jmuhlich

Try this: https://alexsaveau.dev/blog/projects/performance/files/fuc/f...

> The key insight is that file operations in separate directories don’t (for the most part) interfere with each other, enabling parallel execution.

It really is magically fast.

EDIT: Sorry, that tool is only for local copies. I just remembered you're doing remote copies. Still worth keeping in mind.

CPython still needs to update reference counts after the fork even if your python code itself doesn't modify any data, making Copy on Write less helpful than it would first appear. Monolithic memory buffers like numpy arrays can be shared efficiently via CoW, but rich data structures with many PyObjects suffer from the refcounting issue.

The systems of differential equations produced by these types of models tend to be analytically intractable, so "numerical methods" are absolutely essential here. The phrase you want to look for is "initial value problem" (IVP): you know the initial values of the quantities described by the equations and you want to find out how those quantities evolve over time.

https://en.wikipedia.org/wiki/Numerical_methods_for_ordinary...

I like pdbpp. Make sure to install from source as there hasn’t been a release in a while.

https://github.com/pdbpp/pdbpp

To make this more clear as I have had the same experience and want more people to enjoy it: Mamba is more or less a drop-in replacement for the conda tool with the guts rewritten in C++ and as a result is much faster to resolve dependencies -- I want to say an order of magnitude faster or more. It's even noticeably faster at downloading and installing. You just

  conda install -c conda-forge mamba

in your base environment and then you can start running mamba instead of conda.

TIFF will support that size and larger. It can use JPEG or ZLIB compression as standard, or even WEBP or JPEG-XL which are unofficially supported by a small number of TIFF libraries.

I’ve used those Snappy brand cutters in a handheld cordless drill. It’s workable but tricky. You need to start with light pressure and high speed until the cutters score a clean circle, then gradually increase the pressure. The workpiece absolutely needs to be clamped down with zero movement, and it helps if you can find a way to brace your upper body and the drill as you work. The trickiest part is keeping the cutter perfectly square to the work in both axes —- if you tip it too far you’re likely to shatter the cutter as the metal is very brittle.

That species of moth doesn’t exist on the American continent but other similar and related ones do, whose range does overlap with that of hummingbirds. For example:

https://en.m.wikipedia.org/wiki/Hemaris_thysbe

This post claims the history of which releases closed which holes is a bit more complicated and that 8u191 really is the first release to prevent this particular exploit. However it also points out that it’s still possible to achieve RCE via log4j template expansion in certain Tomcat and Websphere configurations:

https://mbechler.github.io/2021/12/10/PSA_Log4Shell_JNDI_Inj...

You might be interested in this 100% compatible SB 1.0 reimplementation (including 8051 of course) : https://github.com/schlae/snark-barker

Also if the upper tray is height-adjustable, check the rubber one-way valve flaps in the middle of the supply tube that runs up the inside rear wall of the basin. If any of those flaps are damaged or missing, water will leak out there instead of being directed to the spray arm depending on which height position you're using. If you're lucky, you might still have one usable height position without replacing anything. Otherwise you need a whole new supply tube ($60-100) as the rubber part doesn't seem to be sold separately.

Nice work avoiding the copula there :)

PDF had support for JS at least since the late ‘90s as part of the scriptable forms system.

Dask does the lazy evaluation and query planning thing on numpy arrays and pandas dataframes, and can execute in parallel. It mimics most of their native interfaces which makes it a pretty easy drop-in.

https://docs.dask.org/en/latest/

True, but running the %debug magic command in IPython immediately after an unhandled exception drops you into the debugger at the point of the raised exception, complete with the full interpreter state including the stack. It doesn’t permit edit-and-continue but you can still evaluate expressions (including calling any function) to explore what went wrong without having to recreate the situation inside an explicit pdb session.

Yeah, that part of Cambridge was originally all industry. It’s been changing slower than you would think.

Shellac has nearly zero resistance to water vapor so it wouldn’t prevent corrosion at all. Nitrocellulose lacquer isn’t a lot better.

Maybe even more so the sequel Syndicate Wars

https://youtu.be/ECHmCzXMFb8

I have fond memories of using Swarm in an undergrad course on agent based simulation ~20 years ago. Just tweaking the heatbugs example in various ways was already enlightening. Thanks for your work!

I’m an American with a very limited space to work in and I love Paul’s approach.

Regarding finishes, I highly, highly recommend the book Understanding Wood Finishing by Bob Flexner. Flexner applies a very no-nonsense scientific approach which greatly demystifies the process. For example he explains what that pre-stain conditioner actually is (chemically), what it achieves (with photos), and exactly when you would want to use it. There is a lot of marketing and folklore around wood finishing which makes it a lot more confusing than it is, and this book helps cut through all of that.

The Duron had a set of circuit traces exposed on the ceramic chip carrier that were cut by laser at the factory to set the maximum allowed clock multiplier (and thus its final speed). You could reconnect all the cut traces with pencil lines and use a motherboard that let you set the multiplier directly to overclock freely.

Sorry, I thought Feather supported random row access but it turns out it only supports random column access.

For fst, I only played with the R interface, which would be called like this to retrieve row 12345 from the "fingerprint" column:

  read_fst("library.fst", columns="fingerprint", from=12345, to=12345)

However fst didn't offer a raw/binary column datatype last I checked, which is frustrating. It has chr (string) but R can't have embedded NUL bytes in strings, so that was a dead-end for efficient storage of binary fingerprints for R. I didn't check if the underlying fstlib structures accept NULs in string columns.

Hi Andrew! Looking into similar requirements I came across Feather and fst. They both basically let you efficiently slice into compressed on-disk DataFrames. Feather already supports Python, but fst is just for C++ and R at the moment.

https://blog.rstudio.com/2016/03/29/feather/

http://www.fstpackage.org/ https://github.com/fstpackage/fstlib

And it flat out refuses to run without an internet connection to fetch ads!

All those dashboards use the APIs underneath. The cost explorer dashboard certainly does incur API calls on my account.

You can fix some freeze-over issues by taking a length of stiff wire (14 gauge solid copper is great if you have some, a wire coat hanger is an OK substitute), wrapping one end tightly around the nearest part of the defrost cycle heater element, and routing the other end through the area where you are getting ice build-up. Some freezer designs allow ice to build up out of the defrost heater’s reach, and the added wire just conducts that heat where it’s needed.

I once used a computer lab where the AIX workstations each had their local filesystems exported to the entire campus via AFS. Even /dev. Including /dev/audio. Which was world-writeable. And there was no apparent volume control for the chassis speaker.

It’s a treasure from the “old web”. Information dense, simple markup, possibly updated by hand...