tanelpoder.bsky.social 🇺🇦
@TanelPoderCreator of https://t.co/w1Pz4s2EhV and a long time computer performance geek. Blog: https://t.co/IOSFP4GmxN
Similar User
@OracleAskTOM
@JLOracle
@FranckPachot
@connor_mc_d
@SQLMaria
@fritshoogland
@MartinDBA
@oraclebase
@DBoriented
@ludodba
@KnownAsMarkus
@ChrisRSaxon
@POUG_ORG
@thatjeffsmith
@petefinnigan
New blog: Testing the Silk Platform in Google Cloud: Achieving 20 GiB/s I/O Throughput in a Single Cloud VM tanelpoder.com/posts/testing-… Impressive results and the testing was lots of fun!
📢 It's here! The second edition of my book "Performance Analysis and Tuning on Modern CPUs" is out NOW! Dive into low-level optimizations and learn to write fast code like a pro. Grab your copy: amzn.to/4fsroBs Please RT to spread the word! #NewBook #BookRelease 🧵1/3
So LLM is hitting the wall and everyone love agents. What’s next? Knowledge encoding, multi step Agents Verification, RAG eval, agenetic a/b testing.Can’t wait to share what we are cooking :-) 🌳 - there is so much work to do. And we are hiring our first AI Agent engineer to…
I have a presentation about #oracle #Exadata which ends with "35GB/s is a bit tricky in the cloud, & quite niche" Read this blog post by @TanelPoder It's excellent & perfect for non-Exa databases (e.g. PG) needing 20GB/s Who needs 35GB/s for their database? A few I know...
This is a screenshot of a single VM running on Google Cloud and driving a phenomenal >20GB/sec of data throughput 🏎️, courtesy of a blog post by my friend @TanelPoder . For those who know, that number is only half the story - the other half is the 0.42ms latency 🤯
New blog: Testing the Silk Platform in Google Cloud: Achieving 20 GiB/s I/O Throughput in a Single Cloud VM tanelpoder.com/posts/testing-… Impressive results and the testing was lots of fun!
The FlameGraph technique is a great way to visualize metrics in nested hierarchies. @P99CONF speaker @tanelpoder shows how it can be used to measure the response time used by individual execution plan operators. ow.ly/Nm3K50TJEtX #ScyllaDB #P99CONF
Postgres pgvector 0.8.0 released! Iterative (vector) index scans that can be combined on-the-fly with any other WHERE clause filters is a major improvement: github.com/pgvector/pgvec…
this pytorch flame graph is wild, wonder what the @__tinygrad__ equivalent would look like
"Today, we’re pleased to announce that Trillium, our sixth-generation TPU, is now available to Google Cloud customers in preview. "cloud.google.com/blog/products/… "Today, we’re thrilled to announce the ... the first Axion-based VM series ..." cloud.google.com/blog/products/…
If you put a ThreadLocal reference, or any other object that strongly reaches it, into that ThreadLocal itself, it is a guaranteed memory leak in current OpenJDK implementation, at least until the thread exits itself. (Don't ask how we know.)
The presentation covers: 1. Where profiling sits relative to metrics and traces. 2. Why CPU profiling and Flame Graphs are not enough. 3. How to implement Off-CPU profiling and visualize the results. 4. Integration with distributed tracing. Enjoy! youtube.com/watch?v=53UIPZ…
Profiling has traditionally been a tool for analyzing CPU and memory usage issues. However, next-gen tools now enable profiling to address latency problems through deep integration with distributed tracing. Check out my colleague @richardstartin's talk at @P99CONF to learn more!
TIL your BlueSky profile name tanelpoder.bsky.social also works as an URL! I'll hang out there from now on, looks like there's a recent mass movement of data people to it, so plenty of interesting activity. See you there!
Did you catch our #eBPF sessions from @lizrice, @tanelpoder, and @jrfernandez at #P99CONF? If you missed them or need a refresher, check out the overviews, bonus readings, and recordings on demand. ow.ly/yvjU50TVpR6 #ScyllaDB #DevOps #techtips #optimization
"In fact, experienced engineers are being encouraged to train generative AI in a way they were never encouraged to invest time in training juniors." - @trisha_gee oreilly.com/radar/rift-bet…
My talk starting in 3 minutes!
United States Trends
- 1. Tyson 515 B posts
- 2. Karoline Leavitt 19,7 B posts
- 3. #wompwomp 7.062 posts
- 4. Paige 8.069 posts
- 5. Syracuse 22,8 B posts
- 6. Pence 62,6 B posts
- 7. Kiyan 27,3 B posts
- 8. Kash 109 B posts
- 9. The FBI 277 B posts
- 10. Frankie Collins N/A
- 11. Debbie 36,2 B posts
- 12. Jarry N/A
- 13. Whoopi 119 B posts
- 14. #TOKKIVSWORLD 1.922 posts
- 15. #LetsBONK 16,8 B posts
- 16. Dora 24,9 B posts
- 17. White House Press Secretary 22,8 B posts
- 18. Villanova 1.880 posts
- 19. Ace Bailey N/A
- 20. End of 1 23,9 B posts
Who to follow
-
Ask TOM for Database/SQL
@OracleAskTOM -
Jonathan Lewis
@JLOracle -
Franck Pachot
@FranckPachot -
Connor loves SQL , 🍸and ☕️
@connor_mc_d -
Maria Colgan
@SQLMaria -
Frits Hoogland
@fritshoogland -
Martin Bach
@MartinDBA -
Tim Hall ∞🐵+∞⌨️
@oraclebase -
Oren Nakdimon
@DBoriented -
Ludovico Caldara
@ludodba -
Markus Michalewicz
@KnownAsMarkus -
Chris Saxon
@ChrisRSaxon -
ORA-600
@POUG_ORG -
SQL.ProductManager.Oracle.Databases
@thatjeffsmith -
Pete Finnigan
@petefinnigan
Something went wrong.
Something went wrong.