| 1 Aug 2025 |
 dramforever | https://web.archive.org/web/20250106210601/https://cohost.org/dram/post/5512696-the-nix-os-loongson-2 | 12:01:24 |
 emily | of course it's Lemote | 12:01:37 |
| dramforever | a certain famous guy has been known to use a variant of this with a larger screen | 12:02:20 |
| emily | what's the deal with LoongArch64 btw. is it MIPS with extra steps or RISC-V with extra steps | 12:02:20 |
 K900 | It's MIPS | 12:02:42 |
| emily | oh you mention it here | 12:02:43 |
| K900 | No extra steps | 12:02:47 |
| K900 | It's just MIPS but they're not paying for it | 12:02:55 |
| emily | there must be some extra step | 12:02:55 |
| emily | for it to get accepted separately into Linux | 12:02:59 |
| dramforever | i mean... riscv is mips with extra steps right? | 12:03:02 |
| K900 | The extra step is 1) make a lot of it 2) sell to Chinese hyperscalers | 12:03:17 |
| emily | in the case of RISC-V that's a lot of extra steps | 12:03:17 |
| dramforever | no i don't think loongarch is any more mips than riscv is | 12:03:19 |
| emily | like it's MIPS brain but I don't think you can call it MIPS | 12:03:25 |
| K900 | Loongarch is compatible with MIPS64 | 12:03:54 |
| dramforever | the extra step is the chinese company has control over the whole thing now | 12:03:54 |
| K900 | (though they'll never admit it) | 12:03:59 |
| dramforever | rather than licensing it from whoever else | 12:04:01 |
| K900 | I think they have some things on top now | 12:04:06 |
| emily | fair warning I am removing lib.systems.platforms.fuloong2f_n32 | 12:04:25 |
| emily | but it was broken anyway | 12:04:28 |
| emily | you shouldn't need it if you build a more normal system | 12:04:35 |
| dramforever | In reply to @k900:0upti.me
Loongarch is compatible with MIPS64 source? compatible in what sense? | 12:04:42 |
| emily | fwiw K900's statements match what I heard | 12:04:54 |
| emily | that LoongArch just decided to fork MIPS | 12:05:04 |
| dramforever | i've never been able to track down a source on "compatible" | 12:05:27 |
| dramforever | yeah, "fork" is more like it | 12:05:38 |
| K900 | It's the exact same instruction encodings | 12:05:58 |
| K900 | I think they might have diverged somewhat by now? | 12:06:09 |
