There are still some gains to be had on the CPU. I'm at ~570 MH/s on a 5950x

I'm a bit late to the game, but I managed to squeeze out about 1 GH/s on an i7-13700K.

I'm glad that it works better for you, but I have literally none of the problems you list.

I have never in my life typed out a URL (with an id, google.com and stackoverflow.com don't count).

I don't put naked ids alone in spreadsheets. If, once a year, I do need to make a spreadsheet, I put full URLs there.

Length of these ids is comparable to that of base-10 numbers and neither are "very long".

And so on.

thanks for the feedback, you're right, the "url safe" isn't the main focus (well, shouldn't, it's just that it's the last issue I had ;)

For further context, we are working on online campaigns (petitions, sending emails to elected representative...) and everything has either ids (numbers) that we generate (a campaign, a action, a supporter...) or given externally (eg. the elected representatives or vote reference).

Most of these ids are in the 1k to 100k range, that happen to often be the same range as the number of actions taken or emails sent) and mixing these KPIs and ids on the same lists happen to be confusing for users.

It might be that we shouldn't try to put these ids in lists and summaries, but they are super convenient there sometimes and unavoidable when exporting to excel, so dxid was my solution to work around our potentially poor ux design ;)

> participants believe that if they get a lot of points they have easier time finding a well-paying job

It’s true, though. I have a high-rep account and I’ve had a few jobs fall in my lap because of that.

- hey, can I copy your homework? - sure, but change it so it doesn't look like a copy

So, in this "next level" database you have to do the resharding yourself? That is, manually specify which keys go where?

Way to go, my friends :)