Hacker News: genpfault

New comment by genpfault in "Qwen3.6-27B: Flagship-Level Coding in a 27B Dense Model"

genpfault — Wed, 22 Apr 2026 15:52:14 +0000

Getting ~36-33 tok/s (see the "S_TG t/s" column) on a 24GB Radeon RX 7900 XTX using llama.cpp's Vulkan backend:

    $ llama-server --version
    version: 8851 (e365e658f)

    $ llama-batched-bench -hf unsloth/Qwen3.6-27B-GGUF:IQ4_XS -npp 1000,2000,4000,8000,16000,32000 -ntg 128 -npl 1 -c 34000
    |    PP |     TG |    B |   N_KV |   T_PP s | S_PP t/s |   T_TG s | S_TG t/s |      T s |    S t/s |
    |-------|--------|------|--------|----------|----------|----------|----------|----------|----------|
    |  1000 |    128 |    1 |   1128 |    1.529 |   654.11 |    3.470 |    36.89 |    4.999 |   225.67 |
    |  2000 |    128 |    1 |   2128 |    3.064 |   652.75 |    3.498 |    36.59 |    6.562 |   324.30 |
    |  4000 |    128 |    1 |   4128 |    6.180 |   647.29 |    3.535 |    36.21 |    9.715 |   424.92 |
    |  8000 |    128 |    1 |   8128 |   12.477 |   641.16 |    3.582 |    35.73 |   16.059 |   506.12 |
    | 16000 |    128 |    1 |  16128 |   25.849 |   618.98 |    3.667 |    34.91 |   29.516 |   546.42 |
    | 32000 |    128 |    1 |  32128 |   57.201 |   559.43 |    3.825 |    33.47 |   61.026 |   526.47 |

New comment by genpfault in "Framework Laptop 13 Pro"

genpfault — Tue, 21 Apr 2026 20:08:11 +0000

Hear, hear!

New comment by genpfault in "Google releases Gemma 4 open models"

genpfault — Fri, 03 Apr 2026 02:14:33 +0000

llama.cpp (b8642) auto-fits ~200k context on this 24GB RX 7900 XTX & it shows a solid 100+ tok/s ("S_TG t/s") on the first 32k of it, nice!

    ./llama-batched-bench -hf unsloth/gemma-4-26B-A4B-it-GGUF:UD-Q4_K_XL \
    -npp 1000,2000,4000,8000,16000,32000,64000,96000,128000 -ntg 128 -npl 1 -c 0
    |    PP |     TG |    B |   N_KV |   T_PP s | S_PP t/s |   T_TG s | S_TG t/s |      T s |    S t/s |
    |-------|--------|------|--------|----------|----------|----------|----------|----------|----------|
    |  1000 |    128 |    1 |   1128 |    0.416 |  2404.87 |    1.064 |   120.29 |    1.480 |   762.20 |
    |  2000 |    128 |    1 |   2128 |    0.755 |  2649.86 |    1.075 |   119.04 |    1.830 |  1162.83 |
    |  4000 |    128 |    1 |   4128 |    1.501 |  2665.72 |    1.093 |   117.08 |    2.594 |  1591.49 |
    |  8000 |    128 |    1 |   8128 |    3.142 |  2545.85 |    1.114 |   114.87 |    4.257 |  1909.47 |
    | 16000 |    128 |    1 |  16128 |    6.908 |  2316.00 |    1.189 |   107.65 |    8.097 |  1991.73 |
    | 32000 |    128 |    1 |  32128 |   16.382 |  1953.31 |    1.278 |   100.12 |   17.661 |  1819.16 |
    | 64000 |    128 |    1 |  64128 |   43.427 |  1473.74 |    1.453 |    88.12 |   44.879 |  1428.89 |
    | 96000 |    128 |    1 |  96128 |   82.227 |  1167.50 |    1.623 |    78.86 |   83.850 |  1146.42 |
    |128000 |    128 |    1 | 128128 |  133.237 |   960.69 |    1.797 |    71.25 |  135.034 |   948.86 |

New comment by genpfault in "CodingFont: A game to help you pick a coding font"

genpfault — Mon, 30 Mar 2026 16:04:35 +0000

Doesn't seem to serve rendered samples so you have to set "browser.display.use_document_fonts" to "1" to see anything useful.

New comment by genpfault in "Intel Announces Arc Pro B70 and Arc Pro B65 GPUs"

genpfault — Thu, 26 Mar 2026 15:27:41 +0000

600 GB/s of memory bandwidth isn't anything to sneeze at.

~$1000 for the Pro B70, if Microcenter is to be believed:

https://www.microcenter.com/product/709007/intel-arc-pro-b70...

https://www.microcenter.com/product/708790/asrock-intel-arc-...

New comment by genpfault in "Parallel Perl – Autoparallelizing interpreter with JIT"

genpfault — Fri, 20 Mar 2026 19:12:57 +0000

Not to be confused with GNU parallel[1], written in Perl.

[1]: https://en.wikipedia.org/wiki/GNU_parallel

New comment by genpfault in "Can I run AI locally?"

genpfault — Sat, 14 Mar 2026 05:40:24 +0000

Speculative decoding[1]?

[1]: https://github.com/ggml-org/llama.cpp/blob/master/docs/specu...

New comment by genpfault in "The dead Internet is not a theory anymore"

genpfault — Wed, 11 Mar 2026 21:58:30 +0000

> A social networking system simulates a user using a language model trained using training data generated from user interactions performed by that user

Google People[1]?

[1]: https://qntm.org/perso

New comment by genpfault in "Why applicant tracking systems are broken by design"

genpfault — Thu, 19 Feb 2026 18:55:53 +0000

> "A system's purpose is what it does"

POSIWID: https://en.wikipedia.org/wiki/The_purpose_of_a_system_is_wha...

New comment by genpfault in "Art of Roads in Games"

genpfault — Mon, 09 Feb 2026 03:01:36 +0000

https://www.redblobgames.com/articles/curved-paths/

New comment by genpfault in "Qwen3-Coder-Next"

genpfault — Wed, 04 Feb 2026 03:18:26 +0000

Nifty, thanks for the heads-up!

New comment by genpfault in "Qwen3-Coder-Next"

genpfault — Tue, 03 Feb 2026 18:17:38 +0000

Nice! Getting ~39 tok/s @ ~60% GPU util. (~170W out of 303W per nvtop).

System info:

    $ ./llama-server --version
    ggml_vulkan: Found 1 Vulkan devices:
    ggml_vulkan: 0 = Radeon RX 7900 XTX (RADV NAVI31) (radv) | uma: 0 | fp16: 1 | bf16: 0 | warp size: 64 | shared memory: 65536 | int dot: 1 | matrix cores: KHR_coopmat
    version: 7897 (3dd95914d)
    built with GNU 11.4.0 for Linux x86_64

llama.cpp command-line:

    $ ./llama-server --host 0.0.0.0 --port 2000 --no-warmup \
    -hf unsloth/Qwen3-Coder-Next-GGUF:UD-Q4_K_XL \
    --jinja --temp 1.0 --top-p 0.95 --min-p 0.01 --top-k 40 --fit on \
    --ctx-size 32768

New comment by genpfault in "JuiceSSH – Give me my pro features back"

genpfault — Mon, 26 Jan 2026 19:24:05 +0000

> ConnectBot

https://f-droid.org/en/packages/org.connectbot/

https://github.com/connectbot/connectbot

New comment by genpfault in "Killing the ISP Appliance: An eBPF/XDP Approach to Distributed BNG"

genpfault — Fri, 23 Jan 2026 18:38:26 +0000

Because TFA never bothered to define it:

Broadband Network Gateway (BNG)[1]

[1]: https://github.com/codelaboratoryltd/bng#bng-broadband-netwo...

New comment by genpfault in "Task-free intelligence testing of LLMs"

genpfault — Thu, 08 Jan 2026 23:50:34 +0000

tap tap tap tap tap

New comment by genpfault in "Nvidia Nemotron 3 Family of Models"

genpfault — Thu, 18 Dec 2025 05:56:25 +0000

Getting ~150 tok/s on an empty context with a 24 GB 7900XTX via llama.cpp's Vukan backend.

New comment by genpfault in "Multivox: Volumetric Display"

genpfault — Thu, 04 Dec 2025 20:31:12 +0000

It was right there[1] in the assembly video.

[1]: https://youtu.be/pcAEqbYwixU?t=1038

New comment by genpfault in "Customize Nano Text Editor (2022)"

genpfault — Tue, 04 Nov 2025 19:43:05 +0000

> Ctrl+S to save

XOFF ignored, mumble mumble

New comment by genpfault in "NPM flooded with malicious packages downloaded more than 86k times"

genpfault — Thu, 30 Oct 2025 23:43:43 +0000

> triple-backtick code blocks

If only :(

New comment by genpfault in "How to save the world with ZFS and 12 USB sticks: 4th anniversary video (2011)"

genpfault — Sat, 11 Oct 2025 17:03:07 +0000

> Seagate sometimes has decent prices on new.

Make sure to check the "annual powered-on hours" entry in the spec sheet though, sometimes it can be significantly less than ~8766 hours.