| 24 Nov 2022 |
 pie_ | idk te enterprise gpu stuff, k20s and k40s or soemthing, also something about xeon phi | 02:30:06 |
 tpw_rules | i intend to write up and publish how i did that at some indefinite point... | 02:30:07 |
| tpw_rules | ooh | 02:30:17 |
| pie_ | * idk the enterprise gpu stuff, k20s and k40s or soemthing, also something about xeon phi | 02:30:20 |
| tpw_rules | xeon phi is firmly in the nerd kitsch category at this point i'm pretty sure | 02:30:36 |
| pie_ | but yeah its ~10 year old hardware | 02:31:20 |
| pie_ | anyway | 02:31:23 |
| pie_ | I am trying to rebuild the world, at least as far as im needing it, on musl | 02:31:38 |
| pie_ | tats most of the pain so far | 02:31:43 |
| pie_ | i mean nix is at least making it feasible | 02:31:47 |
| tpw_rules | why | 02:31:47 |
| tpw_rules | that's the real danger of nix | 02:31:56 |
| pie_ | because i need a libc that works with an old kernel? | 02:32:09 |
| tpw_rules | oh yes that makes sense | 02:32:20 |
| pie_ | apparently musl supports everything more or less | 02:32:39 |
| pie_ | gcc broke compatibility at some point | 02:32:47 |
| pie_ | * glibc broke compatibility at some point | 02:32:56 |
| pie_ | musly explicitly supoprts 2.6.36 or 2.6.39 i think | 02:33:17 |
| pie_ | * musl explicitly supoprts 2.6.36 or 2.6.39 i think | 02:33:25 |
| pie_ | and ajs pointed out to me that we have a special glibc courtesy of vcunat IIRC that would work with rhel6 but I dont think I have a highe nough patch level | 02:34:02 |
| pie_ | * and ajs pointed out to me that we have a special glibc courtesy of vcunat IIRC that would work with rhel6 but I dont think I have a high enough patch level | 02:34:12 |
| pie_ | the more you know shrug | 02:34:19 |
| pie_ | though my guess is rhel is patchd to suppot the libc then and not so much the other way around, idk | 02:35:08 |
| pie_ | currently im
- mutilating fhsuserenv trying to get it to eval withmusl
- trying to get some python stuff to build
| 02:35:55 |
| pie_ | appppaaaarrently fhsuserenv only works with glibc? Im not sure | 02:36:12 |
| pie_ | I tried to remove all the i686 stuff but something is still hanging on and preventing me from evaling | 02:36:30 |
| pie_ | (i think I just waited for 3 fortran stage compiles on the side or something...blas and lapack are compiling..) | 02:37:15 |
| pie_ | In reply to @tpw_rules:matrix.org
that's the real danger of nix Im wondering if this is a more feasible way of upgrading userspace if, and until, we can upgrade the kernel | 02:37:58 |
| pie_ | its an open question why its so old, we are supposed to get custory of the machine...at some point | 02:38:38 |
| pie_ | * its an open question why its so old, we are supposed to get custody of the machine...at some point | 02:38:45 |
