sylware

28 ms·

661.
by sylware 1y ago
Last time I heard about that it was for "super computers": nearly or even faster than the alternatives with a massive energy consumption advantage.
662.
by sylware 1y ago
This is exactly what could imped LLM training dataset in the western world, which will mechanically lead to "richer" LLM training dataset in countries where some PI is not walling that data for training. But then, the countries wi
663.
by sylware 1y ago
I don't recall properly (because I did selve my mapping projects for the moment), but don't openstreet map core data distribution format based on protobuffers?
664.
by sylware 1y ago
For tons of software out there, but not all, development time is minuscule compared to the life cycle.
665.
by sylware 1y ago
More real life C compilers is always a good thing. But remember: to be able to build a linux kernel, you will need zillions of gcc extensions... (I have a sweet spot for the alignment attribute of stack variables in printk). That said, &quo
666.
by sylware 1y ago
I have 32 bits support on my x86_64 gaming ring, _ONLY_ for the steam client. The "steam client" is still a 32 bits ELF executable, which statically loads openGL and x11 libs... (namely not even a wayland->x11 fallback or a ope
667.
by sylware 1y ago
If you could avoid to post direct X/twitter links: they do not work with noscript/basic (x)html browsers. Thx.
668.
by sylware 1y ago
Oh, I remember I did a plain and simple C port of an old gperf, cgperf https://www.rocketgit.com/user/sylware/cgperf Ofc, I did add my own bugs.
669.
by sylware 1y ago
Weird, I was thinking the same abouth the 1kW chips, but for GPUs instead of CPUs. Actually, I guess we were _really_ many thinking "1kW" ? I wonder if "1kW" will stay niche, presuming all issues which the 1kW barrier w
670.
by sylware 1y ago
It is amazing that risc-v is getting performant large implementations in 'a market' completely dominated by others (with extremeley toxic PI locks). It is really refreshing to see that people with significant resources are trying
671.
by sylware 1y ago
follow the money... and you'll really know.
672.
by sylware 1y ago
I remember, nokia about to release the first smartphone, full linux with a clean plain and simple C written OS/platform... and... then nothing happened. I recall the user groups being puzzled on what was waiting nokia, they were ready,
673.
by sylware 1y ago
They can exploit local priviledge escalation flaws without "RCE". And you are right, kernel anti-cheat are rumored to be weaponized by hackers, and making the previous even worse. And when the kid is playing his/her game at h
674.
by sylware 1y ago
I forgot about that: all the nice game binaries from them running directly on nearly all systems...
675.
by sylware 1y ago
Won't help: I get scans and script kiddy hack attempts from digital ocean, microsoft cloud (azure, stretchoid.com), google cloud, aws, and lately "hostpapa" via its 'IP colocation service'. Ofc it is instant fail-to
676.
by sylware 1y ago
ucloud ("based in HK") has been an issue (much less lately though), and I had to ban the whole digital ocean AS (US). google cloud, aws and microsoft have also some issues... hostpapa in the US seems to become the new main issue (
677.
by sylware 1y ago
omg... that's why my self-hosted servers are getting nasty trafic from SC all the time. The explanation is that easy??
678.
by sylware 1y ago
Finally fixing gov web interop? Aka noscript/basic (x)html browser support for critical online services (at least)? They have to keep in mind, 99% of the time won't be writing web stuff... but keeping the sites available and safe,
679.
by sylware 1y ago
If they manage to plug their microarch design on RISC-V ISA (yes, they will throw away a ton of things), they will be ready, performance-wise. This real hard part is transitioning the software stack, including games...
680.
by sylware 1y ago
I would say otherwise. The future, if sane, is certainly not with a PI locked ISA like ARM all over again (look at x86). Actually, it looks like a super bad move from intel.
681.
by sylware 1y ago
? RISC-V ?
682.
by sylware 1y ago
oh, you are new to HN, because you would not need to ask such question if you were reading HN in the last few years... You can start on risc-v wikipedia page and/or on the official risc-v web site.
683.
by sylware 1y ago
Whitelisting seems to be the way to go. With IPv6 and OS generated IPs (up to what the ISP domestic router allows) could be very efficient.
684.
by sylware 1y ago
It should be RISC-V... who is in charge at Intel?? Is this related to the rumors of softbank (ARM) money injection in Intel?
685.
by sylware 1y ago
Yep, look at nasm SDK, horrible, with tons of perl generators, etc, etc. From the worst.
686.
by sylware 1y ago
You could use GAS, FASM2, or write an ffmpeg specific one: writting a real-life assembler is orders of magnitude simpler than writting a real-life compiler... Usually that implementation complexity is very dependent on the macro pre-precess
687.
by sylware 1y ago
There is serious abuse of nasm macro-preprocessor. Going to be tough to move away to another assembler.
688.
by sylware 1y ago
But depends on the whatng cartel web engines... Nothing to be proud of, mate.
689.
by sylware 1y ago
Come on, with a bit of irony, they have their first major bug: they are pulling a c++ compiler (either gcc or clang).
690.
by sylware 1y ago
Huh? This is not related to the LICENSE, it is related to the technical dependency on a whatng cartel web engine (geeko/blink/webkit).

More ›