Profile cover photo
Profile photo
Lorena Barba
357 followers -
Engineering professor, computational scientist, jazz buff, techie, mac fan, academic writer and font geek.
Engineering professor, computational scientist, jazz buff, techie, mac fan, academic writer and font geek.

357 followers
About
Lorena's posts

Post has attachment
I wrote a short piece for Science — a story of struggles towards reproducible CFD ... I hope you like it.

Post has attachment
"But, can I copy text from Wikipedia?"—Confused Student.
"No."

Post has attachment
"Numba: Tell those C++ bullies to get lost" is the title of our SciPy 2016 tutorial (11 July, Austin, TX).

Here's a blog post by +Gil Forsyth  explaining some changes in the tutorial, due to technical reasons—and imminent feature improvements in Numba because of it.

Post has attachment
We have just finished writing up a rather painful study, submitted our paper (pre-print on arXiv), and would be delighted to get comments from our friends and colleagues ... Disqus comments accepted on our group website. Twitter convo also welcome.

Post has attachment
Reactions to my tip on how I use figshare
A strategy that serves both our open-science and reproducibility goals, also helps with keeping copyright and releasing figures under CC-BY

Post has attachment
"Using AmgX to Accelerate PETSc-Based CFD Codes"
—This technical report supplements the presentation by Pi-Yueh Chuang at the 2016 GPU Technology Conference, on Thu. April 7th: session S6355. (Video will soon be available.)

In this work, we set out to accelerate PetIBM using the NVIDIA AmgX library on multi-GPUs. We wrote a wrapper code to couple AmgX and PETSc, so that we require few modifications in PetIBM. Our AmgX wrapper is able to handle the situation where the PETSc-based code launches more MPI processes than the number of GPUs available, achieving heterogeneous speed-up. It is open source and we hope others may find it useful to accelerate their PETSc-based parallel programs with multiple GPUs.

Post has attachment
I just published this piece discussing the popular message of the "computational thinking" buzzword, and the deeper, more powerful idea that Seymour Papert had in mind when first using the phrase 30 years ago ... 

Post has attachment
"Real or AI?" ... Ha ha! This is really hilarious!
I love it.

Post has attachment
I wrote this as a motivation document for the panel I'm chairing next Tuesday at the III Arnold Workshop: Modeling & Code, at the headquarters of the AAAS in Washington DC.

If you have comments that you'd like me to bring to the panel discussion, I would be glad to!

Post has attachment
Please consider an end-of-year, tax-deductible donation to NumFOCUS, the 501(c) charity and umbrella organization providing fiscal sponsorship to open-source software projects like AstroPy, IPython, Julia, Jupyter, Matplotlib, NumPy, pandas, PyTables, rOpenSci, Sympy and training organizations Data Carpentry and Software Carpentry.
Only 1 day left to reach the $5,000 goal, doubly matched by two anonymous donors!
https://www.flipcause.com/secure/cause_pdetails/NDE2OQ==
Wait while more posts are being loaded