[ad_1]
Asian Scientist Magazine (Sep. 04, 2023) — The last few decades of human history represent a tiny fraction of our time on Earth. However, the technological breakthroughs within this short period have transformed the human experience, from space probes venturing beyond our solar system in search of the universe’s origin to the countless artificial intelligence (AI) innovations that enable more accurate medical diagnoses, new communication methods and efficient renewable energy systems.
But this breakneck speed of technological growth and implementation comes at a cost. Many are left behind when technology races ahead of society’s ability to adapt to the new changes. This lag in learning poses a challenge for policymaking and labor markets, and also perpetuates pre-existing inequalities, particularly in developing nations.
High-performance computing (HPC) is especially susceptible to this challenge. While its lightning-fast capabilities have empowered us to solve a myriad of scientific conundrums, maintaining this momentum is critical for the next generation of HPC scientists, experts and the public to keep HPC literacy on pace with its rapidly evolving landscape.
With recent achievements in exascale computing, many challenges and questions are emerging: What educational strategies are necessary to bridge the ever expanding gap between progress in HPC technology and our understanding thereof? How can we ensure HPC literacy becomes a priority in our educational systems? And, finally, why is bridging this gap so crucial?
DECODING THE DIGITAL DIVIDE
HPC powers many applications—from operating nuclear fusion pilot plants to producing 3D animated films. At the forefront of today’s digital renaissance, technology has suffused numerous industries and will likely remain a cornerstone of innovation and progress.
HPC systems—featuring constellations of powerful computers interconnected through blazingly fast networks and sophisticated algorithms—require a deep understanding and a high level of craftsmanship to master. This encompasses a wide range of skills and expertise, spanning hardware, software and specialized domain knowledge.
What is more, the intricacies of HPC demand not only proficiency in programming languages and parallel computing techniques, but also a thorough understanding of numeral algorithms, data management and optimization strategies.
It is for this very reason that cultivating a foundational understanding of HPC is of paramount importance. Piecing together the factors contributing to the burgeoning gap currently faced in the industry is therefore essential, as it threatens our ability to harness the full potential of HPC systems and leverage their vast capabilities in the long term.
One key facet is the rapid pace of innovation in HPC technologies, which often surpasses the ability of community members and the public to acquire and adapt to new knowledge.
Unsurprisingly, the repercussions of this growing HPC literacy gap are manifold. The most pronounced and prevalent consequence is the suboptimal utilization of HPC resources. Many users, ill-equipped with the necessary technical know-how, may not fully comprehend and subsequently deploy these systems to their fullest potential. For instance, bioinformaticians might run genomic simulations on a high-performance cluster using only a fraction of its available cores, leaving a substantial portion of the machine’s computational power untapped.
This, in turn, can lead to far-reaching ripple effects—thwarting growth and stymieing research and development as businesses and researchers drown in a sea of HPC speak, jargon and technicalities.
Exacerbating the literacy gap are the increasingly high barriers to entry imposed by the ever-advancing complexity of HPC technologies. To illustrate, modern systems based on graphics processing units (GPUs), which are exceptionally adept at parallel processing, require a high level of technical understanding that can be daunting to newcomers. This situation significantly impacts the training and development of the next generation of computer scientists in the field, making it more challenging for them to acquire the necessary skills and expertise.
“Even with STEM education at the tertiary level, there remains a gap between the knowledge acquired by students—including at the PhD level—and the engineering required for cutting-edge developments in computing technology,” said Professor Sean Smith, Director of the National Computational Infrastructure (NCI) at the Australian National University, in an interview with Supercomputing Asia. “In many cases, on-the-job training is relied upon to provide this education, which can be quite happenstance,” added Smith.
On the software front, a growing number of scientific domains are pivoting towards exploiting configurations heavy on GPUs, which in turn requires the re-architecting of many algorithms and software platforms.
“This can be a challenge, especially in areas like climate and weather modeling, where extensive software platforms have taken decades to develop. Reconstructing such codes to take advantage of the most efficient hardware is not an easy feat and presents a significant roadblock for many domains,” noted Smith.
BUILDING BRIDGES TO BOOST ACCESS
As with many contemporary issues, a multifaceted approach is key, including participation of and collaboration among multiple stakeholders that encompass academia, industry and government.
When executed effectively, education and training programs can serve as powerful tools for fostering HPC literacy. For instance, under the AI & HPC-enabled Education and Talent Development for Singapore initiative unveiled at the SupercomputingAsia 2023 (SCA23) conference, several institutions—including the National Supercomputing Centre (NSCC) Singapore; Institute of Technical Education (ITE); Republic Polytechnic (RP); AI Singapore (AISG); Singapore Polytechnic (SP); Singapore Institute of Technology (SIT); and the Institution of Engineers, Singapore Incubator & Accelerator (IES-INCA)—are joining forces to develop talent in HPC among its populace.
These collaborations aim to establish new curricula, training courses, workshops and student competitions in areas such as HPC, AI, data science, analytics, and advanced simulation and modeling. By nurturing talent in these fields, the workforce will be better equipped to support relevant industry sectors and navigate the complex landscape of HPC.
Low Khah Gek, Chief Executive Officer of ITE, highlighted that the partnership with NSCC offers valuable opportunities for students to learn about HPC applications and gain first-hand experience using a supercomputer. This collaboration equips students with in-demand skills in HPC, preparing them for the jobs of the future economy.
Previously, ITE collaborated with NSCC to offer the SkillsFuture Certificate of Competency (CoC) course, which equips participants with fundamental knowledge of HPC and access to Singapore’s first national petascale supercomputer. Co-trained by ITE lecturers and NSCC specialists, students can learn how to harness an HPC system remotely to perform complex calculations at blazing speeds, expediting the development of deep learning AI applications.
Similarly, Soh Wai Wah, Principal and Chief Executive Officer of SP, emphasized his institution’s commitment to raising awareness of HPC resources among staff and students and showcasing HPC’s potential to efficiently tackle large and complex problems.
In addition to education and training programs, providing access to advanced HPC resources will help level the playing field within the HPC domain in academia. Announced at the SCA23 conference, Singapore’s next-generation national supercomputer, the Advanced Supercomputer for Petascale Innovation Research and Enterprise 2A (ASPIRE 2A), has been commissioned and will be made available to the city state’s research communities.
By granting access to the supercomputer’s resources, local institutions can advance teaching and education in fields such as AI and machine learning. This initiative also supports professionals in advanced research and IT by offering them the opportunity to hone their skills using cutting-edge HPC resources.
ASPIRE 2A’s accessibility enables all researchers to utilize the supercomputer’s resources for projects spanning areas such as climate change, biomedical science and smart nation activities. This increase in access to HPC resources and the collective efforts to develop talent in HPC-related fields exemplify a forward-looking approach to bridging the HPC literacy gap and preparing for a future driven by advanced technologies.
Singapore is not alone in its mission to foster HPC literacy. In Thailand, the National Science and Technology Development Agency (NSTDA) Supercomputer Center, commonly known as ThaiSC, not only provides cutting-edge HPC services for Thailand’s research and development community, but also actively engages current and prospective HPC users through a variety of tailored programs, workshops and events.
“To raise awareness, share knowledge and promote best practices in HPC, we regularly organize webinars featuring industry partners and academic institutions,” explained Dr. Piyawut Srichaikul, Director of the NSTDA Supercomputer Center to Supercomputing Asia. “Our latest LANTA supercomputer, which utilizes Hewlett Packard Enterprise’s advanced computer architecture, is also open for visits from schools and institutions for educational purposes.”
FORGING ALLIANCES FOR A THRIVING LEARNING ECOSYSTEM
At Australia’s NCI, outreach activities and internal training programs are conducted to help engineers embrace HPC and AI technologies. Through comprehensive surveys across its user base, NCI identified a significant gap in knowledge for students transitioning from domain-specific fields to HPC as they join the workforce.
Although many of these students have backgrounds in science or engineering, they often lack formal computer-science training. NCI addresses this gap with targeted interventions, providing solutions that differ from conventional university teachings.
“We are partnering with organizations like Intersect Australia, a leading provider of digital-skills training for researchers in Australia, to assist users in becoming acquainted with entry-level programming,” explained Smith. “As for advanced-level training, we formulate coursework in collaboration with specialized research groups across the nation, often through internship programs that leverage the expertise of these teams.”
Smith underscored the alliance between academia, industry and government as a vital catalyst in nurturing HPC literacy. Synergistic partnerships would give rise to platforms that allow diverse entities to pool their expertise and resources, which could further bolster HPC education and research.
With respect to government, NCI is not the only Commonwealth-funded research infrastructure working to bridge the HPC literacy gap in Australia. Its Perth counterpart, the Pawsey Supercomputing Research Centre, also operates an active skills training program to improve HPC literacy.
The Australian Research Data Commons, another government-funded project, also focuses on upskilling data literacy across various domains throughout the country. Through these concerted efforts, the Australian government is actively contributing to closing the gap in HPC and data literacy.
Collaborations in HPC extend well beyond national borders. The fifth iteration of the regional APAC HPC-AI Student Competition, jointly organized by NCI and NSCC Singapore, recently concluded. Over a challenging six-month period, the event brought together graduate, advanced degree and undergraduate students from across the region.
The competition provided an opportunity for participants to develop their skill sets and deepen their understanding of HPC and AI technologies, all while showcasing their mastery of these disciplines on an international stage.
Rikky Purbojati, Associate Director of Research Computing at the National University of Singapore, told Supercomputing Asia, “This competition has united students, researchers and industry professionals to address a specific set of contemporary challenges. It’s an excellent platform for educating a significant subset of the community on the potential and capabilities of HPC systems to tackle many of today’s critical issues, such as climate change and AI, among others.”
Srichaikul also shared some of ThaiSC’s global engagements, including their active role in the ASEAN HPC Task Force. Just late last year, with support from the European Union, Thailand hosted the second edition of the EU-ASEAN HPC School, a week-long intensive course on HPC and its applications to applied scientific research.
“These events provide a platform to explore methods of sharing HPC resources across various centers, addressing both technical and governance issues to bridge accessibility gaps,” added Srichaikul.
SETTING FARTHER SIGHTS
Looking beyond HPC industries has its benefits too. Gleaning indispensable insights from and emulating established frameworks of other industries can help the HPC community to draw upon proven models of publicprivate partnerships and interdisciplinary initiatives that have successfully overcome similar conundrums.
Smith points to colleagues in the quantum computing sector who have made significant strides in developing seminars, workshops and public dissemination programs to educate the general public in quantum computing—what they are, their capabilities and limitations.
“Enhancing public literacy is equally crucial, as it fosters greater awareness, interest and support for research in various domains, including HPC,” added Smith. “Learning from examples like these can help further bridge the gap between technology and public understanding.”
When it comes to educating the public about HPC, harnessing the power of narrative storytelling can be an effective conduit through which HPC knowledge is disseminated in a fun and engaging manner. Coupled with the strategic use of social media, such techniques can demystify the subject and spark curiosity among a wider audience, who might have otherwise been walled off by the technical humdrum often associated with HPC.
Embracing this approach is particularly appealing to the younger generation, who tend to be more receptive to dynamic and interactive learning methods. This not only paves the way for their understanding of HPC, but also better prepares them for the emergence of novel technologies, ensuring they remain adaptive and informed as the field evolves.
LIGHTING A PATH FOR HPC
The future of HPC is brimming with exciting emerging trends and technologies that are poised to further revolutionize the field. From exascale computing to quantum computing and advanced AI techniques, the future promises unprecedented computational power, transforming our ability to address complex scientific, engineering and societal challenges.
In parallel with these technological advancements, ongoing efforts to advance HPC literacy are evidently more critical than ever. As the field of HPC continues to evolve, so too must the resources that support HPC literacy.
Keeping up to date with the latest HPC developments and actively pursuing education and training initiatives ensure that both current and future generations of users are equipped with the knowledge and skills required to effectively harness emerging technologies.
To stay at the forefront of technological advancements, Purbojati suggested, “Individuals can keep tabs on HPC publications, follow relevant social media channels and keep an eye on academic journals for deeper insights.”
“Communities can also play an important role by forming organizing bodies that distribute regular newsletters and hosting engaging events, such as user group meetings, to keep everyone informed about current technologies and what the future holds,” added Purbojati.
Augmented by interdisciplinary approaches that foster strategic collaborations between academia, industry and government, the collective efforts of the HPC community can help cultivate a thriving and supportive ecosystem that nurtures talent and promotes innovation in HPC and beyond.
–
This article was first published in the print version of Supercomputing Asia, July 2023.Click here to subscribe to Asian Scientist Magazine in print.
Copyright: Asian Scientist Magazine.
Disclaimer: This article does not necessarily reflect the views of AsianScientist or its staff.
[ad_2]
Source link
