity of parallel processors, few software developers know how to program them.  Worse, prior to OpenCL, each processor company provided its own preferred parallel programming technology, which fragmented the industry.
The Khronos OpenCL standard was developed by major software and processor companies (including Intel, AMD, Apple, Adobe, NVIDIA, and many others) to provide a common way to write parallel software that will work on any processor, including a CPU, GPU, or FPGA. OpenCL includes its own parallel language, OpenCL C, that is designed to smoothly integrate with existing applications written in any language that can bind to C.  Today, OpenCL is leveraged by major companies to reduce operational costs and obtain peak software performance.
Despite the importance of parallel programming and OpenCL there are few educational resources on the subject.  YetiWare is pleased to offer a combined course on OpenCL and parallel programming so that developers can learn about these important advancements from an expert.  Our training targets professional software developers that want to learn new techniques from another software professional.
<H2>Testimonials</H2>
"AJ's course on OpenCL is a cut above other courses in that he brings with him tremendous background and experience, and delivers it in an easily understandable way that is reachable for novice, but still remains interesting for experts. This is more than just a one-way presentation, but a true tutorial with labs and hands-on components that truly improves your skill in this domain using skills that are now highly in demand across many industries, making this course pay for itself. He also makes it fun with anecdotes and entertaining stories."
Michael Wong. Former CEO OpenMP. ISOCPP.org Director, VP, VP of Research &amp; Development Codeplay Software 
 
"SciNet is Canada’s largest supercomputer centre, providing Canadian researchers with computational resources and expertise necessary to perform their research on scales not previously possible in Canada. SciNet powers work from the biomedical sciences and aerospace engineering to astrophysics and climate science.
 In today’s heterogeneous architecture world, AJ’s training in OpenCL is extremely valuable, for OpenCL allows one to extract performance from the heterogeneous systems, while preserving  code portability across architectures, thus helping future-proof application codes.”
Dr. Daniel Gruner, CTO (Software), SciNet High Performance Computing Consortium, Compute/Calcul Canada 
 
 
"The traditional thinking has been that parallel programming is difficult to learn. But AJ and his team have put together clear and effective course material that takes participants from learning fundamentals to delivering real life, relevant applications quickly.  As a result, participant are able to reap the benefits from the course quickly. The use of moderated lab sessions and opportunity to expand on existing applications makes it much more conducive for participants to make effective use of OpenCL in short order."
I-Cheng Chen, P.Eng., Fellow, Platform Architecture, Advanced Micro Devices Inc.
 
 
"My research deals with building infrastructure and approaches to making heterogeneous computing easier, especially with respect to integrating FPGAs into computing systems.  I was happy to meet AJ and find out that we have very similar visions, even though we come from the opposite ends of the stack -- I come from the hardware and architecture end and AJ from the CS end.  AJ is clearly an expert in OpenCL and more generally, the issues of heterogeneous computing from the programmer's perspective, which helps me understand what's needed in the hardware to make the programming easier." 
Professor Paul Chow, Dusan and Anne Miklas Chair in Engineering Design, University of Toronto 
 
<H2 CLASS="MsoNormal">What You Will Learn</H2>
<UL>
<LI>Computer processor design trends that motivated parallel programming and the OpenCL standard</LI>
<LI>The potential of parallel programming contrasted with reality today</LI>
<LI>The processor mechanisms that provide parallel programming</LI>
<LI>Why parallel programming is hard, from a computer science theory perspective</LI>
<LI>How the OpenCL standard is structured</LI>
<LI>The key abstractions of OpenCL (host, device, memory, and execution)</LI>
<LI>How to use the standardized C API to control OpenCL from your own applications</LI>
<LI>How to program devices using OpenCL C</LI>
<LI>How to think about parallel programming in general</LI>
<LI>Introduction to modern processor architectures and OpenCL abstractions</LI>
<LI>How to profile code</LI>
<LI>The memory model of OpenCL 1.x and 2.x</LI>
<LI>How to recognize a good parallel algorithm when you see one</LI>
</UL>
<H2>About You and What You Should Already Know</H2>
YetiWare provides training for software professionals, so please do make sure that you have the right background to attend to get the most out of it. If you are a tad rusty, you should consider our review session option so that you are ready to go for the training. If you would really like to attend but don't have the right background, reach out to us and we can let you know about other workshops that provide a gentler introduction.
 
<H3 CLASS="western">Software Development</H3>
<UL>
<LI>
You have a solid background in C
<UL>
<LI>
You can write and compile a C program without help
</LI>
<LI>
You have written a non-trivial C program at some point in your career
</LI>
<LI>
You understand pointers, pointers-to-pointers, conditionals, looping, and all that basic stuff
</LI>
<LI>
You know how to use gdb if you have to, and maybe valgrind
</LI>
<LI>
You know what a function pointer is and how to declare and use it
</LI>
<LI>
You know what a call stack is and how recursion is implemented
</LI>
<LI>
You understand bitwise operators and how they work
</LI>
</UL>
</LI>
<LI>
You understand memory management concepts
<UL>
<LI>
You know that processes have memory that is partitioned into a stack and heap
</LI>
<LI>
You know about virtual memory as provided by the operating system
</LI>
</UL>
</LI>
<LI>
You understand systems programming
<UL>
<LI>
You know how to use malloc and free and, roughly, how the OS implements them
</LI>
</UL>
</LI>
<LI>
You understand the role of machine code and how instructions are executed by processors
</LI>
<LI>
You know what an asynchronous call is
</LI>
</UL>
 
<H3>Computer Science</H3>
<UL>
<LI>
You understand data-structures and algorithms
<UL>
<LI>
Arrays, and linked-lists that are implemented with pointers
</LI>
<LI>
Interesting data-structures like balanced trees, graphs, hash tables, and so on
</LI>
</UL>
</LI>
<LI>
You understand the difference between deterministic and non-deterministic algorithms
</LI>
<LI>
You understand the asymptotic growth of functions and big-O notation
</LI>
<LI>
You understand basic probability
</LI>
<LI>
Optionally, you understand complexity theory, Turing Machines, and the complexity class P
</LI>
</UL>
 
<H3 CLASS="western">Computer Architecture</H3>
<UL>
<LI>
You understand what a processor register is and why it is needed
</LI>
<LI>
You know what a processor ISA is
</LI>
<LI>
You understand that the CPU has a stack and how functions are, roughly, called
</LI>
</UL>
 
<H3 CLASS="western">Linux</H3>
<UL>
<LI>
You know how to use the shell to compile and execute programs
</LI>
<LI>
You are sufficiently proficient with Linux to use the training server to do exercises, over SSH or with a BASH shell
</LI>
</UL>
 
<H2>About The Instructor (AJ Guillon)</H2>
AJ Guillon is a Khronos member and contributed to OpenCL 2.1, OpenCL 2.2, and the OpenCL C++ kernel language. He has worked extensively with OpenCL since December 2008, when the first OpenCL 1 .0 specification was released. AJ is the founder of YetiWare Inc, a software company developing innovative parallel platforms built on top of OpenCL. He is available to share his experience and knowledge of the OpenCL standard through consulting and training, or public speaking events. AJ is a masters swimmer, water polo player, and enjoys rock climbing when time permits. He lives in Toronto, Ontario, Canada and earned an Honors Bachelor of Science from the University of Toronto, where he specialized in Computer Science.
 
<H2>What To Bring</H2>
You will need to bring your own laptop with a wireless network card to connect to our training server.  The training server runs X2Go on Linux and contains AMD FirePro GPUs for you to try out.  You will need to install X2Go on your laptop to be able to remotely connect to the system.
Optional - Review Day (extra fee) - seperate EventBrite registration for this session
If you are attending our OpenCL course and feel that you would benefit from reviewing some concepts, then you can optionally purchase a seat at our review day.  This will be held the day prior to the start of the training, and will review background concepts required to effectively participate in the training.
The following general areas will be reviewed:
<UL>
<LI>C programming (syntax review)</LI>
<LI>CPU architecture (registers, instruction set, clock, memory, etc.)</LI>
<LI>Computer science (complexity, algorithms, asymptotic notation, Turing Machines, complexity classes, etc.)</LI>
<LI>Low-level programming (stack and heap, memory allocation, operating system concepts)</LI>
</UL>
The instructor will spend approximately half of the review session presenting material, and half of the review session answering questions or taking up exercises.
 
<H2>FAQs</H2>
Are there ID requirements or an age limit to enter the event?
Yes, please bring your printed ticket to the event so we may verify it against the confirmed and paid list of training attendees.  YetiWare does not accept registrations or payment on site.
 
What are my transport/parking options getting to the event?
We will provide links to the closest available hotel or parking lot.  Attendees who are staying over should check rates with their hotel as well.
 
What can/can't I bring to the event?
Check list above, but it is very important to bring your own laptop with X2Go configured and installed as no computers are available from YetiWare for this training.  Instead, you are connecting to a local server with X2Go. We will provide for power so you can plug in your laptops.
 
Where can I contact the organizer with any questions?
You may send any questions to <A HREF="mailto:john.reynolds@yetiware.com" TARGET="_blank" REL="nofollow noopener noreferrer">john.reynolds@yetiware.com</A> regarding the OpenCL training course
 
Is my registration/ticket transferable?
You may transfer your ticket to another team member in the same company who uses the same corporate email domain address.  Refunds for training are not available but you may register for another one of our training sessions.   This option is open for 18 months from the date of the end of the training.  We will work with you to ensure you are aware of future training sessions as we start to expand our geographic footprint.
 
Can I update my registration information?
Yes, but your company information and email address of the company will be used to verify and confirm attendance at the OpenCL training event.
 
Do I have to bring my printed ticket to the event?
Yes please, it will make the registration go a lot smoother for both parties.
 
What is the refund policy?
There is no refund for OpenCL training, however a same value credit is given for 18 months at any of our training sessions offered.  We will work with you to ensure you are aware and communicated with on future OpenCL training sessions.  Should the session be canceled a full refund will be given to the registrants.
 
The name on the registration/ticket doesn't match the attendee. Is that okay?
No, the attendee must be the person who has registered as verified by providing valid ID the day of registration.
 
What if I want to attend, but Sunnyvale is too far away for me?.
If appropriate, we can plan a specific corporate session for your company and the various team members and come to your location, contact us for details and pricing.  You can also email us and let us know where you are, and we will let you know if we are doing training in your area in future.
 
Is there an academic discount?
We love academia, and yes, we do offer discounted academic training, such as our previous course at the <A HREF="http://www.fields.utoronto.ca/activities/15-16/computing" TARGET="_blank" DATA-MSYS-CLICKTRACK="0" REL="nofollow noopener noreferrer">Fields Institute</A>.  We do not offer an academic discount for this particular training event, instead, we want to hear from you to talk about a discounted training rate for your institution or research group.  If you are on your own, please do contact us, and we will let you know about future academic training events in your area.
 
I'm an open source contributor and cannot possibly afford this, can you do something for me?
We may offer a discount, or waive fees, for open source contributors completely at our discretion.  There is an application process so that we can assess the open source project and your contributions to it.  If you are interested, please contact us, and we'll help you through the process, and let you know what we can do for you.
 
What about quantum computing?
While quantum computing looks interesting for NP-complete problems, we do not think you should wait for a quantum computer to solve your hard problems.  In fact, quantum computers are really just another form of parallelism, and at some level of theoretical abstraction likely equivalent to the parallelism we teach.
 
Should I take the review day?
The <A HREF="https://www.eventbrite.ca/e/opencl-training-review-day-tickets-25959863651" TARGET="_blank">review day</A> is offered before the course starts to make sure you have all of the background that you need to take the course, and to get the most out of it.  We will not cover any material from the training in the review day, instead, we are just going to answer questions and address gaps in knowledge prior to the start of the training.
 
<HR>
OpenCL and the OpenCL logo are trademarks of Apple Inc. used by permission by Khronos.  All other product names, trademarks, and/or company names are used solely for identification and belong to their respective owners.

Biltmore Hotel

Our {communityGuidelinesLink} describe the sort of content we prohibit on Eventbrite. If you suspect an event may be in violation, you can report it to us so we can investigate.

To report other categories of prohibited or illegal content, submit {link}