simd

Though we include Boost JSON as a dependency for benchmarking purposes, we do not include it as part of our benchmarks currently.

Scenario

A user is ingesting some data (~300M rows) into a table with 270K distinct symbol values. For each of these 270K symbols we end up with at least one posix_fallocate() necessary to add a new value block. See the below flamegraph - it's quite illustrative.

An important part of the scenario is non-default index block capacity which is set to 2048. With the default capacity (w

Our users are often confused by the output from programs such as zip2john sometimes being very large (multi-gigabyte). Maybe we should identify and enhance these programs to output a message to stderr to explain to users that it's normal for the output to be very large - maybe always or maybe only when the output size is above a threshold (e.g., 1 million bytes?)

I am finding a product which can replace Elasticsearch.

varying float<3> col = { 0, 0, 0 };
uniform float <3> reduce_col;
reduce_col.x = reduce_add( col.x );
reduce_col.y = reduce_add( col.y );
reduce_col.z = reduce_add( col.z );
reduce_col.w = reduce_add( col.w );

This currently does not produce an error. I expect attempting to access the w component of a three component type to be an error.

Hi,

thank you very much for this cool library.

I noticed that simde_mm512_load_ps is missing while simde_mm512_loadu_ps is implemented. However, both seem to exist for Intel intrinsics (see screenshot). Maybe I oversaw one or maybe they do the same anyways?

Thanks
Lukas

![Screenshot from 2021-10-25 13-28-29](https://user-images.githubusercontent.com/14993470/138687358-aa6b531e-bdc