iiinova — where your digital self keeps evolving

Giulia Guidi @giulsguidi.bsky.social · 30/09/2026

ICS is launching its 1st Junior Program Committee! 🔥 Early-career researchers new to the ICS PC can gain hands-on #HPC and #CompArch reviewing experience, guided by senior researchers. Please nominate yourself by Oct 15, 2026 👇 docs.google.com/forms/d/e/1F... #ICS27 #SIGARCH @sighpc.bsky.social

010

Giulia Guidi @giulsguidi.bsky.social · 30/09/2026

I’ll see you in Atlanta next June for #FCRC! I’ve got food recs!

001

Giulia Guidi @giulsguidi.bsky.social · 30/09/2026

I'm looking forward to seeing the great work our communities are doing. Please consider submitting, and help spread the word to your colleagues and students! #ICS27

100

Giulia Guidi @giulsguidi.bsky.social · 21/09/2026

Yifan Li is going to present our recent work titled "Sparse Linear Algebra Accelerates Genotype Representation Graph Computation at Biobank Scale" at the #Sys4Health Workshop at #SOSP 2026! Preprint: www.biorxiv.org/content/10.6... LinkedIn post: lnkd.in/dahXztbu Code: github.com/CornellHPC/M...

Rob Patro @robp.bsky.social · 19/09/2026

Reposted by Giulia Guidi

That magical feeling when you’ve submitted a @biorxivpreprint.bsky.social preprint and you’re awaiting screening so you can share the work with everyone. BioRxiv preprint screening is like “stochastic Christmas morning” 🤣.

1151

Giulia Guidi @giulsguidi.bsky.social · 19/09/2026

Nina going at her brother 😅

000

Giulia Guidi @giulsguidi.bsky.social · 17/09/2026

I'm really looking forward to visiting @inf.usi.ch in less than a month! I'll be talking about our recent work Mikado: www.biorxiv.org/content/10.6... #HPC #SparseLinearAlgebra #Genomics #PopulationGenetics #GenotypeRepresentationGraph

000

Reposted by Giulia Guidi

arXiv cs.DC Distributed, Parallel, and Cluster Computing @csdc-bot.bsky.social · 24/03/2026

Julian Bellavita, Lorenzo Pichetti, Thomas Pasquali, Flavio Vella, Giulia Guidi: Communication-Avoiding SpGEMM via Trident Partitioning on Hierarchical GPU Interconnects arxiv.org/abs/2603.21444 arxiv.org/pdf/2603.21444 arxiv.org/html/2603.21444

arXiv cs.DC Distributed, Parallel, and Cluster Computing @csdc-bot.bsky.social · 22/04/2026

Reposted by Giulia Guidi

Yifan Li, Giulia Guidi: Ocean: Fast Estimation-Based Sparse General Matrix-Matrix Multiplication on GPU arxiv.org/abs/2604.19004 arxiv.org/pdf/2604.19004 arxiv.org/html/2604.19004

arxiv cs.DS @arxiv-cs-ds.bsky.social · 28/08/2026

Reposted by Giulia Guidi

Aaron Li, Yifan Li, Drew DeHaas, Giulia Guidi Cheaper by the Batch: Shared Traversal for Genotype Graph Editing arxiv.org/abs/2608.26488

Rob Patro @robp.bsky.social · 16/09/2026

Books from the 1400 at the Biblioteca Malatestiana! 😱

000

Reposted by Giulia Guidi

Look at Giulia, continuing to kick ass and take names (and she's almost out of names)! Congrats!

121

You’re too kind! Thank you!

000

In addition, many of my students will be at #SC26 as well, presenting their work in various formats. More on this soon.

000

GeneticsPapers @asimresearcher.bsky.social · 14/09/2026

I look forward to attending #SC26 in Chicago in November, where I’ll also have the honor of standing on the main stage for an invited talk in a session focused on the growing role of advanced computing and artificial intelligence in biomedical research.

100

Reposted by Giulia Guidi

Sparse Linear Algebra Accelerates Genotype Representation Graph Computation at Biobank Scale www.biorxiv.org/content/10.64898/20…

April Wei @aprilwei.bsky.social · 14/09/2026

Reposted by Giulia Guidi

New preprint from our collaborators @giulsguidi.bsky.social, and her PhD students Yifan Li and Qingyao Sun, which enabled GRG-based statistical genetics across different computer architecture🔥. So the practical win for GRG users: you can pick the matvec backend to best leverage your machine &task.

biorxiv.org

Sparse Linear Algebra Accelerates Genotype Representation Graph Computation at Biobank Scale

Biobank-scale genomic analyses are increasingly constrained by computational costs, as hundreds of thousands to millions of samples and variants must be analyzed together. The genotype representation ...

164

Reposted by Giulia Guidi

bioRxivpreprint @biorxivpreprint.bsky.social · 12/09/2026

Sparse Linear Algebra Accelerates Genotype Representation Graph Computation at Biobank Scale www.biorxiv.org/content/10.64898/20…

021

bioRxiv Genomics @biorxiv-genomic.bsky.social · 12/09/2026

This project taught me a lot about the challenges of working with sensitive genomics data on cloud infrastructure, and I now have some ideas for how DOE and NIH could work together to make HPC resources more accessible for biobank analyses. The scale of these data is not getting any smaller!

000

Reposted by Giulia Guidi

Sparse Linear Algebra Accelerates Genotype Representation Graph Computation at Biobank Scale www.biorxiv.org/content/10.64898/20…

097

Reposted by Giulia Guidi

AJ Lauer @ayjaylauer.bsky.social · 17/11/2025

Wise words from @giulsguidi.bsky.social while talking about her career story at the #SC25 @womeninhpc.org workshop: "Your path keeps evolving as you meet the right people" and "if in doubt, yet a dog. Or two." ❤️🦮

Giulia Guidi speaking at the Women in HPC workshop at SC25.

062

April Wei @aprilwei.bsky.social · 12/04/2026

Ok, that wasn't the right png, but this one is!

000

Reposted by Giulia Guidi

Very proud to share our new work on General, orders-of-magnitude faster whole-genome analysis with genotype representation graphs (GRG). We topped ourselves in this one 🚀 and made GRG a practical foundation for biobank-scale population and statistical genetics. www.biorxiv.org/content/10.6...

biorxiv.org

24618

Great work by Yifan Li, Qingyao Sun, Drew DeHaas, Max Xiaohang Zhao, Adam Boyko, Shaila Musharoff, and @aprilwei.bsky.social! #HPC #SparseLinearAlgebra #ComputationalBiology #Genomics #BiobankScaleGenomics #GWAS

000

If you work in population genetics, population statistics, or with biobank data and your analyses are bottlenecked by matrix–vector multiplies, Mikado may help. I'd be happy to chat more!

110

Preprint: biorxiv.org/content/10.6... Code: github.com/CornellHPC/M... The name Mikado comes from the pick-up-sticks game: a stick can be lifted only once nothing rests on it, just as a block is dispatched only once its dependencies resolve!

200

This caps off a year of wonderful collaboration with April Wei and Shaila Musharoff (plus a few months' delay courtesy of the UK Biobank close call—IYKYK).

100

Mikado achieves up to 470x on the core primitive and orders-of-magnitude end-to-end speedup for PCA and BOLT-LMM at biobank scale, including All of Us.

200

That exposes the parallelism GPUs need while preserving GRG compression, using only portable sparse primitives and no custom kernel, so it stays efficient as hardware evolves.

100

Under a reverse topological ordering, the GRG adjacency matrix is strictly lower-triangular. So the genotype matrix-vector product (the primitive behind PCA, BOLT-LMM, and heritability estimation) becomes a pipelined sequence of blocked sparse matrix-vector multiplies.

110

Our preprint, "Sparse Linear Algebra Accelerates Genotype Representation Graph Computation at Biobank Scale," introduces Mikado, a method that reformulates genotype representation graph (GRG) traversal as a sparse triangular solve. I'm thrilled it's finally out!

210