dkechag

The primes one is my least favourite one indeed, I left it in just because I happened to include it in the very first version and I am thinking it just counts for 5% in the end... The regex ones are "micro" yet quite important, dkbench it's a Perl (and C)-based benchmark (reflects our main code), and the regex engine is the most highly optimized part of the language so regex speed is a good representation of text processing speed in Perl. As I said, the overall score correlates well to SPEC/Geekbench so as a suite it works well. For compiler comparisons I usually compile a language like python or perl as a test, but I did not want to add something like that, to keep it fast with many smaller benchmarks.

It came up recently in a post I did about how bad Geekbench 6 multi-core is, and some are saying they still use it for single-thread as it correlates well with SPEC. I thought I'd mention the one I built and use, which also correlates well with SPEC (and Geekbench 5+) but is open source and scales well up to hundreds of cores.

It came up recently in a post I did about how bad Geekbench 6 multi-core is, and some are saying they still use it for single-thread as it correlates well with SPEC. I thought I'd mention the one I built and use, which also correlates well with SPEC but is open source and does not have scaling issues.

They warn they will reclaim. I had two accounts, one processing data for a free weather service I volunteer for, so it's not idle and has had no issues for a couple of years now. The other for personal projects, so at times it stays idle for a while and I would get these email warnings which made me switched it to paid. I have not paid a cent, again for a couple of years now.

I had a bit of a hard time when I first signed up. It turns out the problem was I had included "oracle" in my email (I give a different email on my domain to every provider), so some part of the system considered it an internal email and others part did not, so I had a weird account that could not do anything. Took them a month of two to figure out what was wrong...

I meant pretty much anything Oracle... except their free/cheap VMs :)

The testing was from mid-September to mid-January, Turin was released on Azure end of January. They simply did not make the cut, unfortunately I could not wait forever for them, they always seem to lag behind at least half a year :(

I might do an addendum just for Azure...

Huh, I have not been able to provision any newer CPUs after dozens of tests, certainly not Emerald Rapids. And that blog post is weird, their charts don't even have a key shown, it's like they bought a few CPUs and threw that quickly together to get people's hopes up. A real shame, I am still running DO droplets, but they are behind the times...

Yeah, that's the main reason. Not sure why the ban, medium etc are much worse...

The second reason is I've never had a slashdot submission accepted and I saw that after posting this the page suggested I share it to increase chances of the editors picking it up, but I don't really use social media so I though hey why not HN.

There's a table of contents near the top. Use it to jump to the performance per dollar.

> No theory. No fluff. Just production.

ChatGPT tells me "no theory, no fluff" all the time :D