# S2 E1: Bit: What's the hype with Hyper?

> This is content from just-tim, the data-and-analytics channel by Tim Ngwena (formerly 'Tableau Tim'). Tim has 12+ years of hands-on BI experience and covers Tableau most of all, plus Power BI, Looker, Hex, SQL and data modelling, the analytics industry, and the craft of doing the job — always tool-agnostic and honest about the trade-offs.

- **Author:** Tim Ngwena (just-tim, https://just-tim.com/about)
- **Published:** 2024-04-22
- **Format:** Video · 51 min watch · transcript available
- **Topics:** Data engineering, Industry trends, Tool strategy
- **Tools:** Tableau (hyper, performance, prep)
- **Canonical:** https://just-tim.com/posts/s2-e1-bit-whats-the-hype-with-hyper
- **Watch:** https://www.youtube.com/watch?v=JKD1uboSz68

In this season opener of our 'What So What Now What' podcast, my co-host Ravi and I unpack Tableau's Hyper engine. We trace its academic origins, simplify the technical reasons it's so fast, and look at what it enables for the future of the Tableau platform.

## Key takeaways

- Hyper was acquired by Tableau around March 2016 and introduced in Tableau 10.5, replacing the older TDE extract format; it was built by academics in Munich, including query-optimisation expert Dr Thomas Neumann.
- Hyper's speed comes from four pillars, not one: query optimisation, query compilation via an LLVM intermediate step, doing as much processing as possible in memory, and morsel-driven parallelisation.
- Because RAM is dramatically cheaper and CPUs more capable than 20 years ago, Hyper could be designed from scratch to exploit modern hardware, something legacy enterprise databases struggle to retrofit.
- Morsel-driven parallelisation splits work into millions of tiny pieces rather than fixed per-core chunks, avoiding the diminishing returns Amdahl's Law predicts when you simply add more cores.
- Hyper aims to collapse separate transactional, analytical and beyond-relational systems into one, enabling features like Tableau Prep, live tooltips and loading data into extracts without a refresh.

## Chapters

- 0:11 What Hyper is and where it came from
- 3:00 Beyond speed: the four founding principles
- 6:30 How modern hardware changed the game
- 13:25 Query optimisation explained
- 18:09 Query compilation and LLVM
- 23:54 Why memory beats the hard drive
- 30:11 Morsel-driven parallelisation
- 35:33 Tableau's competitive advantage
- 41:51 The future Hyper enables
- 46:46 Further reading and wrap-up

Watch the full video, read the transcript and use chapter deep-links on the page: https://just-tim.com/posts/s2-e1-bit-whats-the-hype-with-hyper

---
just-tim — Data and analytics, with a point of view. · https://www.youtube.com/channel/UC7HYxRWmaNlJux-X7rNLZyw · https://twitter.com/TableauTim · https://www.linkedin.com/in/timngwena