ColonelPhantom

5 ms·

1.
Supporting ATI TeraScale GPUs from 2007-2009 in RPCS3(blog.rpcs3.net)
5 points by ColonelPhantom 2mo ago|1 comments
2.
by ColonelPhantom 2mo ago
Or HOGS? HTMX-OS-Go-Sqlite. While having "OS" in there is kind of redundant, it does make for a nice and general acronym.
3.
by ColonelPhantom 3mo ago
Which model class requires an 80 GB VRAM GPU? From my perspective, popular models seem to be either in the ~30B range (Qwen3.6, Gemma 4), while the larger models (MiniMax, MiMo, StepFun, Deepseek) are in the multiple hundreds of billions pa
4.
by ColonelPhantom 3mo ago
> the most likely experts Is that how MoEs work? I though that an important constraint for MoEs is that experts need to be uniformly used to make sure they can be used effectively. If there is a 'common subset' that, if anythin
5.
by ColonelPhantom 3mo ago
Deepseek V4 Flash still has 13B active params though? That is about half as many as Qwen3.6-27B (and much more than Qwen3.6-35B-A3B). Given that RAM (even on a base M4 or 'regular' Intel/AMD system) is like an order of magnit
6.
by ColonelPhantom 4mo ago
Carp is memory safe via linear types + references, similar to Rust, so I would not describe it as C-like but rather Rust-like.
7.
by ColonelPhantom 5mo ago
But what _is_ a "Text User Interface"? Google Images just returns what is being discussed here: "GUIs" that run in some kind of text mode. And to me, that's also what a TUI is. A more textually oriented environment
8.
by ColonelPhantom 5mo ago
You mentioned Strix Halo, which also has off-die memory. Strix Halo does have a real advantage from its wider memory bus (four channels for 256 bit instead of 128 bit), but Strix Point is equivalent-ish to Intel's platforms like Panthe
9.
by ColonelPhantom 5mo ago
Hilariously, those AMD chips are way behind the Intels in terms of memory. First off, I believe that Intel has its memory far more "unified". AMD typically has a stricter VRAM/RAM 'tradeoff' setting that does not ex
10.
by ColonelPhantom 5mo ago
Nvidia Turing (RTX 20) definitely marked a major shift IMO. - It was the first card to enable real-time ray-traced effects. - Mesh shaders are a significant overhaul of the geometry pipeline that's only recently getting real traction.
11.
by ColonelPhantom 6mo ago
838 seems to be the real INT8 TOPS number for the 5090; going from 800 to 3400 takes an x2 speedup for sparsity (so skipping ops) and another x2 speedup for FP4 over INT8. So it's closer to half the speed than a tenth. Intel also seems
12.
by ColonelPhantom 6mo ago
Aren't Intel Xeon Rapids and Intel Xeon Forest just different target markets? Rapids has fewer but faster cores in general, and more special-purpose accelerators (e.g. AMX, QAT), while Forest is focused on maximum compute density (just
13.
by ColonelPhantom 6mo ago
Very cool! I am wondering one thing: how fast is it? Much of the "secret sauce" of the Voodoo is its high speed: a first-gen Verite or (God forbid) any ViRGE takes many more cycles for common operations like, say, Z-buffered pixel
14.
by ColonelPhantom 6mo ago
GPT-OSS is tailored to be extremely memory efficient. Not only is it natively using the 4.25 bit per token MXFP4 format, but it also uses sliding window attention for half of its layers. It also doesn't have that many layers, only 36 f
15.
by ColonelPhantom 6mo ago
Interesting; I was not aware of those "universal synthetics" but they make sense: a stronger reasoning base would make modeling tasks easier. Thanks for the link! Again, though, if those work I assume they will be used for the slo
16.
by ColonelPhantom 6mo ago
If generating synthetic data is such a great way to improve performance, why would it not be applied to the slowrun? Especially for the unlimited compute track, you should have plenty of time to generate as much synthetic data as your heart
17.
by ColonelPhantom 6mo ago
I like the idea of LLM-calling as an automation-friendly CLI tool! However, putting all my agents in ~/.config feels antithetical to this. My Bash scripts do not live there either, but rather in a separate script collection, or prefera
18.
by ColonelPhantom 6mo ago
Interesting read! One remark though: I'm not too familiar with the architecture of a Google TPU, but comparing the TPU's VMEM with Nvidia's shared memory feels wrong to me. Looking at the size, and its shared nature, it feels
19.
by ColonelPhantom 7mo ago
The reason for that is that most memory bandwidth bumps come with new memory generations. For example an early DDR4 platform (e.g. Intel Skylake/Core iX-6000) and a late one (e.g. AMD Zen3/Ryzen 5000) only differ by 1.5x as well,
20.
by ColonelPhantom 7mo ago
It's not just Qwen; we also recently had GLM-4.7-Flash in the same roughly 30B-A3 range. Seems to me like there's no shortage of competition for good old GPT-OSS 20B (not just Qwen3.5-35B and GLM-4.7-Flash, but also Qwen3(-Coder)-
21.
by ColonelPhantom 8mo ago
> Meanwhile, European makers are stuck not knowing what to do, make Americans happy or compete with the Chinese. Huh? This comment sounds extremely America-centric to me. Porsche sells more cars into Europe than North America, despite ta
22.
by ColonelPhantom 9mo ago
Each layer of the LM is also at most 16 KiB, so if you want to minimize bank switching, I think making sure each layer is in one bank would be enough? Bank switching shouldn't give much overhead anyway unless it complicates an inner lo
23.
by ColonelPhantom 9mo ago
Touchegg kinda sucks (gestures are not 1:1 but rather just "triggered"), and you also don't need it. KDE and Gnome (as well as some WMs like Niri) have native touchpad gesture support on Wayland. Using my touchpad for history
24.
by ColonelPhantom 10mo ago
I have no idea what you are saying (with "is"??), but I don't think this is true: KDE Dolphin is very full-featured and runs natively on Wayland.
25.
by ColonelPhantom 10mo ago
That's fair! I believe that window positioning also works on XWayland, though, so running your file manager that way should still work with the rest of the system being Wayland (and Gnome has no plans to drop XWayland afaik). I belie
26.
by ColonelPhantom 10mo ago
I think FTE is mostly used as a 'unit'. E.g. if two people work on something 50% of the time, you get one "FTE-equivalent", as there is roughly one full-time employee of effort put in. Though in this context it just seem
27.
by ColonelPhantom 10mo ago
Anecdotally, I strongly doubt this is true, although my environment is probably quite biased. I know a ton of people who use Gnome, some who use KDE, and I think roughly all of these people use them with Wayland. The standalone-WM users I k
28.
by ColonelPhantom 10mo ago
Rumors of Gnome's demise seem greatly exaggerated to me. It's still the default DE in nearly all major distributions, and it doesn't seem to have incurred major mindshare or marketshare hits recently. I feel like most of the
29.
by ColonelPhantom 10mo ago
Where is this Wayland black box then? If anything, Wayland made this situation significantly better: the X11 server was exactly this 'single point of failure black box' you are describing. Wayland replaces this with a much simpler
30.
by ColonelPhantom 10mo ago
I don't think it's true that anything is architecturally or fundamentally broken in Wayland (though if you disagree, I'm very curious what you think is so deeply broken). Most of the issues and slow adoption were because the

More ›