Sign in

Jorge Luis Gálvez Vallejo

@jorgegalvez.bsky.social
249 followers 695 following 151 posts

HPC Specialist at the National Computational Infrastructure Computational chemist by training, high performance computing by accident. Warhammer 40k, Magic the Gathering, Science Fiction, and Latin American Literature aficionado.

PostsRepliesMedia
Jorge Luis Gálvez Vallejo @jorgegalvez.bsky.social · 22/08/2026
Everything except coupled cluster has analytical gradients and rhf has analytical Hessians. It's connected to a geometry optimizer. I also have FMO, EE MBE, MBE, GMBE. FMO and eembe cannot cut bonds yet. I have pcm too.
000
Jorge Luis Gálvez Vallejo @jorgegalvez.bsky.social · 22/08/2026
If you try it again I've gone to town on my code. I have rhf, uhf, DFT, unrestricted DFT, mp2, rimp2, unrestricted MP2/rimp2, ccsd(T), riccsd(t), for DFT all the way up to double hybrids. Except unrestricted dft everything can be done with density fitting.
100
Jorge Luis Gálvez Vallejo @jorgegalvez.bsky.social · 10/05/2026
I personally find that FMO is a bit painful due to the self consistent charge bit for the monomers, leads to annoying parallelism for the polymers. I think with xtb you can do a very good screening of interactions and then avoid some redundant calculations.
020
Jorge Luis Gálvez Vallejo @jorgegalvez.bsky.social · 10/05/2026
Have you tried a pure MBE instead of FMO? John Herbert's group developed a python library called fragments for this. I wrote github.com/JorgeG94/met... To do fragment based quantum chemistry with xtb.
github.com
GitHub - JorgeG94/metalquicha: A hastily put together framework for fragmented quantum chemistry
A hastily put together framework for fragmented quantum chemistry - JorgeG94/metalquicha
110
Jorge Luis Gálvez Vallejo @jorgegalvez.bsky.social · 11/03/2026
Most of my post PhD career can be summarized by "Turns out doing (insert thing) is hard."
000
Jorge Luis Gálvez Vallejo @jorgegalvez.bsky.social · 01/03/2026
Yeah lack of offer for positions in academia makes someone stay forever unless they're truly tired
000
Jorge Luis Gálvez Vallejo @jorgegalvez.bsky.social · 01/03/2026
This is 100% true
000
Jorge Luis Gálvez Vallejo @jorgegalvez.bsky.social · 01/03/2026
A Start up with terrible management (most) can be horrendous, especially in countries with little protections for workers.
110
Jorge Luis Gálvez Vallejo @jorgegalvez.bsky.social · 31/12/2025
ah yes, commit messages of a well studied person
020
Reposted by Jorge Luis Gálvez Vallejo
Journal of Open Source Software @joss-openjournals.bsky.social · 20/12/2025
Just published in JOSS: 'FSML – A Modern Fortran Statistics and Machine Learning Library' doi.org/10.21105/joss.09058
051
Jorge Luis Gálvez Vallejo @jorgegalvez.bsky.social · 16/12/2025
I've passed on the rec to my wife who can wander while I conference! See you around
000
Jorge Luis Gálvez Vallejo @jorgegalvez.bsky.social · 11/12/2025
Ah that's very annoying.
010
Jorge Luis Gálvez Vallejo @jorgegalvez.bsky.social · 11/12/2025
I thought they'd been doing it for a bit? I've been required to submit my socials since I applied for my first student visa in 2016. Had to do it for my H1, F1 and tourist ones. Is this for ESTAs?
110
Jorge Luis Gálvez Vallejo @jorgegalvez.bsky.social · 06/12/2025
Building MPI today and for the first time I've noticed the Infinite Memory Engine, amazing
010
Jorge Luis Gálvez Vallejo @jorgegalvez.bsky.social · 15/11/2025
There's also so much around when they started playing with computers !
000
Jorge Luis Gálvez Vallejo @jorgegalvez.bsky.social · 14/11/2025
Oh so you're probably going to cover the early history of quantum chemistry! I've been annoyed that there's no sci comm book that covers it
100
Jorge Luis Gálvez Vallejo @jorgegalvez.bsky.social · 13/11/2025
Congrats!! Since you go the cern direction I assume you go more towards particle physics?
110
Jorge Luis Gálvez Vallejo @jorgegalvez.bsky.social · 13/11/2025
When not writing code I try to paint
010
Jorge Luis Gálvez Vallejo @jorgegalvez.bsky.social · 07/11/2025
Thanks!
010
Jorge Luis Gálvez Vallejo @jorgegalvez.bsky.social · 05/11/2025
I go back and forth on whether this code needs to be rewritten from scratch. Today it is a day where I think it should.
000
Jorge Luis Gálvez Vallejo @jorgegalvez.bsky.social · 05/11/2025
If I can't run your software using mpirun -np X your_executable and I need a custom script that goes beyond abstracting a more complex mpirun, like: mpirun -np X --map-by ppn:X:node then I hate your software
000
Jorge Luis Gálvez Vallejo @jorgegalvez.bsky.social · 05/11/2025
How are the works put into the anthology selected? I sadly rarely read these type of pieces because I don't know where they're published :(
100
Reposted by Jorge Luis Gálvez Vallejo
Dr Abeba Birhane @abeba.blacksky.app · 29/10/2025
my favourite type of emails for collaboration/advice is from start-ups and for-profit companies. friend, there’s nothing i would love to do with my time more than review and advise on your profit maximising, socially harmful (useless at best), and borderline pseudoscience tools
132419
Jorge Luis Gálvez Vallejo @jorgegalvez.bsky.social · 27/10/2025
I work with a code that has been in continuous development since the late 1970s and it has almost every flavor of Fortran in it
010
Jorge Luis Gálvez Vallejo @jorgegalvez.bsky.social · 20/10/2025
If your arrays are sorted, certain bits of your "sorting" code won't be called. From "On increasing test coverage", by me.
000
Jorge Luis Gálvez Vallejo @jorgegalvez.bsky.social · 19/10/2025
and because I cannot be bothered - with each push my entire CI is running...my code is getting super tested.
000
Jorge Luis Gálvez Vallejo @jorgegalvez.bsky.social · 19/10/2025
Deploying documentation via github actions and trying to figure out "where in the world" are my images, is a fun Sunday night experience
110
Jorge Luis Gálvez Vallejo @jorgegalvez.bsky.social · 07/10/2025
Talk title: Ab initio simulation of magnetic hysteresis in single-molecule magnets. Please register in the link below! we hope to see you there! Link: anu.zoom.us/meeting/regi... Please share! We hope to see you there! #compchem #chemistry #abinitio #community
010
Jorge Luis Gálvez Vallejo @jorgegalvez.bsky.social · 07/10/2025
We are delighted to invite you to our upcoming AMMA webinar, this time next Monday October the 13th at 12:00 Sydney time. This month we have Dr. Ben Atkinson postdoctoral researcher in the Chilton Lab @nfchilton.bsky.social at the Australian National University
111
Reposted by Jorge Luis Gálvez Vallejo
Andreas Herten @andih.bsky.social · 28/07/2025
How about writing up something exciting about #HPC accelerators? Or fancy system tests? We’ve extended the deadline of the #SC25 #WACCPD workshop – and #HPCTESTS is also still open. Both close Friday! - WACCPD: waccpd.org/call-for-pap... - HPCTESTS: olcf.github.io/hpc-system-t... #GPU #Benchmarking
212
Jorge Luis Gálvez Vallejo @jorgegalvez.bsky.social · 03/07/2025
this is what a true commit message looks like
010
Jorge Luis Gálvez Vallejo @jorgegalvez.bsky.social · 17/06/2025
The website for MM25 is up! Come down to beautiful Melbourne, Australia to enjoy a conference for all sorts of computational chemists - from MM to QM, from B3LYP to wB97x-D3; molecular or periodic! More information to come! #compchem #chem #conference #chemistry #australia #melbourne #molecular
mgmsamma.org
mm25
Molecular Modelling Conference Website
010
Jorge Luis Gálvez Vallejo @jorgegalvez.bsky.social · 11/05/2025
This is a fantastic opportunity for soon to be PhD grads and early career postdocs! Everyone is welcome :)
000
Jorge Luis Gálvez Vallejo @jorgegalvez.bsky.social · 11/05/2025
Please join us today for a lunchtime (in Australia) discussion on moving on from research roles to software engineering roles joined by my former colleague Dr Melisa Alkan at NVIDIA and Dr. Cameron Hyde at QCIF. Link: anu.zoom.us/meeting/regi... #compchem #amma #webinar #software #hpc
100
Jorge Luis Gálvez Vallejo @jorgegalvez.bsky.social · 08/05/2025
Lol delivered 2021
010
Jorge Luis Gálvez Vallejo @jorgegalvez.bsky.social · 17/04/2025
Big G at RSC today in Canberra
000
Jorge Luis Gálvez Vallejo @jorgegalvez.bsky.social · 09/04/2025
You're better off not looking at it haha
010
Jorge Luis Gálvez Vallejo @jorgegalvez.bsky.social · 08/04/2025
Have you seen that there's a sub stack they contribute to?
100
Jorge Luis Gálvez Vallejo @jorgegalvez.bsky.social · 03/04/2025
The link to register for the talk in the first post! Feel free to come by and share with your networks! #compchem #physchem
010
Jorge Luis Gálvez Vallejo @jorgegalvez.bsky.social · 03/04/2025
AMMA cordially invites everyone to our monthly webinar series! The talk is at 12:00 Sydney time and will be recorded! We are still ironing out the kinks for the recording distribution. anu.zoom.us/meeting/regi... Prof. Haibo Yu will talk to us about Biological Simulations!
120
Jorge Luis Gálvez Vallejo @jorgegalvez.bsky.social · 31/03/2025
Indeed, not all code needs gpus, forgot to clarify
000
Jorge Luis Gálvez Vallejo @jorgegalvez.bsky.social · 31/03/2025
Go get lunch, if after lunch it is still restarting, power cycle. This is my approach, if I'm going to brick my computer I'd rather have had lunch.
000
Jorge Luis Gálvez Vallejo @jorgegalvez.bsky.social · 31/03/2025
Being able to run on GPUs is quite important, though. Finding good ways to do it from Fortran is a very fun problem!
110
Jorge Luis Gálvez Vallejo @jorgegalvez.bsky.social · 26/03/2025
La reputísima madre que los parió. A chingar a su madre. Pardon my Spanish.
000
Jorge Luis Gálvez Vallejo @jorgegalvez.bsky.social · 13/03/2025
They should cover the USA...
010
Jorge Luis Gálvez Vallejo @jorgegalvez.bsky.social · 13/03/2025
It looks like a kitchen bench. Maybe the martians place a barbecue right by it for some snags
010
Jorge Luis Gálvez Vallejo @jorgegalvez.bsky.social · 07/03/2025
Oh this is great to know! Will build mpich then
010
Jorge Luis Gálvez Vallejo @jorgegalvez.bsky.social · 05/03/2025
Most excited for the ABI!
110
Jorge Luis Gálvez Vallejo @jorgegalvez.bsky.social · 05/03/2025
Excited to try this later this year!
100
Jorge Luis Gálvez Vallejo @jorgegalvez.bsky.social · 25/02/2025
#chemsky #compchem #scicomm
010