skip to Main Content
ANTICIPATING OUR INVOLVEMENT IN THE 2019 RICE UNIVERSITY OIL AND GAS HPC CONFERENCE

ANTICIPATING OUR INVOLVEMENT IN THE 2019 RICE UNIVERSITY OIL AND GAS HPC CONFERENCE

TL;DR: RiceU’s is the de facto oil/gas event. Sylabs to sponsor. Singularity workshop 3/6/2019, 9 am. See you there!

As I wrote on my personal blog in 2015:

RTM [Reverse Time Migration] has a storied history of being performance-challenged. Although the method was originally conceived by geophysicists in the 1980s, it was almost two decades before it became computationally tractable … of particular note are the consistent gains being made since the introduction of GPU programmability via CUDA, as innovative algorithms for RTM can exploit this platform for double-digit speedups.

For many within the community, the introduction to computational exploitation of GPUs would have been the Rice University Oil & Gas HPC Workshop in 2008. Some eight years later, I’d expend some algorithmic effort myself refactoring RTM’s I/O bottleneck through novel use of Apache Spark. Although the industry isn’t as focused on RTM specifically in more recent times, computational, I/O, and other bottlenecks remain points of contention in Full Waveform Inversion (FWI) and elsewhere as even more recent programs of the RiceU event bear testimony.

For over a decade then, the Rice University Oil & Gas HPC Conference (as it became known in 2016) has been a microcosm of the broader HPC space. If you scan the programs over the past decade plus, you’ll observe another salient data point: starting in 2016, the industry initiated its application of Machine Learning to problems of relevance to the energy-exploration sector. In other words, the scope of the RiceU event broadened and deepened once again to accommodate Enterprise Performance Computing (EPC) – a term coined by us to highlight the compute-driven emphasis of HPC plus Deep Learning workloads within the private sector.

Owing to our common interest in EPC then, Sylabs is proud to have been recently identified as the first “Launch Pad” level sponsor of the 2019 edition of this popular event. In another first for the RiceU event, this brand-new tier allows start-ups like Sylabs to actively participate in an entry-level sponsorship role.

In addition to sponsoring the event, members of the Sylabs team will deliver a post-conference workshop on Singularity containers for oil and gas EPC workloads. As someone who’s spoken about containerization at this same event on two separate occasions in the past (here and here), it’s my impression that previous introductions of this technology failed to ignite serious interest within the oil and gas industry. Because Singularity containers were designed with HPC firmly in mind from the outset, the prospects for containerized EPC workloads are, well, singularly different now. As detailed in the description for the workshop, our plan is demonstrate this difference by applying Singularity to real-world EPC use cases of relevance to the oil and gas community. Because Singularity is readily available as open source software, the on-ramp for adoption is surprisingly gentle – even when it comes to real-world applications involving Deep Learning frameworks such as distributed TensorFlow.

This year’s event will be held March 4-6 at Rice University’s downtown Houston campus. If you’re already in the area, you likely know about the event, and have plans to attend; all we want to do here is encourage you to drop by the start-ups’ area (in the main auditorium lobby area) and to join our Singularity workshop. If you’re not local to Houston, you might want to make plans to attend, as the RiceU event also has a reputation for drawing in an increasingly international mix of presenters and attendees. We look forward to seeing you there!

There’ll be plenty more to share leading up to and following the 2019 Rice University Oil and Gas HPC Conference. Please stay tuned.

[Featured Image: The BioScience Research Collaborative (BRC) Building at Rice University. Taken by the author in March 2016.]

Back To Top