twa927

8 ms·

31.
by twa927 7y ago
It works well for tight loops processing much data, or heavy object-orientation (multiple levels of class hierarchies). It probably won't work well for regular Django webapps or scripts. Also, real-world Python numerical/AI code u
32.
by twa927 7y ago
CPython is mostly reference-counted.
33.
by twa927 7y ago
I meant the inside of a VM (its implementation), which is coded in C/RPython. Java's VM is coded in C++, but I don't think C++ is used for any regular library functions, while C is used heavily for Python's stdlib.
34.
by twa927 7y ago
I started doing it but yes, it's too much effort to get two full benchmarks, I'm sorry :). But I think it went down to the inefficiency of String.split(): https://stackoverflow.com/questions/37007189/stri
35.
by twa927 7y ago
I don't have access to this codebase now but I'll try to write some benchmark.
36.
by twa927 7y ago
I was trying to speed up a log processing service running on PyPy by rewriting it in Java. I was surprised that the result was about twice slower (I know Java quite well and I didn't see obvious optimizations; most of the time was spen
37.
by twa927 7y ago
> The only reason for a KV cache is if you're running a slow language like Ruby, PHP, Python, where the memory overhead for objects is large and the speed of a remote cache is roughly the same as one in memory. If you find yourself
38.
by twa927 7y ago
Maybe it's time to start hosting olympics in the same place? And make all countries participate in costs? There are multiple arguments for this: https://www.google.com/search?q=hosting+olympics+in+the+same...
39.
by twa927 7y ago
That's interesting, I'm wondering if this happens also to e.g. "Cloud Specialists" or happened to "Object Orientation Specialists" 20 years ago...
40.
by twa927 7y ago
Yes, I run it several times, but each time the effect wasn't noticeably better. And I would prefer to have an option for an explicit configuration of the rendering properties (like "hinting: slight/medium/none") ins
41.
by twa927 7y ago
> Also, I saw this in the news a while ago: https://www.phoronix.com/scan.php?page=news_item&px=Fedora-C... . Hmm, that's interesting, so this is the same ClearType algorithm as the one currently used in Windows
42.
by twa927 7y ago
> "Previously, those were kind of off the books, under the table. You know, 'Don't ask don't tell deployments.' "But now suddenly, it's the AI team and they've got to be supported." Are AI dev
43.
by twa927 7y ago
Can someone say if the font rendering is improved compared to previous releases? This might sound like nitpicking but it's the thing that kept me from switching from Ubuntu, which generally has superb font rendering.
44.
Ask HN: Better GitHub issue/comment search
1 points by twa927 7y ago|0 comments
45.
by twa927 7y ago
Ok, so this depends on what algorithms are used for the feature detection ("edges in an image, and an eye detector, and a texture analyser for fur"). I'm guessing hand-coding an algorithm for detecting edges in an image can b
46.
by twa927 7y ago
Hmm, I think there's no "machine learning" here. There's a human hard-coding some thought process, using mostly some simple statistics/thresholds to e.g. define what a "fur texture" looks like.
47.
by twa927 7y ago
I think expert systems consist of a "rule engine" where rules can be added dynamically?
48.
by twa927 7y ago
> Until about 2013, If you wanted to make a software system that could, say, recognise a cat in a photo, you would write logical steps. You’d make something that looked for edges in an image, and an eye detector, and a texture analyser f
49.
by twa927 8y ago
And it looks like TokuDB is no longer developed?
50.
by twa927 8y ago
> Memory is a difficult one, especially in garbage-collected languages which have a habit of filling up the heap even when it's not used, so it's not always obvious how much memory is actually being used without having language
51.
by twa927 8y ago
Ok thanks. I'm looking for a way to run a service with memory peaks - 50pct is 200MB and 99pct is 2GB. It looks like no current serverless solution would handle it without overpaying by allocating 2GB for all executions.
52.
by twa927 8y ago
Does max. RAM usage need to be specified upfront, or what's billed is the live-allocated amount?
53.
by twa927 8y ago
Does max. RAM usage need to be specified upfront, or what's billed is the live-allocated amount? This is important for scenarios like running Chrome Headless when the RAM usage varies a lot between invocations. AFAIR AWS Fargate requir
54.
by twa927 8y ago
Yep, you're describing a typical design of a DB. I'm looking for an alternative design. I'm guessing it could use multiple small files (a few MBs) and merge them to avoid fragmentation.
55.
by twa927 8y ago
> Shouldn't plain vacuum be enough? The space is going to be reused for new items anyway, it won't take more of the space not already reserved by the database? In theory, yes. In practice, I've seen a lot of bloat being pr
56.
by twa927 8y ago
Regular VACUUM only marks space as available for reuse. VACUUM FULL returns space to the OS but it requires an exclusive lock and a multi-hour run for a large table.
57.
Ask HN: SQL engine reclaiming space from DELETEs
12 points by twa927 8y ago|18 comments
58.
by twa927 10y ago
Isn't OpenStack just an open source project? Ie. it doesn't sell anything.
59.
by twa927 10y ago
Thanks for the example. I expected a higher price for the self-hosted version: it's only 3 months worth of the highest saas plan.
60.
Ask HN: SaaS pricing vs. self-hosted pricing
2 points by twa927 10y ago|8 comments

More ›