Just published in JOSS: 'jax-smfsb: A Python library for stochastic systems biology modelling and inference' https://doi.org/10.21105/joss.07491
February 5, 2025 at 5:29 PM
my limited exposure to scicomp has been more than enough to convince me that we've got to stop letting scientists maintain their own code *immediately* because it's a miracle they haven't yet caused the collapse of the entire scientific endeavor
February 26, 2025 at 11:13 PM
my limited exposure to scicomp has been more than enough to convince me that we've got to stop letting scientists maintain their own code *immediately* because it's a miracle they haven't yet caused the collapse of the entire scientific endeavor
Proudly open our freshly printed chapter "Laser Capture Microdissection for Quant Shotgun Lipidomics" @springernature.com. The best #lipidomics protocols we developed & optimized over years in collaboration with SciComp group @mpi-cbg.de - cudos to Silvia Radrezza!
link.springer.com/protocol/10....
link.springer.com/protocol/10....
March 23, 2026 at 10:09 AM
Proudly open our freshly printed chapter "Laser Capture Microdissection for Quant Shotgun Lipidomics" @springernature.com. The best #lipidomics protocols we developed & optimized over years in collaboration with SciComp group @mpi-cbg.de - cudos to Silvia Radrezza!
link.springer.com/protocol/10....
link.springer.com/protocol/10....
cannot fucking believe how well this is working. I think I can publish this application this week. (this is my self-hosted Qwen2.5-VL-3B-Instruct-powered handwritten note data mining app.)
![(maximally loaded RTX 5060 Ti w/ 16GB RAM)
sudo nvtop Aes/onyx/Note Max/Notebook + Device 0 [NVIDIA GeForce RTX 5060 Ti] PCIe GEN 4@ 8x RX: 210.5 MiB/s TX: 33.52 MiB/s GPU 2805MHz MEM 13801MH TEMP 56°C FAN 53% POW 59 / 180 W GPU[ |||| 100 GPU0%- GPUO mem% 75 50 25 0 ||||96%] MEM[ ||| | |15.611Gi/15.929Gi] L44s- -335- -225- -11s- -05- PID USER DEV TYPE GPU GPU MEM CPU 76281 sysop 0 Compute 32% 5168MiB 32% 100% HOST MEM Command 2977MiB /home/sysop/RectangularFile/v 76299 sysop 0 Compute 32% 5168MiB 32% 100% 76300 sysop 0 Compute 32% 5168MiB 32% 100% 2296MiB /home/sysop/RectangularFile/v 2465MiB /home/sysop/RectangularFile/v F2Setup F6Sort F9Kill F10Quit F12Save Config](https://cdn.bsky.app/img/feed_thumbnail/plain/did:plc:34d6gxlllrndltpas55xfedz/bafkreic25xfkwkhl3ofo4b76zzxuhwazq33xafvvrema6j5lsdlnq6svfi)
June 11, 2025 at 8:19 AM
cannot fucking believe how well this is working. I think I can publish this application this week. (this is my self-hosted Qwen2.5-VL-3B-Instruct-powered handwritten note data mining app.)
Qual semestre de scicomp a gente aprende isso
June 2, 2025 at 10:19 PM
Qual semestre de scicomp a gente aprende isso
In general I might be tempted to do a project like this in pieces or stages.
Stage 1: Requirements translated into a rigorous test suite.
Stage 2: Backend that does the hard scicomp parts in a way compliant with the test suite.
Stage 3: Visualization and UI.
Stage 1: Requirements translated into a rigorous test suite.
Stage 2: Backend that does the hard scicomp parts in a way compliant with the test suite.
Stage 3: Visualization and UI.
May 10, 2026 at 8:06 PM
In general I might be tempted to do a project like this in pieces or stages.
Stage 1: Requirements translated into a rigorous test suite.
Stage 2: Backend that does the hard scicomp parts in a way compliant with the test suite.
Stage 3: Visualization and UI.
Stage 1: Requirements translated into a rigorous test suite.
Stage 2: Backend that does the hard scicomp parts in a way compliant with the test suite.
Stage 3: Visualization and UI.
Gokberk Kabacaoglu has wrapped up our #SC25 booth presence in collaboration with @ucl-arc.bsky.social and @cake-dri.bsky.social as well as other UK universities at scicomp.webspace.durham.ac.uk/scicomp-at-s... Thanks for dropping by!
SciComp at Supercomputing 2025 - Scientific Computing at Durham
A Landmark Week for UK HPC Collaboration in St. Louis at SC25 The SciComp team has returned from an inspiring and action-packed week at Supercomputing 2025 (SC25) in St. Louis. Together with colleague...
scicomp.webspace.durham.ac.uk
November 26, 2025 at 12:53 PM
Gokberk Kabacaoglu has wrapped up our #SC25 booth presence in collaboration with @ucl-arc.bsky.social and @cake-dri.bsky.social as well as other UK universities at scicomp.webspace.durham.ac.uk/scicomp-at-s... Thanks for dropping by!
The cool part about scratching the surface of knowledge, though, is that there is always more knowledge to gain, and as I keep digging deeper, I feel like I've gained more knowledge from two months on this scicomp track than I gained from two years of trying to push gamedev past hobbyist stuff.
At this point I know SO much about the stuff they make and do that if they did end up interviewing me, I'm starting to run out of stuff I'd need to ask them
April 7, 2026 at 5:18 PM
The cool part about scratching the surface of knowledge, though, is that there is always more knowledge to gain, and as I keep digging deeper, I feel like I've gained more knowledge from two months on this scicomp track than I gained from two years of trying to push gamedev past hobbyist stuff.
If what I described is your use profile I would consider one of:
- Avoid agentic and just ask specific questions about specific pieces and paste those in where relevant.
- See if you can somehow prompt it into a better "basin of attraction" where it acts like a scicomp-knower.
- Come back later.
- Avoid agentic and just ask specific questions about specific pieces and paste those in where relevant.
- See if you can somehow prompt it into a better "basin of attraction" where it acts like a scicomp-knower.
- Come back later.
May 10, 2026 at 7:58 PM
If what I described is your use profile I would consider one of:
- Avoid agentic and just ask specific questions about specific pieces and paste those in where relevant.
- See if you can somehow prompt it into a better "basin of attraction" where it acts like a scicomp-knower.
- Come back later.
- Avoid agentic and just ask specific questions about specific pieces and paste those in where relevant.
- See if you can somehow prompt it into a better "basin of attraction" where it acts like a scicomp-knower.
- Come back later.
SciComp research group social with guests at #SIAMCSE25 in Texas: next time we should check if there are vegetarians among us. But it was very very good.
March 7, 2025 at 1:27 PM
SciComp research group social with guests at #SIAMCSE25 in Texas: next time we should check if there are vegetarians among us. But it was very very good.
j'attends avec grande impatience l'édition 2024 de Rust pour cette unique raison l'offloading GPU et la différentiation automatique 🤩
Je ne vais plus pouvoir me passer de ce langage haha !
github.com/rust-lang/ru...
Je ne vais plus pouvoir me passer de ce langage haha !
github.com/rust-lang/ru...
Expose experimental LLVM features for automatic differentiation and GPU offloading · Issue #109 · rust-lang/rust-project-goals
Metadata Owner(s) @ZuseZ4 Team(s) compiler, lang Goal document 2024h2/Rust-for-SciComp Summary Expose experimental LLVM features for automatic differentiation and GPU offloading. Tasks and status D...
github.com
December 2, 2024 at 9:06 PM
j'attends avec grande impatience l'édition 2024 de Rust pour cette unique raison l'offloading GPU et la différentiation automatique 🤩
Je ne vais plus pouvoir me passer de ce langage haha !
github.com/rust-lang/ru...
Je ne vais plus pouvoir me passer de ce langage haha !
github.com/rust-lang/ru...
It was great to host Eleni Vlachopoulou from AMD in our SciComp seminar today. Full house and interesting talk on Optimizing k-Nearest Neighbours Algorithms. Join us next week for further talks by friends from GeorgiaTech @gtresearchnews.bsky.social: scicomp.webspace.durham.ac.uk/events/semin...
Seminar Series | Scientific Computing at Durham
The Scientific Computing seminar series takes place on Fridays at 13:00. Join on zoomContact: anne.k.reinarz@durham.ac.ukUpcoming and past events #
scicomp.webspace.durham.ac.uk
February 12, 2025 at 8:28 PM
It was great to host Eleni Vlachopoulou from AMD in our SciComp seminar today. Full house and interesting talk on Optimizing k-Nearest Neighbours Algorithms. Join us next week for further talks by friends from GeorgiaTech @gtresearchnews.bsky.social: scicomp.webspace.durham.ac.uk/events/semin...
Spent evening chatting with scicomp postdocs; feel like an explorer stumbling on primitives untouched by modern tech
We're sitting here making type systems for subtle memory ownership; their most recent bug was bcs Fortran doesn't check functions are called with the right number of parameters
We're sitting here making type systems for subtle memory ownership; their most recent bug was bcs Fortran doesn't check functions are called with the right number of parameters
June 7, 2025 at 5:01 AM
Spent evening chatting with scicomp postdocs; feel like an explorer stumbling on primitives untouched by modern tech
We're sitting here making type systems for subtle memory ownership; their most recent bug was bcs Fortran doesn't check functions are called with the right number of parameters
We're sitting here making type systems for subtle memory ownership; their most recent bug was bcs Fortran doesn't check functions are called with the right number of parameters
We are searching for Expressions of Interest for Newton International Fellowships 2026: scicomp.webspace.durham.ac.uk/newton-inter...
Newton International Fellowship (NIF) 2026 - Scientific Computing at Durham
Call for Expressions of Interest: Newton International Fellowship 2026 Join the Scientific Computing (SciComp) Group at Durham University’s Department of Computer Science Are you an exceptional resea...
scicomp.webspace.durham.ac.uk
November 11, 2025 at 4:33 PM
We are searching for Expressions of Interest for Newton International Fellowships 2026: scicomp.webspace.durham.ac.uk/newton-inter...
how many people went through the passionate artist to miserable scicomp student pipeline
September 2, 2026 at 2:40 AM
how many people went through the passionate artist to miserable scicomp student pipeline
Could I get some criticism on my first real library, SciComp? https://old.reddit.com/r/csharp/comments/1o2u91g/could_i_get_some_criticism_on_my_first_real/
#csharp #dotnet
#csharp #dotnet
October 10, 2025 at 7:23 AM
Could I get some criticism on my first real library, SciComp? https://old.reddit.com/r/csharp/comments/1o2u91g/could_i_get_some_criticism_on_my_first_real/
#csharp #dotnet
#csharp #dotnet
Could I get some criticism on my first real library, SciComp? https://old.reddit.com/r/dotnet/comments/1o2u5so/could_i_get_some_criticism_on_my_first_real/
#dotnet
#dotnet
October 12, 2025 at 8:44 PM
Could I get some criticism on my first real library, SciComp? https://old.reddit.com/r/dotnet/comments/1o2u5so/could_i_get_some_criticism_on_my_first_real/
#dotnet
#dotnet
Could I get some criticism on my first real library, SciComp? https://old.reddit.com/r/dotnet/comments/1o2u5so/could_i_get_some_criticism_on_my_first_real/
#dotnet
#dotnet
October 10, 2025 at 7:24 AM
Could I get some criticism on my first real library, SciComp? https://old.reddit.com/r/dotnet/comments/1o2u5so/could_i_get_some_criticism_on_my_first_real/
#dotnet
#dotnet
Suitability for SciComp vs ML is a key debate for Posits. There are trade-offs in error propagation & precision that must be carefully weighed depending on the specific application's requirements. #NumericalAnalysis 2/6
June 20, 2025 at 11:00 PM
Suitability for SciComp vs ML is a key debate for Posits. There are trade-offs in error propagation & precision that must be carefully weighed depending on the specific application's requirements. #NumericalAnalysis 2/6
HN discusses Posit floating-point numbers as an alternative to IEEE floats. Key topics: design, trade-offs, applications (SciComp/ML), hardware challenges, and accuracy. Insights from a creator included. #FloatingPoint 1/6
June 20, 2025 at 11:00 PM
HN discusses Posit floating-point numbers as an alternative to IEEE floats. Key topics: design, trade-offs, applications (SciComp/ML), hardware challenges, and accuracy. Insights from a creator included. #FloatingPoint 1/6
Deadline 31 Dec 2025: EMS/ECMI Lánczos Prize nominations.
Recognising outstanding mathematical software with important applications.
Details: euromathsoc.org/news/reminde...
#MathSoftware #SciComp #AppliedMath
Recognising outstanding mathematical software with important applications.
Details: euromathsoc.org/news/reminde...
#MathSoftware #SciComp #AppliedMath
Reminder: Deadline 31 December 2025 for EMS/ECMI Lánczos Prize nominations
Markus Juvonen
euromathsoc.org
December 19, 2025 at 4:35 PM
Deadline 31 Dec 2025: EMS/ECMI Lánczos Prize nominations.
Recognising outstanding mathematical software with important applications.
Details: euromathsoc.org/news/reminde...
#MathSoftware #SciComp #AppliedMath
Recognising outstanding mathematical software with important applications.
Details: euromathsoc.org/news/reminde...
#MathSoftware #SciComp #AppliedMath
Great talk by Thomas Flynn in our @durham-comp-sci.bsky.social SciComp seminar on his work within the @shareing.bsky.social grant.
February 4, 2026 at 4:09 PM
Great talk by Thomas Flynn in our @durham-comp-sci.bsky.social SciComp seminar on his work within the @shareing.bsky.social grant.
It was great to have Prakash Murali from @cst.cam.ac.uk in our SciComp seminar today. Interesting ideas on full stack design in quantum computing.
April 30, 2025 at 3:53 PM
It was great to have Prakash Murali from @cst.cam.ac.uk in our SciComp seminar today. Interesting ideas on full stack design in quantum computing.
Thanks to Gokberk Kabacaoglu's initiative, we now finally have our book impact section up: scicomp.webspace.durham.ac.uk/scicomp-cont... One of the books originates from our module Parallel Scientific Computing in level 3, one from @miscada.bsky.social's professional skills module, and one from AI.
SciComp Contributes to Three New Springer Publications - Scientific Computing at Durham
SciComp Contributes to Three New Springer Publications SciComp is proud to announce its contribution to three new books published by Springer, each inspired by the MISCADA initiative. The publications...
scicomp.webspace.durham.ac.uk
April 3, 2025 at 12:20 PM
Thanks to Gokberk Kabacaoglu's initiative, we now finally have our book impact section up: scicomp.webspace.durham.ac.uk/scicomp-cont... One of the books originates from our module Parallel Scientific Computing in level 3, one from @miscada.bsky.social's professional skills module, and one from AI.