Alexander Gounares
CEO @ Concurix Corporation
Concurix
Alex Gounares is the founder and CEO of Concurix Corporation. Concurix is a new startup focused on building an extremely scalable and high performance operating system for modern data centers. Concurix is headquartered in Kirkland, Washington.
Prior to Concurix, Gounares served as AOL's Chief Technology Officer. In this role, he lead all aspects of AOL's technology strategy, platform development and external technology partnerships. He was responsible for all of AOL's global engineering, IT, and operations functions. In addition, he served as a member of the company's Global Executive Operating Committee.
Gounares joined AOL from Microsoft, where he was Corporate Vice President and Chief Technology Officer for the company's Online Services Division. During his tenure at Microsoft, Gounares led significant strategic and technical operations for some of the company's most important projects including Microsoft's global advertising platform, Bing search, MSN and Microsoft Virtual Earth. Gounares also served for three years as Technology Advisor to Microsoft Chairman and founder Bill Gates, as well as Corporate Vice President of Corporate Strategy in Microsoft's Finance Department.
Prior to joining Microsoft in 1993, Gounares worked at Los Alamos National Laboratory. He has founded four startups and is also an inventor on more than 130 U.S. patents filed and pending. Gounares holds a bachelor's degree cum laude in Computer Science from Princeton University.
Alexander Gounares is Giving the Following Talks
Scaling Erlang In The Multicore Era
Modern CPUs have continued to follow Moore’s Law, continuously offering stunning increases in capability with every generation. However, software in general has not kept pace, particularly in terms of fully leveraging the power of the latest multi-and many-core CPUs.
The root of the challenge is Amdahl’s Law—the speedup in a parallel system is governed by the amount of serialization in the code. Writing parallel algorithms can be hard—and hard to write correctly. And even when written correctly, locking conventions for shared state in traditional programming languages can introduce serialization bottlenecks.
Erlang has proven to be a phenomenal platform for scaling on multi and many-core systems, thanks to its inherent concurrency. However, even then there are challenges to fully harnessing the power of modern chips.
Alex's talk will cover the latest work in scaling Erlang, and in particular will focus in on using big data analysis techniques to better understand and improve Erlang programs.