fuy

5 ms·

1.
by fuy 29d ago
yes, if you control the clients and do client-side pooling.
2.
Modern GPU Programming Book(mlc.ai)
2 points by fuy 3mo ago|1 comments
3.
by fuy 3mo ago
The book is a companion to Machine Learning Systems https://mlsyscourse.org/ CMU course.
4.
by fuy 5mo ago
There's been many examples of societies where killing or abusing people was legal etc. Law is not math, it can be (and often is) wrong; in many cases a law is just a way for ruling class to make money/keep power etc. It's com
5.
by fuy 6mo ago
claude code easily uses 10+GB in single session :) 1Gb sounds very efficient by comparison
6.
by fuy 6mo ago
almost all Spotify playlists/radio features are extremely stupid. Radios all repeat the same 2-3 songs per artist, doesn't matter where you start. Classical pieces in radios are always 1-2 minutes long etc etc. It was like before
7.
by fuy 7mo ago
sorry about that, I appreciate the reminder. Sometimes it's hard to engage in conversations with obvious bots/bad actors in good faith.
8.
by fuy 8mo ago
CLUSTER command is not the same as index organized tables, it's a one-time "physical sort" operation. New data is not organized until you run CLUSTER again. Index organized tables are maintained automatically by Oracle/S
9.
by fuy 8mo ago
Not saying you're wrong, but solo developers building (relatively) greenfield projects are the best bet for increased AI productivity. Solo dev projects are usually reasonably sized (< million LOC), style is more uniform, there&#x27
10.
by fuy 9mo ago
isn't this template supposed to mean that all the previous considerations are now obsolete?
11.
by fuy 10mo ago
It's web in a (limited) sense that there's probably a web frontend somewhere, but this "somewhere" is usually pretty far away from where most of the code is developed. Most of the backend logic is not related to serving
12.
by fuy 1y ago
Looking at first optimization, I wonder if double-checking after acquiring exclusive lock brings any performance benefits. The whole premise is that cache access is read-heavy, so not acquiring exclusive locks for reads eliminates by far th
13.
by fuy 1y ago
perf and look at stack traces (or off-cpu events for waits/locks). also, ebpf
14.
by fuy 1y ago
Hey Tanel - I wanted to thank you for that blog post and psn tool - it recently helped me in a tricky performance investigation.
15.
by fuy 1y ago
Had similar situation a few years before - switched a (now) billion revenue product from Read Committed to Read Committed Snapshot with huge improvements in performance. One thing to be aware when doing this - it will break all code that re
16.
by fuy 1y ago
isolation levels, that is!
17.
by fuy 2y ago
Top 5% overall or top 5% of dev jobs? Former sure, latter might be not? I live in Prague, and there are definitely multiple companies that pay more than 90k to middle positions. especially if it's before taxes.
18.
by fuy 2y ago
what about npgsql? it's pretty much official and is developed by Microsoft people
19.
by fuy 2y ago
And also these old C hands don't seem to get paid (significantly) more than a regular web-dev who doesn't care about hardware, memory, performance etc. Go figure.
20.
by fuy 2y ago
Thomas Neumann and Alfons Kemper papers can hardly be called "home-grown", their stuff has been implemented in multiple industrial systems. Postgres optimizer, though, is very bad even with simple correlated subqueries, let alone
21.
by fuy 2y ago
I love and use Postgres daily for many years, but: Performance monitoring is pretty much absent, all you have is pg_stat_statements and friends. So for any serious scale you need 3d party solution (or you're writing your own) straight
22.
by fuy 2y ago
Are you saying medieval Germany was richer than present-day Germany or USA? That's of course not true. And it's also not clear why you call it a survivorship bias - these villages were probably above-the-median in their days, but
23.
by fuy 3y ago
Aurora is using native Postgres planner, I believe, probably with some minor enhancements.
24.
by fuy 3y ago
One other thing about JIT that I feel is pretty crazy is that the generated code is not cached. I mean it's the most expensive part of the query execution a lot of the time, how come it's not cached? I couldn't find good reas
25.
by fuy 3y ago
Or it could work in a way that the Planner has access to data about previous runs of each query, and it can use this data to change plans that were proven bad during execution. This way, the first execution would be slow, but Planner could
26.
by fuy 3y ago
+. And then you have to figure everything one more time when you decide to move to (or to add support for) Azure/GCP.
27.
by fuy 3y ago
and doesn't scale well. Meaning using it can take your Postgres server down way before regular SQL workload would.
28.
by fuy 3y ago
Hey Franck, just wanted to say I appreciate your database writings. I read a whole bunch of them over the years, and always found them interesting and educational.
29.
by fuy 3y ago
The common sentiment on HN I'd say is the opposite: ORM's are needed for junior devs who didn't have time to learn SQL properly. After NoSQL hype died out (around 5 years ago), the HN attitude towards both ORM's and usin
30.
by fuy 3y ago
a bit off topic, but do you know of any good resources on tuning battery life and CPU fan control for system76 laptop?

More ›