temac

64 ms·

601.
by temac 7y ago
Not so sure given the current temporal context we are discussing about. Paul Graham has some interests in Lambda School (YC S17), has also personally promoted it (e.g. https://twitter.com/paulg/status/1175434253741
602.
by temac 7y ago
The mistake is being nullable/optional by "default", that is with the least amount of effort for programmers using such a language. Or worse only ever nullable (like Java is except for its built-in scalars I think?). There is
603.
by temac 7y ago
That's part of the worst writing I have read about people. Writing a complete armchair analyst pamphlet indicating among other things than some people are "less than a man" reflects rather poorly on the author. I hope I'
604.
by temac 7y ago
It's not only about the constant factor, but also about the fact that O is only valid toward infinity, and potentially about how computer actually work (e.g. the extreme discrepancy between the latency of L1 vs RAM)
605.
by temac 7y ago
The quite detailed root cause (technical and even economical) was already known after the first crash. So if that's "capitalism" working that it needed a second one, I actually don't like "capitalism" very much
606.
by temac 7y ago
> Airbus may eventually have that problem with the A320 family. I really don't know. But they don't have it yet. Airbus is already fly-by-wire on the A320, and they know how to do it not too stupidly. They can provide the sam
607.
by temac 7y ago
I've got a huge issue with all the stories about the "software" being supposedly at fault. The problem was never about the software per se; it was about the avionic specs on system level. You can do very complex things withou
608.
by temac 7y ago
Measurement of what? You don't have to measure logic. Of course you can be unscheduled for any reason in userspace, of course you can be rescheduled with completely arbitrary latency esp. compared to targeted wake up, and of course pol
609.
by temac 7y ago
Not in every cases, but both are certainly suboptimal compared to a diretected wake up.
610.
by temac 7y ago
> #StoreRelease memory barrier even on Intel What is that? x86 is TSO... Do you have an example of the full acquire and release sections in x86 assembly?
611.
by temac 7y ago
Completely fair locks can have extremely terrible throughput performance, so depending what you are doing with them, they might not be a good idea...
612.
by temac 7y ago
Does it really sleep for a specified period instead of doing directed wake-ups? If so that's very far from ideal...
613.
by temac 7y ago
A spinlock in the kernel typically only spins. You use it for the cases when you can't schedule... So it better will be for a short time only. But the concept of short time does not even exist deterministically in userspace, usually, b
614.
by temac 7y ago
That's not extremely low power compared to real low power states. The main advantage of PAUSE is the scheduling of the other hyperthread (if it exists) and maybe not generating a gratuitous L1 / MESI workload at a crazy rate (well
615.
by temac 7y ago
> France does not extradite its own citizens. By saying that they wouldn't extradite him they have just repeated the legal position. Inconsequential nitpick: I'm not even sure this is a position formalized by anything, IIRC it
616.
by temac 7y ago
Also, a yielded thread has absolutely no reason to be woken up with low latency, and this has nothing to do with the supposed scheduler quality. I'm not sure a kernel blindly prioritizing otherwise cpu-bound thread just because they yi
617.
by temac 7y ago
Besides OS where they are often needed, spinlocks can have their place in a few very specific (and most of the time isolated on dedicated computers), but I agree that too many people are trying to play with them now, without enough knowledg
618.
by temac 7y ago
If you are thinking about the fear of contraband imports of yellowcake, well none of that has ever been found and its highly unlikely to have existed. The removed yellowcake had been stockpiled prior to the first Gulf War, and was declared
619.
by temac 7y ago
There is no multitasking for regular programs in DOS: if a program launches another one, the parent is simply paused until the child exits. Also DOS used various interrupt handlers (not just the synchronous int 21h)
620.
by temac 7y ago
I believe pocket and especially sync do nothing if you don't use them. This is in absolute contrast with Chrome that ties your whole navigator to your google account if you ever log on gmail...
621.
by temac 7y ago
> Bloomberg > After being annihilated after the story’s publication, Supermicro’s stock has bounced back. You can quickly form a theory that would be interesting to investigate...
622.
by temac 7y ago
How would hackers later have fun if you made a chip that works all the time?
623.
by temac 7y ago
Your source for the belief tha Iraq had WMD is an article explaining they had yellowcake? Even a youtuber made some and processed it (and this was completely legal)
624.
by temac 7y ago
> The idea of the generic CPU may very well become seen as a luxurious, wasteful idea once Moore's law properly runs out of road in a few years. Specialization will breed new ISAs, even boring ones. This is already the case and I su
625.
by temac 7y ago
The problem has not been solved mainly by the store though. It has been solved by sandboxing.
626.
by temac 7y ago
I was trying to cite a typical full L1 hit latency... I thought store-load fw simply avoid having to flush the complete write buffer before the access is even possible, which risk to take far more than ~3 cycles. Now maybe it can be faster
627.
by temac 7y ago
The Linux approach reflects the social structure this project has been developed in. The Linux project needs to be able to evolve independently of other projects, so they do just that. This is a classic case of technical architectures follo
628.
by temac 7y ago
> I don't know. Neither do others outside of the project. > I find these baseless micro-aggressions against Go missplaced and unfruitful. Hu? Ok then Go is perfect because it is developed in secret. We are discussing here, I&#x27
629.
by temac 7y ago
If you have to decompress it at startup, you will typically do it to anonymous memory. You can attempt to be fancy at user-level by trying some silly tricks like putting them to shared memory, although I don't know the API of common on
630.
by temac 7y ago
Yes and the security mainly comes from the sandboxing, not from some hypothetical stores of websites making random judgments (also I also know that the reputation and blocking of website could depend on the presence of hostile JS, but given

More ›