Sam Arch 🇦🇺

🧵Congrats to the #VLDB2025 award winners at @VLDBconf (London)! 🏆 • Best research papers: “Diva” (@niv_dayan @UofT + @KTHuniversity) & “AnyBlox” (@JanaGiceva @tum_db) • Runner-up: “The Key to Effective UDF Optimization” by @CMUDB @andy_pavlo @pateljm

The latest paper from the #1 CMU-DB PhD student @SamArchDB's is wild compilation DB magic! He automatically makes UDFs run 300x faster on @SQLServer and 1.3x faster on @duckdb. Code: github.com/SamArch27/PRISM Paper: vldb.org/pvldb/vol18/p1…

Vol:18 No:1 → The Key to Effective UDF Optimization: Before Inlining, First Perform Outlining vldb.org/pvldb/vol18/p1…

youtube.com/watch?v=XMIEkn… by @SamArchDB

@erikdarlingdata @wiscdb @GraySystemsLab @uw_db Hi Erik, VLDB still needs to post the paper on its website. In the meantime, you can check out a recording of my talk for NYC Systems. youtube.com/watch?v=XMIEkn…

@SamArchDB @wiscdb @GraySystemsLab @uw_db Link?

The UDF world tour continues. This week, I'm stopping at UW Madison, Microsoft's Gray Systems Lab, and The University of Washington. See you all there. 10/24 @wiscdb 10/25 @GraySystemsLab 10/25 @uw_db

I am flying to the Bay Area to give talks about my new VLDB 2025 paper on UDFs with @andy_pavlo and @pateljm. With our new technique (UDF outlining), queries run up to 1000× faster than FROID. The paper will drop soon. 9/10 @databricks 9/11  @UCBerkeley 9/12 #HTAPSummit2024

This speaks to me: it’s the essence of building hard things that take time to have impact, but then matter in a big way - you have to love the process, not just the outcome!

VLDB'24 Paper #1: Collecting training data for ML models with DBs is $$$/slow. @wanshenl's Boot framework uses @PostgreSQL extensions to cutoff redundant queries. Offline training goes from weeks to hours! • Code: github.com/lmwnshn/boot • Paper: x.com/pvldb/status/1…

Video of @SamArchDB 's talk from NYC Systems August 2024 is now up! Dear UDFs, I Broke Up With You, But Now I'm Ready To Give You A Second Chance. Will You Take Me Back? Sincerely, SQL youtube.com/watch?v=XMIEkn…

🇦🇺s incoming

@_TylerHillery We have three more bombs dropping this month from @lmwnshn + @SamArchDB + William Zhang. @abigale_kim just submitted her VLDB paper last week too. It's going to be a blow out year.

It took three years to finish, but our follow-up to the 2006 "What Goes Around Comes Around" is finally out! Stonebraker and I examine the last 20 years in databases and discuss why relational databases + SQL will continue to remain on top. 📄PDF: db.cs.cmu.edu/papers/2024/wh…

We are proud to release the first major version of DuckDB, v1.0.0, codenamed "Snow Duck". This version is a culmination of almost six years of research and development. Today we are shipping an innovative database system with a backwards-compatible storage format. Check out our announcement blog post: duckdb.org/2024/06/03/ann…

Here is a short video from my recent interview with @SCSatCMU's media team on the one year anniversary of the infamous @Jeopardy "matrix" scandal: instagram.com/reel/C6Y5SCPrF…

I am excited to announce that I will start my PhD in database systems at @wiscdb and work with @xiangyao_yu beginning Fall 2024! I'm super grateful to everyone who has supported me along the way :)

My #1 PhD student @butro successfully completed his PhD defense. Thanks to the committee (@pateljm @justinesherry @samrmadden). Matt's thesis is on accelerating databases with eBPF (telemetry, proxies, OLTP stores). You have 60 days to hire him. Expect fierce competition.