snvzz

6 ms·

61.
by snvzz 2mo ago
The LLM can run elsewhere, such as the cloud. For the part that runs locally, efficiency might matter.
62.
by snvzz 2mo ago
In seL4, a VMM escape, hard as that'd be, yields nothing. VMM handles all VM exceptions, and is just another user program. It has no higher capabilities than the VM itself.
63.
by snvzz 2mo ago
Eh, you'd have to pay me to use anything from Analogue. ModRetro put their effort where it matters[0], whereas Analogue aren't even honest about their hardware, with baseless claims of 100% compatibility, recurring shipping delays
64.
by snvzz 2mo ago
"open source" is the string that's missing in that announcement. The historical source code (AmigaOS git history up to at least when Commodore went bankrupt) should be MIT and in e.g. github. But that doesn't seem to be
65.
by snvzz 2mo ago
seL4 is open source and pretty small. What's stopping you?
66.
Why Were the Dabaos Late?(crowdsupply.com)
1 points by snvzz 2mo ago|0 comments
67.
by snvzz 2mo ago
No source code, no binary, just a page with some claims about code that might even exist. Eh.
68.
by snvzz 2mo ago
Censorship is now being called `guardrails`.
69.
by snvzz 2mo ago
RISC-V conveniently provides a `wfi` (wait for interrupts) instruction or hint -- it can act as a nop when not implemented.
70.
by snvzz 3mo ago
As of the recent launch of the ESP32-S31, all the ESP32 product lines have been migrated to RISC-V. Unsurprisingly, the old, non-standard, encumbered architecture is not missed.
71.
ESP32-S31 now in mass production and available for purchase(blog.adafruit.com)
3 points by snvzz 3mo ago|0 comments
72.
by snvzz 3mo ago
If anything, there's a strong argument to switch to seL4.
73.
by snvzz 3mo ago
To record audio, sample it with an ADC. Then store digitally.
74.
by snvzz 3mo ago
>For one, it had a forbid() system call which would stop the scheduler from pre-empting the current running process. This is, by the way, equivalent to SCHED_FIFO on Linux. There are very legitimate reasons a hard realtime task may need
75.
by snvzz 3mo ago
The CUDA addition (a vendor-specific API) is definitely a mistake.
76.
by snvzz 3mo ago
Sounds very Whataboutism. This is about the latter, and not just the correctable ones; Detecting them also has worth.
77.
by snvzz 3mo ago
>the thing about ECC memory is that it doesn't support reporting of ECC errors Wait that. It does, and that's the point. Although the CPU and motherboard do of course also have to support it, but with AMD it is possible. I do i
78.
by snvzz 3mo ago
Non-ECC memory. And into the trash it goes. Refer to https://news.ycombinator.com/item?id=48969530
79.
by snvzz 3mo ago
The keyword being latency. Even a powerful PC, such as running WinNT or Linux on my 9800x3d, doesn't feel as tight as my pretty basic Amiga 500 does. It's the difference between a RTOS proper (AmigaOS, QNX, BeOS) and a toy (Window
80.
by snvzz 3mo ago
Easier than ever, with RISC-V Vector (RVV), which is part of RVA23.
81.
by snvzz 3mo ago
Drivers run in supervisor mode as part of Linux, unfortunately.
82.
by snvzz 3mo ago
LLMs gave that one last push: Linux is not sustainable. We need to start looking at seL4 as the base, instead.
83.
by snvzz 3mo ago
RISC-V is inevitable. RISC-V is rapidly growing the strongest ecosystem.
84.
by snvzz 3mo ago
However, if on LineageOS, which mind you is the only way to get updated (and thus safer) android versions in many devices, they simply block you. Awful.
85.
by snvzz 3mo ago
I run a 9800x3d with 96GB ECC (2x 48GB DDR5-5800). It was a pain to select a compatible board and source the RAM but glad I did, in June 2024... the RAM prices went up immediately after. non-ECC should be a kernel taint in Linux. Newer vers
86.
by snvzz 3mo ago
>On 32-bit (and thats where density matters) it's sill worse then Thumb even with compression extension. Thumb is really good, but this is no longer true as of late 2021, when bitmanip and the code reduction extensions landed. RISC-
87.
by snvzz 3mo ago
>trying fix the code density issue with billions of extensions Not sure what you're on about. By the time the spec was first ratified (2019), RV64GC was already the densest 64bit ISA, and it's not even close.
88.
by snvzz 3mo ago
>and it's basically a joke. Look up the fab process node and die area of that chip, and think again.
89.
by snvzz 3mo ago
>Are there any RISC-V "machines"? that one can use to do something useful or fun with that someone here could recommend? Multiple boards based on the RVA23 spacemiT K3 are shipping as of recently. They are usefully performant.
90.
by snvzz 3mo ago
With or without DRM?

More ›