nomennescio

Compression factor 7.4 ``` -rwxrwx--- 1 root vboxsf 459713896 Feb 15 19:21 factor.basis+core+extra.image -rwxrwx--- 1 root vboxsf 61979751 Feb 15 19:39 factor.basis+core+extra.image.compressed ``` ``` /media/sf_project/software/factor.fork$ time for i in {1..100};do...

@mrjbq7 I expect to improve upon these figures if I can get the version to work which uses a preallocated buffer. The interesting thing is that you can see a...

And I think a more meaningful benchmark than just executing an empty script can show the difference between having all objects in memory versus only having them as memory mapped...

And of course, I would like to see these benchmarks run on real systems, not just a VirtualBox... Something you can do?

Getting ZSTD to allocate as little memory as possible, and to avoid memcpy-s turns out to be quite experimental, and very non-trivial. But this is exactly where the expected performance...

You mean the second one? Or both?

Found another example: ``` { { 0 1 } { 2 3 } } ``` is pushed onto the stack and displayed in the Listener as ``` { ~array~ ~array~...

> Why not add them to the math.unicode vocabulary? I was not aware that would be the location to do so, I will have a look at that.

There starts the first issue already. When I `USE: math.unicode` and look in the browser at math.unicode, I see many glyphs not being rendered, but a "box" is shown in...

BTW I see all `ALIAS:` and `CONSTANT:` being declared `inline` in that vocab. Is that required for efficient code? Why should it NOT be inline?