My research spans several aspects of high-performance computing. I am particularly interested in topics such as: domain-specific languages (DSL), automatic parallelisation, heterogeneous computing, compilers, embedded computing, machine learning and robot vision. At Imperial I am implied in an exciting EPSRC project called PAMELA. What is PAMELA? All about high-performance robot vision!
I received my Ph.D. degree in computer science from the universities Pierre et Marie Curie and Conservatoire National des Arts et Métiers, France, in 2011. My thesis introduces a DSL research in variational data assimilation with special emphasis in ocean modelling. The result of this research is the open source software YAO, which integrates algorithms allowing the automatisation of tasks such as the automatic parallelisation of code on shared memory architectures. YAO is currently used and developed by several labs in different countries.
Before joining Imperial, I have been a permanent researcher at Murex S.A.S., France. At Murex I have been working on MACS, a GPU-accelerated real-world financial mathematics library used by thousands of people around the world. My main focus was on extending the OpenCL standard to run MACS on clusters of GPUs.
- 02/06/2017 – The slides of my talk at iWAPT/IPDPS in Orlando, Florida, are on-line here.
- 06/03/2017 – My Advanced Computer Architecture (CO323) lecture at Imperial on GPGPUs is on-line.
- 01/03/2017 – Our “Algorithmic Performance-Accuracy Trade-off in 3D Vision Applications Using HyperMapper” (arXiv) is accepted at iWAPT/IPDPS.
19/01/2017 – The slides of my seminar at Stanford in Palo Alto are on-line here.
- 16/01/2017 – The slides of my seminar at Argonne National Lab in Chicago are on-line here.
- 01/2017 – Intel/Movidius is showcasing our SLAMBench at the CES tech show, watch the video here.
21/11/2016 – Dr Sajad Saaedi published the video on our last paper, available here.
- 29/09/2016 – My talk at the GTC-EU in Amsterdam is available on-line. The video will be available soon!
- 28/09/2016 – I am an exhibitor at the first GPU Technology Conference Europe (GTC-EU) in Amsterdam, come to visit at booth D3.
- 08/06/2016 – The schedule for the DoC RA symposium is out. The talks look amazing with our special guest Raia from Google DeepMind, join us!
- 29/04/2016 – Want to win an ACACES grant? Guess which HiPEAC member I am reconstructing in CSWSpring16. Link here.
- 20/02/2016 – The programme for our Architectures and Systems for Real-time Mobile Vision (ASR-MOV-2016) workshop hosted at CGO is out.
- 19/02/2016 – Thanks to NVIDIA for the hardware donation, I have just received a TX1, this is much appreciated!
- 03/11/2015 – I have just been elected at the DoC to represent all Research Associates and Research Assistants. Being a rep will be a lot of fun!
- 23/10/2015 – The PENCIL language workshop today was full of interesting news and a new SLAMBench implementation written in PENCIL!
- 15/09/2015 – CFP Int. Workshop on Programmability and Architectures for Heterogeneous Multicores (MULTIPROG-2016) at HiPEAC now open.
- 13/07/2015 – Download the SLAMBench Android app and share the results to rank your device, it is awesome!
- 12/05/2015 – ARM is demonstrating our SLAMBench at the Embedded Vision Summit in Santa Clara, California.
- 03/2015 – I have been wandering around in the Silicon Valley invited by AMD, Google and Qualcomm for seminars. Check out the slides here!
- 03/2015 – At the GTC we announced a new release of our SLAMBench software. Time to refresh your install for more features and performance.
- 03/2015 – I have been invited by NVIDIA at the GPU Technology Conference (GTC) in San Jose, check out the video!
- 03/2015 – Just received new Jetson Tegra K1 hardware offered by NVIDIA for our research on 3D scene understanding. Thank you guys.
- 02/2015 – My paper on SLAMBench has been accepted at ICRA 2015, in May I will be talking at this exciting robotics conference in Seattle!
- 01/2015 – We are in the news of the January HiPEAC newsletter (number 41 at page 11).
- 01/2015 – I am leading a new biweekly software performance optimisation reading group at Imperial, pop in if you like.
- 11/2014 – We are organising the 2015 Compilers for Parallel Computing (CPC) workshop at Imperial (Jan 7-9 2015).