simonbyrne

7 ms·

1.
by simonbyrne 5y ago
If that's seriously what you're looking for, you should really consider Fortran. There is a reason it is still very widely used in scientific domains.
2.
by simonbyrne 5y ago
You can just wrap the literal in a conversion function, eg Float32(0x1p52), which should get constant propagated at compile time.
3.
by simonbyrne 5y ago
That's fascinating thread, thanks: https://github.com/WebAssembly/design/issues/148
4.
by simonbyrne 5y ago
Do you know what happens when you have ops with different flags? e.g. if you have (a + b) + c, where one + allows reassoc but one doesn't?
5.
by simonbyrne 5y ago
From personal experience, yes: I've seen multiple cases of scientists finding the ultimate cause of their bugs was some fast-math-related optimization. The problem isn't necessarily the code they wrote themselves: it is often that
6.
by simonbyrne 5y ago
I tried to lay out a reasonable path: incrementally test accuracy and performance, and only enable the necessary optimizations to get the desired performance. Good tests will catch the obvious catastrophic cases, but some will inevitably be
7.
by simonbyrne 5y ago
I'm not exactly sure what you're asking here, but the point is that "to machine precision" is relative: if f(x) and g(x) are O(1e200), then the absolute error of each is still O(1e185). However f(x)/g(x) will still
8.
by simonbyrne 5y ago
The key thing about floating point is that it maintains relative accuracy: in your case, if you have say f(x) and g(x) are both O(1e200), and are correct to some small relative tolerance, say 1e-10 (that is, the absolute error is 1e190). Th
9.
by simonbyrne 5y ago
Herbie is a great tool, especially for teaching.
10.
by simonbyrne 5y ago
In theory, every function should do that to check things like rounding mode etc. But that would be pretty slow, especially for low-latency operations (modifying mxcsr will disrupt pipelining for example).
11.
by simonbyrne 5y ago
-ffp-contract=fast will enable FMA contraction, i.e. replacing a * b + c with fma(a,b,c). This is generally okay, but there are a few cases where it can cause problems: the canonical example is computing an expression of the form: a * d -
12.
by simonbyrne 5y ago
My point isn't that fast-math isn't useful: it very much is. The problem is that it is a whole grab bag of things that can do very dangerous things. Rather than using a sledgehammer, you should try to be selective and enable only
13.
by simonbyrne 5y ago
Not necessarily: if your cospi(x) function is always returning 1.0 ( https://github.com/JuliaLang/julia/issues/30073#issuecomment... ), but you wrote your code assuming the result was in a different interval, t
14.
Beware of fast-math(simonbyrne.github.io)
189 points by simonbyrne 5y ago|107 comments
15.
by simonbyrne 5y ago
> But then people latched on to the idea that it was caused by some kind of sonic device or hitherto unknown weapon targeting embassies. Possibly prompted by historical precedent: https://en.wikipedia.org/wiki/Moscow
16.
by simonbyrne 6y ago
One of the interesting things about the 80-bit format was that allowed the use of 64-bit integers on a 16 or 32-bit machine. This was what Thomas Nicely was using them for when he found the Pentium FDIV bug: https://faculty.lynch
17.
by simonbyrne 7y ago
> or wait for optimizations to be done, e.g have pandas, numpy etc handle multi-core processors etc ? All those exist already. Indeed, other than DataFrames.jl (the pandas equivalent) they are part of the language itself.
18.
by simonbyrne 7y ago
I disagree about practical and accessible: - its exposition is complicated (trying to prove everything in a general base makes it difficult to understand) - it's woefully out of date (lack of guard digits haven't been an issue for
19.
by simonbyrne 7y ago
Deep learning is the obvious answer: 10 years ago it was largely confined to a small number of research groups and niche conferences.
20.
by simonbyrne 7y ago
Thank you! The Goldberg article is a terrible way to learn about floating point, and the frequency with which it is referred to on StackOverflow is really disheartening.
21.
by simonbyrne 7y ago
I seem to recall that part of the DoE's selection criteria is to ensure a competitive market for future contracts (since they will always be buying more supercomputers). Given the recent dominance of Nvidia in the market (see https:&#
22.
by simonbyrne 7y ago
Additionally the language and penalties for fare evasion often don't match that of say parking tickets, which I would put as roughly similar in terms of moral violations (i.e. taking advantage of a common good, but not endangering the
23.
by simonbyrne 7y ago
Not to mention a Cray 1 used 115kW: the new raspberry pi uses 7W peak.
24.
by simonbyrne 7y ago
It's difficult to examine his comments without historical context: I haven't dealt with any code from that era, but there is still plenty of Fortran 77 code from ~20 years later where the use of gotos makes it difficult to underst
25.
by simonbyrne 7y ago
Am I missing something? Other than the doc generator (which is in Python), none of the repos have been touched for a year or so.
26.
by simonbyrne 7y ago
Also, reminded me of the old CB1 cafe, which labelled itself "UK's oldest internet cafe", which amusingly hadn't updated their website since 2005: http://www.cb1.com/
27.
by simonbyrne 7y ago
I used to live around the corner: I always wondered about that. I do miss Norfolk Street Bakery though (which I hope is still going).
28.
ReviewNB: Code Reviews for Jupyter Notebooks(reviewnb.com)
2 points by simonbyrne 7y ago|0 comments
29.
by simonbyrne 8y ago
Interesting: I've heard anecdotes about similar experiences in southern California, where the "good" school districts often don't allow apartments or smaller subdivided buildings (duplexes/triplexes) to be built.
30.
by simonbyrne 8y ago
Obviously that question depends on your motivation, but one good reason is that it is a great way to learn how software actually works. R makes it very easy to see the underlying R code (you just type the function name), until you get to a

More ›