aseipp

27 ms·

301.
by aseipp 2y ago
Nvidia has decades of experience selling hardware to people with all the pains that entails, support, sales channels, customer acquisition, software, it's something you don't just do overnight, and it does cost money. Google'
302.
by aseipp 2y ago
Linux is behind other OSs at this point in terms of desktop security and hardening; there are of course ever-evolving mitigations for things like memory safety issues, but everyone has that, and more broadly the standard Unix permission mod
303.
by aseipp 2y ago
> I’m not that interested in hacking on the instruction set it runs on. This makes zero sense, frankly. Do you also feel this way about an x86 laptop, that x86 isn't worth learning anything about? Because it might run macOS? It&#x27
304.
by aseipp 2y ago
Honestly, I think the "stringly-typed targets" thing isn't too bad, having used Buck2 quite a bit, and being a Nix user for 10+ years. If anything, it's a small price to pay for some of the other smart things you get in
305.
by aseipp 2y ago
Nix does not solve the problem of "running things in a chroot." It solves the problem of having to reliably and consistently build software in a repeatable, reproducible, programmable way, up-to and including all the fundamental s
306.
by aseipp 2y ago
> Has anyone looked at creating a new object format for Linux? No, because that's overkill for this problem, because the de facto standard is "What do modern free Unicies do", where those are basically Linux/BSD&#x2
307.
by aseipp 2y ago
They use UEFI (like previous Snapdragons did) and EL2 is apparently unlocked on the SXE series (unlike previous Snapdragons.) It's mostly a matter of getting the device trees in place, I guess (I don't think these use ACPI nativel
308.
by aseipp 2y ago
It is not an impossible situation, courts have settled it, and what you describe is not how the law works (despite how many computer engineers think to the contrary.)
309.
by aseipp 2y ago
CUDA has been compatible with the open kernel modules for a while, in fact now that I looked it up, there's at least a handful of features that requires the open kernel drivers, including 1 CUDA specific one (DMABUF support): https:&#
310.
by aseipp 2y ago
FWIW, there is a port of EDK2 to the RK3588, which is what powers this board. Most of the peripherals work. So that's nice. But the thing is, it isn't UEFI so much as the device-tree/ACPI distinction you need to be mindful of
311.
by aseipp 3y ago
If you write a piece of malware that spoofs the secure desktop prompt, then you can steal the users password or whatever. But that does not grant you higher privileges. Only the secure desktop can grant your process higher privileges. The p
312.
by aseipp 3y ago
NVLink doesn't have as strict latency requirements as PCIe does. Stricter latency requirements mean you need more die area per GB/s of bandwidth. In practice this means that if you think of it from an mm^2 perspective, you can get
313.
by aseipp 3y ago
Well, spinning rust HDDs stuck in your server have no actual parallelism and aren't highly available, replicated, etc. How much do you think 1TB of storage should cost?
314.
by aseipp 3y ago
Hardly. This is just a typical double free vulnerability, AKA a normal Tuesday in the C/C++ world, and which people have been exploiting in the kernel (with its complex object lifetimes and semantics) for well over 10 years now. There&
315.
by aseipp 3y ago
The original commit is over 10 years old now, so it's probably lost to the sands of time. I did some looking but couldn't find anything on the old netdev lists. Maybe the patch was sent directly to Pablo Neira Ayuso, the committer
316.
by aseipp 3y ago
It's a traditional proprietary toolchain (about ~900MiB .tar.gz) that uses Synplify for synthesis and a custom fork of VPR for place and route, IIRC. It's reasonably easy to use and most of the UI flow can be scripted by Python, t
317.
by aseipp 3y ago
"Hard" refers to components that are actually part of the ASICs and exist on the die. For example, a "hard DDR memory controller" means that there is a physical part of the die that implements the DDR spec. "Soft&qu
318.
by aseipp 3y ago
I agree that hardware/software codesign are critical to solving things like this, but features like prefetching, speculation, and prediction are absolutely critical to modern pipelines and broadly speaking are what enable what we think
319.
by aseipp 3y ago
Hell, even just being able to use XFS would be an improvement, because ext4 has painful degradation scenarios when you hit cases like exhausting the inode count. (Somewhat related, but there has been a WIP 6.1 kernel for WSL2 "in previ
320.
by aseipp 3y ago
Because the ABI is defined and implemented by shared code in the runtime and compiler. There is no need for cross version compatibility between the two backends, only compatibility within the same version. This isn't particularly new e
321.
by aseipp 3y ago
My suggestion to students who want to hone their skills is to actually spend time programming with whatever tool they like. If, on the other hand, they want their skills to degrade -- perhaps they could instead try something like blogging a
322.
by aseipp 3y ago
Eh, in the past (maybe 5 years or more) THP was a lot fiddlier IME, and occasionally would hit bizarre cliffs in latency and overall performance. A lot of that has gone away. But because of that (among other things) you'd occasionally
323.
by aseipp 3y ago
You can use OpenCL just fine on Nvidia, but CUDA is just a superior compute programming model overall (both in features and design.) Pretty much every vendor offers something superior to OpenCL (HIP, OneAPI, etc), because it simply isn&#x27
324.
by aseipp 3y ago
It is an absolutely massive amount of work to turn something designed for your custom software stack and data centers (custom rack designs, water cooling, etc) into a COTS product that is plug-and-play; not just technically but also things
325.
by aseipp 3y ago
No there isn't. AWS does not use the traditional filesystem layer to store data; that would be a massive mistake from a performance and reliability POV; the POSIX filesystem specification is notoriously vague about things like fsync co
326.
by aseipp 3y ago
Supporting every Linux distribution and their small differences isn't free, and Linux distributions shipping things you haven't tested directly is also a way for users to get bitten by bugs or bad interactions, which they will the
327.
by aseipp 3y ago
In KDE's case the "you can just customize it" works both ways though. They could change the defaults and instead and let old users customize it back to the old way it was, rather than every new user customizing it. It comes a
328.
by aseipp 3y ago
They have accepted bugs from NixOS before (there are nearly a 100 marked as "NixOS Linux" at bugs.kde.org), and I don't see any major reason they would stop. It's true there are many patches for the KDE libraries and Pla
329.
by aseipp 3y ago
> All I'm asking is, if you want to eventually make money on your project, at least be up front about it in the beginning so that your users can make an informed decision when they decide whether to bake it into their stack. That is
330.
by aseipp 3y ago
> how am I supposed to recognise it as a bug or otherwise, when the code to call it hasn't even been written yet? You would just leave a comment on the change, and ask for the author to explain themselves? Or propose they fix it? I

More ›