16 Sep 2024 |
moriel5 | In reply to @staudey:matrix.org Solus Firefox tends to rejigger the settings every once in a while for some reason. At least when it comes to languages. For me, the languages were untouched. | 18:07:19 |
Staudey | There is a fix in place for the language reset thing, but I've seen people complain about it again recently, so I thought it might be related | 18:07:53 |
Reilly Brogan | In reply to @moriel5:matrix.org Unfortunately, not with the I/O and featureset I need. Even what I had listed is 2nd hand/refurbished at these prices. That's fair. I have a lot of NVMe drives (I like to isolate I/O of various things for performance) and I struggle with the number of PCI-e lanes available on consumer CPUs | 18:11:58 |
Reilly Brogan | I think my next platform will have to be threadripper or low-end Epyc just to have enough | 18:12:17 |
Reilly Brogan | In reply to @moriel5:matrix.org
When I have the budget (well, technically I have one, however it is waiting for the budget to build a ThinkPad L15 Gen1 AMD).
I'm not going to switch to a drive that will die on me within a few years. The endurance of modern flash isn't really an issue any more in basically any usecase. Even low-end drives are rated for far more writes than any user is ever likely to use, even with swap | 18:13:41 |
Reilly Brogan | The 500GB SK Hynix Gold P31 for example (an excellent drive that's fairly affordable) uses TLC NAND and is rated for 500TB of writes, which is writing the entire space of the drive 1000x times. Virtually noone out of datacenters using them as caching drives is ever going to come close to that | 18:16:11 |
moriel5 | In reply to @reillybrogan:matrix.org That's fair. I have a lot of NVMe drives (I like to isolate I/O of various things for performance) and I struggle with the number of PCI-e lanes available on consumer CPUs For me it's less about NVMe specifically (although those too), but PCIe in general, as I need as many slots as possible, both for my uses, and for testing (as part of troubleshooting), and some of those PCIe cards could be controllers for other protocols, such as Firewire or serial (if the on-board isn't working for me) or parallel (all which I am still relying upon PCI for). | 18:27:31 |
Staudey | In reply to @reillybrogan:matrix.org The 500GB SK Hynix Gold P31 for example (an excellent drive that's fairly affordable) uses TLC NAND and is rated for 500TB of writes, which is writing the entire space of the drive 1000x times. Virtually noone out of datacenters using them as caching drives is ever going to come close to that Yeah, my SSD has seen 72.58 TB written over 12870 hrs power-on time. Would still be quite a while to get to even half of 500 TB. | 18:30:25 |
moriel5 | In reply to @reillybrogan:matrix.org The endurance of modern flash isn't really an issue any more in basically any usecase. Even low-end drives are rated for far more writes than any user is ever likely to use, even with swap I'm not a normal user. I could easily be compiling something, and then using SDIO to prepare Windows drivers for someone's PC (and the download of the driver packs is via P2P), and I the future I may also so something actually heavy (and I am aiming for at least 10 years of endurance under semi-heavy load). | 18:30:28 |
moriel5 | In reply to @reillybrogan:matrix.org The 500GB SK Hynix Gold P31 for example (an excellent drive that's fairly affordable) uses TLC NAND and is rated for 500TB of writes, which is writing the entire space of the drive 1000x times. Virtually noone out of datacenters using them as caching drives is ever going to come close to that That is certainly a good drive, and if it not for the Firecuda, I would have recommended it. | 18:31:45 |
moriel5 | Apparently I have permission to react to others' posts, however I do not have permission to remove my reaction? | 18:35:58 |
moriel5 | * Apparently I have permission to react to others' posts here, however I do not have permission to remove my reaction? | 18:36:16 |
helium1 | no you can't remove reactions in this room | 18:37:07 |
helium1 | and delete your posts | 18:37:37 |
helium1 | only edit | 18:37:53 |
David Harder | In the past we've had members who would delete their posts sporadically. It was surprisingly disruptive, so it got nerfed | 18:40:17 |
helium1 | Redacted or Malformed Event | 18:41:44 |
Joey Riches | petition granted 📜 🤓 | 18:54:27 |
Joey Riches | (sry xD) | 18:54:33 |
stigarn | Cool that Github autmatically just adds link to cves | 19:31:34 |
stigarn | just added the cve number and then it linked to them | 19:31:54 |
Staudey | The Software Center does too but nobody appreciates it. Poor SC 😥 | 19:34:02 |
| dimasic joined the room. | 19:43:56 |
dimasic | Hi | 19:44:17 |
helium1 | hello new Solus user! | 19:44:36 |
dimasic | Can anyone explain to me what usr merge will be needed for? | 19:45:43 |
TraceyC | If you hadn't already read it, we have a blog post about it. | 19:48:17 |
helium1 | yup! | 19:48:42 |
TraceyC | Beyond anything else, it's becoming an industry standard. It will also be needed to integrate Serpent tooling | 19:48:43 |
Staudey | It's also required for systemd 255 | 19:50:18 |