| 9 Jul 2025 |
 dramforever | ghostwrite is just the most well known of all problems | 10:18:41 |
| dramforever | and i'll just say turning the non-compliant vector extension off is almost a trivial workaround | 10:19:50 |
 zimward | its a bug in silicon. there is just no other workaround lol | 10:20:43 |
 Alex | Yikes. Did they not do enough test simulations while designing the core? | 10:21:50 |
| zimward | no differential fuzzing atleast | 10:22:06 |
| dramforever | what do you simulate? | 10:22:11 |
| dramforever | but also there's two teams involved, the now-called xuantie that made the c920v1 cores, and the sophgo team the made the whole processor by slapping together 16 of 4 core clusters | 10:24:09 |
| dramforever | this ... did not go particularly well | 10:25:38 |
| dramforever | oh, i remembered something i can share: LLVM-compiled multi threaded programs may run very slowly because fence(acq_rel) is compiled to fence.tso | 10:26:41 |
| dramforever | which the cores raise an illegal instruction exception due to a misunderstanding in the specs | 10:27:02 |
| Alex | So the system has to context switch on most fences?
Pure pain. | 10:27:55 |
 Steven Keuchel | I have both the pioneer (sg2042) and the megrez (eic7700x). the c920 cores in the sg2042 are kinda slow. too slow to use as a desktop. and a lot of processes swamp the memory subsystem. the p550 are already much faster. I would estimate that 4x eic7700x might outperform the sg2042, and would probably also be cheaper. | 10:28:36 |
| dramforever | not most fences, just this specific one | 10:28:38 |
| dramforever | just randomly fences taking hundreds of cycles | 10:28:53 |
| dramforever | but it is every time it executes | 10:29:04 |
| Alex | Ah, I'm not very familiar with LLVM's multithreading primitives... | 10:29:12 |
| dramforever | so gcc can do -mno-fence-tso | 10:30:02 |
| dramforever | but i don't think llvm is getting that any time soon, or ever | 10:30:11 |
| zimward | fun stuff. but sounds like it should be an issue in all C920v1 cores | 10:50:04 |
| dramforever | yes | 10:52:34 |
 Tristan Ross | In reply to @skeuchel:matrix.org
I have both the pioneer (sg2042) and the megrez (eic7700x). the c920 cores in the sg2042 are kinda slow. too slow to use as a desktop. and a lot of processes swamp the memory subsystem. the p550 are already much faster. I would estimate that 4x eic7700x might outperform the sg2042, and would probably also be cheaper. Matthew Croughan has a P550 and I've tried compiling things on it. We've seen a lot of things fail on it which makes me think there's something wrong. How does it perform for you? | 13:52:33 |
| dramforever | In reply to @rosscomputerguy:matrix.org
Matthew Croughan has a P550 and I've tried compiling things on it. We've seen a lot of things fail on it which makes me think there's something wrong. How does it perform for you? fail as in...? | 13:59:50 |
| dramforever | random userspace crash? | 13:59:55 |
| Tristan Ross | A lot of test failures like bus errors and io errors | 14:00:21 |
| Steven Keuchel | It's not as stable as I would like because the board crashes about once a week, but I am not running into the problems you describe. I have fewer tests timing out than on the sg2042, and other test failures are clearly software, not hardware related. | 14:06:57 |
| Steven Keuchel | * It's not as stable as I would like because the board crashes about once a week even when idle, but I am not running into the problems you describe. I have fewer tests timing out than on the sg2042, and other test failures are clearly software, not hardware related. | 14:07:12 |
| Tristan Ross | Had to look at the chat history, one of the things we saw were GCC could sometimes throw a bus error. Some git and gettext things had io errors. | 14:12:57 |
| dramforever | anything in dmesg when that happened? | 14:21:34 |
| Tristan Ross | Yeah, there were dmesg things but it'll be hard to scroll back to get that lol | 14:26:00 |
| dramforever | anyway, my point about reliable hardware just got strengthened | 14:31:12 |
