Bull Voices - Podcast

Benchmarking the supercomputers of tomorrow

Episode 6 – 15 minutes – Published september 18, 2026

Meet Larisa in this new episode of Bull Voices. As an applications and performance engineer at Bull, she benchmarks supercomputers that don't exist yet, estimating how future systems will perform before they're even built. What does a typical day look like in this role? How has HPC evolved since she started, from single-core clusters to today's hybrid CPU-GPU systems? And what does it take to project performance for machines like the Alice Recoque national supercomputer? Discover how benchmarking shapes some of the most ambitious computing projects in France and beyond.
 

Highlights of the episode

  • -Larisa's path, from computational physics to a PhD in HPC compilers.
  • -What benchmarking really means: projecting performance for hardware that doesn't exist yet.
  • -How HPC has evolved, from single-core clusters to hybrid CPU-GPU systems and AI integration.
  • -Two contrasting projects: Alice Recoque, France's national supercomputer, and SKA, a modular system for space imaging.
  • -Why energy efficiency is becoming as important as raw performance in benchmarking.
Bull Voices #6 : Benchmarking the supercomputers of tomorrow
  • 16 min
Bull Voices #6 : Benchmarking the supercomputers of tomorrow
Bull Voices | Career stories about HPC, AI, Quantum and Servers
Play

Episode chapters

  • 00 :38 - Presentation & benchmarking
  • 02 :09 - Daily tasks in benchmarking
  • 04 :05 - How does benchmarking actually work ?
  • 05 :25 - Larisa's career path
  • 06 :02 - How has the HPC field evolved since the beginning of her career ?
  • 08 :48 - The future of the field
  • 09 :58 - Most interesting aspects of the job
  • 10 :43 - What benchmarking brings to customers
  • 11 :31 - Alice Recoque's supercomputer use cases
  • 12 :54 - Benchmarking for space imaging
  • 14 :49 - What is it like to work in benchmarking ?
  • 15 :19 - Conclusion

Speaker


larisa_headshot

 

Larisa Soltzfus

Applications and Performance Engineer

Quote highlight

"One of the tricky parts about our job is that a lot of the hardware we are proposing doesn't actually exist yet, so what we end up having to do is project numbers for systems much bigger than systems we currently have access to."

 - Larisa

Stay ahead of innovation

Subscribe to Bull Voices to get the latest insights on AI, HPC, and enterprise computing.

Contact us