Avatar

havoc

havoc

About

Username
havoc
Joined
Visits
4,921
Last Active
Roles
Member, OG, Content Writer, Senpai
Thanked
1647

Comments

  • Haven't yet, but have been watching discussions on it. Best as I can tell iPhone gang in selfhosted space is going for an app called PhotoSync, which has a 6 bucks lifetime tier of some sort to enable more features. Haven't explored it yet, but tha…
  • Decided to build custom scripts that simulate real proxmox tasks. Want to set up tuned ZFS volumes for each so don't think i'll get around real world tasks. So far LXC testing looks like this: Created container in seconds:4.04Started container in s…
  • (Quote) Oh that looks promising. Got an error on trying to compile it. /* Counters */int files_created; /* number of files created */int files_deleted; /* number of files deleted */int files_read; /* number of files read */int files_appended;…
  • P1600X optane drive fio Disk Speed Tests (Mixed R/W 50/50) (Partition /dev/mapper/p1600x-vm--105--disk--0):---------------------------------Block Size | 4k (IOPS) | 64k (IOPS) ------ | --- ---- | ---- -…
  • (Quote) They're all different things. Haven't found a nice diagram that has an overview, but best as I can tell it's like this: Neural network (broad overarching category) * Graph neural nets --> The recommendations somik listed * LLMs (but eve…
  • (Quote) Well sometimes it takes me two tries to figure out whether the script needs to start with #! or !# but yes absolute bash master right here
  • (Quote) Alas I don't think I'm impressing you lot with my janky homeserver. But yeah basically want something that feels snappy for real world homelab proxmox use. So if I'm gonna spend a bunch of time tuning may as well try to measure the right me…
  • Deepseek claims they're profitable, so while there is lots of free VC money shenanigans going on I don't think it's true across the board. That said...deepseek is in China...who knows what sort of indirect subsidies (elec price etc) they didn't cou…
  • (Quote) Now I just need 20 grand worth of GPUs to run it...
  • (Quote) Yeah should be fine though you'd need to quantize it a bit to make space for context. I'd probably do qwen3 models. Either dense one or you can try the MoEs 32B and put it on system mem partially. Will still be fast given low activations in …
  • (Quote) Good tip! Thanks.
  • (Quote) Keep forgetting they don't just help themselves to my wallet...
  • They just charged me a late payment fee :(
  • Kinda curious which country moves fastest to adapt to this entire change. Clearly the genie is not going back in the bottle so we're going have sizable chunks of professions that are going to need a plan B. And not sure how flexible the average ~45 …
  • (Quote) #!/bin/bashapt install build-essential make cmake libcurl4-openssl-dev curl python3-pip pkg-config libopenblas-dev gitgit clone https://github.com/ggml-org/llama.cppcd llama.cpp/cmake -B build -DGGML_BLAS=ON -DGGML_BLAS_VENDOR=OpenBLAScmake …
  • More impressed that someone found a CPU with a sample size of one passmark: (Image)
  • My new janky AF ebay parts homeserver. Absolutely not going to win performance contests - but it is growing on me. All my other homelab stuff is SBC/minipc so was really missing something more ATX sized...pcie slots, satas etc. Highlights - Triple …
  • (Quote) Same. Must admit I'm a little puzzled by storage VPS in general Mission critical stuff goes to big providers (hetzner and up). Linux ISOs & friends I'll keep two copies on local LAN - acceptable risk profile and I'd rather not have linu…
  • (Quote) That looks like half a petabyte of drives? o_O Nice info! Thanks for taking the time Anybody know how this works at hyperscaler level? I heard google does triple region replication for gmail data so guessing it flips back to software at th…
  • Is anyone even doing hardware RAID anymore? I get the sense that everyone moved to software because it is less of a mess on hardware failure (specifically the controller)
  • (Quote) Kagi search has a small-web search feature which is basically a webrings-like content but in centralised curated format by them (Image)
  • That sucks :( Put plans to make a blog on hold because it just feels a bit it would be speaking into the void. I noticed in my own browsing I'm either asking an LLM or googling and often the AI summary on top has the key fact I'm looking for so no …
  • (Quote) haha indeed. (though I do have angry customers in day job that is funding my homelab shenanigans...) Anyway...no storage testing for now. XMP on this board seems broken. So guess I'm figuring out 30+ memory timings settings by hand
  • @host_c Thanks! Use case is various homelab stuff. Relevant one for discussion is probably half a dozen nodes will be using this as s3/object storage backend for k8s hence interest in fast small files and snappiness. NFS too but most of LAN is 2.5g…
  • (Quote) You clearly know a lot about this! But think you misunderstood me. I'm not thinking cache, leaning more towards metadata/small files on optane. i.e. the data never goes to main pool. (Quote) Well given that I did a non-ECC build I think in…
  • Awesome timing. Been wondering about this too for a project. I suspect with main pool being SSD and thus already pretty fast any sort of caching layer beyond RAM is of limited benefit. So I'm thinking more metadata & small files (vaguely recall…