Ph.D. Student at @UoC | past intern @oraclelabs | Graduate Research Assistant @CarvForth Research on LSM-trees, always interested in Storage/OS research!
Prof. Donald Knuth opened his new paper with "Shock! Shock!"
Claude Opus 4.6 had just solved an open problem he'd been working on for weeks — a graph decomposition conjecture from The Art of Computer Programming.
He named the paper "Claude's Cycles."
31 explorations. ~1 hour. Knuth read the output, wrote the formal proof, and closed with: "It seems I'll have to revise my opinions about generative AI one of these days."
The man who wrote the bible of computer science just said that. In a paper named after an AI.
Paper: https://t.co/juSOmK9vOt
Did you know that SSDs use quantum tunnelling to store data? Or that we have to completely rewrite RAM every 30ms to prevent data from just dissipating?
If you did, then you can probably skip this chapter on how computers store data:
https://t.co/d7m0TLXAKv
Latency Profiling and Optimization talk video:
https://t.co/dNN0825qmX
Includes overview of new perf features.
Tagline: Profilers lied to us all that time! ☺
BTW if anyone wants a good intro to database storage / Log structured storage (aka LSM trees), the @CMUDB lecture this fall is a good one: https://t.co/ln6Yi0K1JZ
I was going through the mmap/buffer-cache discussion involving Andy Pavlo and Marc Brooker on my TL and thought of taking a look at some of the details on how LRU based eviction works in PostgreSQL and MySQL InnoDB buffer pools ..
I was aware that they implement different algorithms and none of them were "vanilla" LRU ones - they are customised for better concurrency and scalability. And points to some interesting data structure techniques as well ..
Here's a short thread🧵.. (0/3)
Were you looking for benchmarks for your shell research? Look no further! Our 🐨 Koala 🐨 benchmark suite is out (https://t.co/YUPmN884qW) and it just got a best paper award at the last ATC ever! Check it out :)
On our latest ParadeDB blog post about the LSM tree we built in Postgres, many asked me how we map it down to Postgres pages. We also wrote a blog post about that, which you can find here: https://t.co/qC2st89NBb
Last week I had the privilege to give a keynote at the International Conference on Performance Evaluation (ICPE'25). It wasn't recorded, so I captured the key points as a blog post.
Curious about the *fastest* distributed OLTP?
Think strongest consistency and high availability with 100s of M. txs/sec (even on just a few nodes).
Check Dandelion, our #VLDB’25 paper.
Profiling: you are doing it all WRONG! The SHOCKING truth about CPU vs. real time
Now that I get your attention :)
My perf latency profiling changes merged:
https://t.co/D5713CXKEw
First profiler ever that samples realtime (not CPU time) and suitable for latency optimization
Use the -D and -S options of objdump on an elf with debug symbols to generate listing file with disassembly intermixed with the C Source code.
Here are the commands I used in the video -
$ gcc -g main.c # compile the code
$ objdump -D -S a.out > main.lst
I use these regularly in my projects to be able to look at the assembly for the C source and take a through look at the placement of various symbols.
----
If you are interested in learning C Language by diving into the System level mechanics, take a look at the course I am creating. Here - https://t.co/sgoYLzT5AW