Intel to developers: the zettascale race is on
For a company known to many for its microprocessor hardware and chipset prowess, Intel has always had one of the best tech conference programs.
The Intel Developer Forum (IDF) has arguably / maybe always been a mainstay of the Programmer’s Year.
In the current era where no firm conference bookings take place, the company has hosted its first IntelÂ® Innovation Event, where he claims to have “returned to his developer roots” with a commitment to the community and a developer-centric approach through software and hardware.
We know of course that Intel’s software-centric technology proposition revolves around a node of efficiency and architectural engineering advantages through parallelism, concurrency, and threading, that is -Tell where application execution can be channeled into streams for more efficient processing based on logical separation of software tasks. inside the program or system in hand.
âAs the creator of the original Intel Developer Forum, it is a great honor to once again bring together people from across the ecosystem to explore the future of technology. Developers are the real superheroes of the digitized world – a world backed by semiconductors. We won’t rest until we exhaust the periodic table, unlocking the magic of silicon and empowering developers, âsaid Pat Gelsinger, Intel CEO, or at least its communications function has.
Intel detailed key developer investments, including an updated, unified and more comprehensive development area, oneAPI 2022 toolkits, and new oneAPI Centers of Excellence.
All are intended to improve access to resources and simplify development on central processing unit (CPU) and accelerator architectures:
To improve ease of access to reference designs, toolkits and other assets across AI, client, cloud, 5G / edge and games, the new resource provides developers with access to a consolidated Intel Developer Catalog of major Intel software offerings, as well as an Intel DevCloud Development Environment to test and run workloads on a variety of the latest Intel hardware (CPUs, Graphics Processing Units (GPUs), Matrices programmable on-site doors (FPGAs and accelerators) and software tools.
Intel is preparing to deliver oneAPI 2022 toolkits with 900 new features since its release last year.
This new release adds cross-architecture software development capabilities for CPUs and GPUs via the first unified C ++ / SYCL / Fortran compiler and Data Parallel Python and extends the performance modeling of the Advisor accelerator, including VTune Flame Graph for Visualize performance hot spots and improve productivity with extended Microsoft Visual Studio Code integration and support for Microsoft WSL 2.
Intel suggests that we think about how computing propagates through what it calls “heterogeneous structures” of processors, graphics processors, application accelerators, interconnect processors, computer peripherals. cutting edge and FPGAs – all of which require persistent memory and software to tie these together into a complete solution.
The zettascale race is on
“The race to zettascale is enabled to generate, store and analyze data at scale. It took over 12 years to move from petascale computing to exascale computing. Intel has set itself the challenge of reaching zetta in five years: zetta 2027. At the heart of this goal is Intel’s work with the open ecosystem to ensure developers have optimized software tools and environments. to speed up their deployments, âthe company said in a press release. .
As well explained here on the Next platform, “Zettascale means zettaflops – more precisely, it means 64-bit precision zettaflops and not doing funny reduced precision counting, or retooling lower precision applications and saying it behaves like a machine with 1000 exaflops at 64-bit. “
Intel also pointed out that cloud developers have wide access to the latest 3rd generation Intel Xeon Scalable processors from major cloud service providers including Alibaba, AWS, Baidu, Google, Microsoft, Oracle and Tencent.