• Subscribe

The cost of an imbalanced HPC work force

Speed read
  • HPC lags behind other STEM fields in representation of women
  • Broadening appeal of HPC to women will also attract other groups
  • Recruitment crisis in technology can’t be solved without women

Toni Collis is a physicist by training, with a PhD in computational simulation of biological systems using both classical and quantum mechanics. She is currently an applications consultant at the Edinburgh Parallel Computing Centre (EPCC).

Collis is also co-founder of Women in High-Performance Computing (WHPC), an organization dedicated to raising awareness about diversity in computer science and seeking to encourage more women to participate in HPC. 

We caught up with Collis at PRACEdays17 in beautiful Barcelona.  


When did you first start thinking about gender equality and gender parity in STEM and HPC?

When I first came to EPCC I thought, ‘Wow, HPC is so much more diverse than physics!’ But then I started working on international collaborations. I would attend a project meeting, and I’d be the only woman in the room. It happened every single time. It made me wonder: How many women are there? Were my meetings unusual or the norm?

<Strong>Toni Collis</strong> argues that not only is diversity in the tech world a moral imperative, it will also increase opportunity and productivity — for men and women. Courtesy Toni Collis.

A lot of people use HPC, but that’s not how they brand themselves. That makes it hard to go to a university or company and say ‘Tell me how many women are in your HPC department.’ The counting alone is a huge challenge— actually implementing changes to bring about improvements in diversity is even more so.  

So how do you count women’s participation in HPC?

Our main source of information is HPC conferences, but we don’t know whether they’re representative of the broader community.

Some people say women may attend conferences in greater numbers because women are believed to be better presenters or networkers (a generalization I don’t necessarily agree with!). The other side says that women may be less likely to go to conferences because they’re more likely to have caring responsibilities.

The key point is no one knows.

Big conferences such as Supercomputing have 13 percent attendance by women. PRACEdays last year was 17 percent. But at a more niche event like the annual PGAS conference, the proportion of women drops to as low as five percent.

We need to get out there and tell people, ‘When you use parallel computing you can make the world a better place’ rather than, ‘When you use parallel computing you can use the fastest machine in the world.’~ Toni Collis 

The MPI Forum, which develops the Message Passing Interface standard that most of us use to parallelize the big machines, is hugely influential in our community yet there are only two female contributors on the current standards body.

Let's talk about the confluence of causes — What is causing the disparity in HPC?

Recruitment is probably our biggest hurdle. We aren’t doing as well as our feed-in subjects—we’re not attracting female students at the rate that they’re graduating. We need to get out there and tell people, ‘When you use parallel computing you can make the world a better place’ rather than, ‘When you use parallel computing you can use the fastest machine in the world.’ We need to be diverse about the image we portray.

The other problem is retention. In Europe, 41 percent of women leave the technology sector by the mid-point in their career, compared to just 17 percent of men. The reason seems to be an accumulation of discomfort in the workplace and lack of promotion opportunities and sponsorship.

If you’re in an under-represented group, you’re less likely to be listened to when you have an idea, less likely to have an idea taken forward, and less likely to receive credit if it is. Women have as much ability as their male peers—let’s give them the same opportunities.

One of the arguments for parity is not just that it’s ‘right,’ but that it improves productivity and increases innovation. How will having more women in HPC lead to better outcomes?

If everyone’s the same, there’s no ‘thinking outside the box.’ Diversity of thought means you’re more likely to challenge an idea. You get more publications from diverse teams, better citation rates, and more innovation.

Science Node live. A quick chat with Women in HPC co-founder Toni Collis at PRACEdays17 in Barcelona.

Also, women across the world are the primary purchasers. If only men are designing your products, you are ignoring a minimum of 51 percent of your purchasers.

In some cases, the female perspective is just ignored. For example, when Google Translate ‘turned-on,’ overnight the internet went predominantly male because nobody on the team had thought about female pronouns. Hopefully a more diverse team would have picked that up.

What would you say to someone who’s still resistant to re-balancing?

Technology is facing a recruitment crisis. We need millions of technology workers around the world, and we’re not going to fill those with our current graduates. We are definitely not going to fill those without including women.

Many of the things I’m talking about, particularly about making the tech workplace more appealing, don’t just apply to women. It’s about giving everybody opportunities, and everybody credit where credit is due. It’s about making sure everybody feels included. 

Meet the rest of the WHPC team.

<strong>Alison Kennedy</strong> — Chair of the Board of Directors and Manaaging Director of the <a href= 'http://www.prace-ri.eu '>Partnership for Research Computing in Europe (PRACE).</a>
<strong>Kim McMahon</strong> — President and CEO of McMahon Consulting.
<strong>Kelly Nolan</strong> — Executive director, external affairs for <a href= 'https://www.computecanada.ca'> Compute Canada </a>.
<strong>Sharon Broude Geva</strong> — Director of <a href='http://arc.umich.edu'>Advanced Research Computing (ARC)</a> at the University of Michigan.
<strong>Lorna Rivera</strong> —Research scientist in program evaluation at the <a href='https://www.ceismc.gatech.edu'> Georgia Tech Center for Education Integrating Science, Mathematics and Computing (CEISMC)</a>.
<strong>Marjolein Oorsprong</strong> — Communications Officer at the Partnership for Advanced Computing in Europe (PRACE).
<strong>Athina Frantzana</strong> — PhD research student in <a href= 'https://www.epcc.ed.ac.uk'> EPCC (Edinburgh Parallel Computing Centre)</a> at the school of Physics & Astronomy of the University of Edinburgh.

Join the conversation

Do you have story ideas or something to contribute? Let us know!

Copyright © 2023 Science Node ™  |  Privacy Notice  |  Sitemap

Disclaimer: While Science Node ™ does its best to provide complete and up-to-date information, it does not warrant that the information is error-free and disclaims all liability with respect to results from the use of the information.


We encourage you to republish this article online and in print, it’s free under our creative commons attribution license, but please follow some simple guidelines:
  1. You have to credit our authors.
  2. You have to credit ScienceNode.org — where possible include our logo with a link back to the original article.
  3. You can simply run the first few lines of the article and then add: “Read the full article on ScienceNode.org” containing a link back to the original article.
  4. The easiest way to get the article on your site is to embed the code below.