Jesus Garcia is responsible for IT LNG & Innovation at Repsol, one of Europe’s largest oil and gas companies.

In which application areas do you use HPC?

The primary use of HPC within Repsol is for the processing of seismic data. Repsol uses a range of algorithms some of which are proprietary, for example Reverse Time Migration (RTM). The company also runs some financial risk calculations on a small cluster.

Why do you use HPC ?

HPC is the only way that large amounts of seismic data can be processed quickly. When prospecting for oil and gas, there is significant commercial advantage in knowing where the most promising fields are, given the highly competitive nature of the sector.

What are the cost benefits to your business of using HPC?

As described above, there are significant commercial pressures to get results from prospecting within days rather than weeks.

Which HPC systems do you use in your business? What is their capital value? How scalable are your applications?

Repsol uses resources at the Barcelona Supercomputing Centre (BSC) comprising an IBM system with PowerPC processors. There is also a significant technical collaboration between the two organisations. Repsol also has a cluster of 300 cell processors in Houston, Texas. Seismic processing is highly scalable because different data sets are independent.

For your business does the Cloud offer a viable alternative to owning and managing your own systems?

The Cloud is a very good option for the type of applications Repsol has. However, we do not yet use the Cloud because some further development of languages and programming methodologies is needed to give portability and performance. The Cloud could already be suitable for some smaller applications such as reservoir simulation. Security does not seem to be a major problem for companies such as Repsol.

What are the challenges you see in the development of your HPC capability (e.g. scalability of applications, power consumption, cost of systems)?

The big challenges are portability of applications across different systems and power consumption. Power consumption of HPC systems, in particular, is becoming a more and more important factor. The very rapid development of HPC technology makes investment in hardware very difficult to justify and, in many cases, it is better to hire computing time on centralized facilities such as those at BSC.

Are new languages and programming paradigms needed particularly as we move toward exascale systems?

There is clearly a need to protect the significant investment already made in existing software. Portability is very important. Although performance is also a significant factor, as systems get more and more powerful, it should be possible to trade-off performance against portability.

What are your views on GPGPU computing?

Repsol has been doing some tests with GPGPU systems for seismic processing. The results so far are promising and can complement other established technologies. Again, when using GPGPUs and, indeed, any attached processing devices, there need to be standards to protect the significant investments in software and to minimise the need to adapt code across a range of devices. For GPGPUs and other attached devices, memory bandwidth is also an important issue which has implications for data structures and the overall performance of an application.

What are your views on reconfigurable (i.e. FPGA-based) computing particularly in light of developments at Convey?

Repsol has considered this type of device but does not yet have any implementations on FPGAs. This may be a promising area and FPGAs may be very good for certain types of algorithms such as FFTs which are very common in seismic processing. Of course, memory bandwidth would also be an important consideration here.

Describe the HPC systems you would like to have available in 3, 5 and 10 years' time.

The key features would be portability with no significant performance losses and low power consumption. Half the total cost of ownership (TCO) of HPC systems is the power consumption. A “wonder” compiler that adapt algorithms to parallel devices yet achieving a good performance would also be desirable, but this may be more of a wish than a reality.

In which HPC research areas would EU-funded programmes benefit your business?

There is a lot of HPC expertise in the EU and HPC facilities. This should be better used by industry. The EU has a role to play here. A network of HPC resources, academic experts and industrial users should be set up. A key issue would be how to set up this network and how to ensure that applications could move easily from the systems at one centre to those at another.

There also needs to be research to adapt neural network and artificial intelligence systems to complex real-world applications. Such applications would require significant levels of computer power. Furthermore efforts should be made to port significant libraries such as Matlab and SciLab to high-performance systems.

Our thanks Mr Garcia for such clear presentation of the computing needs of a large energy company.

Comment on this interview