zrm

82 ms·

751.
by zrm 8y ago
The amount of L3 per thread is still 64MB/#threads. Where the difference you're describing most matters is for single-threaded code, where in theory the one thread could otherwise have the entire 64MB. But that isn't the circ
752.
by zrm 8y ago
There are a lot of things that have to go the wrong way at once to get to the point where that really matters. The first is that your working set won't fit in the processor caches and has regular cache misses into main memory -- but mo
753.
by zrm 8y ago
This isn't Bulldozer anymore. Intel has faster single thread performance on some workloads, but the difference for "most workloads" is modest and for the same money AMD is offering more cores, more memory channels and more I&
754.
by zrm 8y ago
Couldn't this also be done for the other reactive components (especially nitrogen and oxygen) before cooling the gas? Removing ~99% of the mass ahead of time would require that much less energy for cooling.
755.
by zrm 8y ago
You would get all kinds of interesting packet reordering issues, which can cause performance problems all their own. I've sometimes contemplated just faking the TCP protocol entirely. Do the TCP handshake, then send "TCP packets&q
756.
by zrm 8y ago
> I wanted to check out Mastodon and looked into setting up a node. It's a nightmare of rube goldberg machine dependency hell. It's not that I can't, but that I don't have time to mess with it. The sad thing is that w
757.
by zrm 8y ago
They make two main points: > By implementing the SSA with a kernel module, developers who wish to use it do not have to link with any additional userspace libraries. It's not clear why this is a major problem to begin with, but sock
758.
by zrm 8y ago
Then why not move the user-space code into the original user process as part of the standard library?
759.
by zrm 8y ago
The concern is that TLS is a complex protocol and if there is a bug in the code then the attacker gets access to the kernel instead of only a user process. There is also the problem of conservative distributions using very old kernels. If y
760.
by zrm 8y ago
If you're going to build something like this, especially when the data sets are typically large, this is a very strong case for erasure coding. If you split a 5GB file into (20) 250MB chunks and then mirror each of them four times, t
761.
by zrm 8y ago
> I'm not very familiar with DNSCurve but I don't think it has any mechanisms to prevent amplification. The DNSCurve page discusses amplification: https://dnscurve.org/amplification.html The main point is that
762.
by zrm 8y ago
DNS-over-QUIC is a plausible alternative to DNSCurve, but you're still looking at a fallback to TCP TLS when you encounter a middlebox which is interfering with all the UDP traffic it doesn't outright block.
763.
by zrm 8y ago
> The transport protocol picked for the DNS, the User Datagram Protocol (UDP), was the best option at the time, but is fraught with issues, including being a major ingredient for DNS-based DDoS events. Only recently has attention turned
764.
by zrm 8y ago
> The MAC is just an example. You need some way to block someone abusing your connection. You're thinking like a sysadmin. Think like an organization. Compare the situation where you have a public space where everyone is welcome exc
765.
by zrm 8y ago
> You'd need to also identify which device was infringing by getting a connection time/destination. I still don't see where it says you have to do that. Your link doesn't seem to say anything about it. I question the
766.
by zrm 8y ago
I feel like this is why the advice is always to consult an attorney. If the law has some easy to fulfill requirement (service policy) then concerned people should have one even if they're only providing access to Uncle Bob and not the
767.
by zrm 8y ago
> As long as you're legally responsible for the traffic coming out of your network, this is not a good thing to do. Unless people explicitly get the same protection an ISP gets, I'll keep advising them to not to share their con
768.
by zrm 8y ago
Even for that, it's not the ideal layer. A basic connection should generally be available for everyone even if it's a rate-limited logically-separated segment that only provides internet access. Then if you want special treatment
769.
by zrm 8y ago
Sometimes I wonder if trying to encrypt WiFi is even worth it. "E2EE or GTFO" is pretty compelling. The counterargument is keeping packet headers (i.e. remote IP addresses) and plaintext DNS queries private, but that's alread
770.
by zrm 8y ago
Because names are scarce and squatting produces inefficiency. Someone who wants to make productive use of the name has to pay a middle man in order to do it, which reduces their resources available to do the useful thing and unjustly enrich
771.
by zrm 8y ago
> But the problem then is that getting enough ethernet ports in the thing to equal the average router is price prohibitive. Unless you actually need the ports to do some kind of network segmentation, one solution is to just plug the insi
772.
by zrm 8y ago
So there are two main issues with consumer routers. The first is that the hardware is garbage. This isn't universally true, but it's a strong general rule, and models get released and discontinued all the time so the short list
773.
by zrm 8y ago
Then why is the effort not to detect/reinitialize modifications to the BIOS or IME instead of this more complicated alternative?
774.
by zrm 8y ago
> If that's the case, why doesn't the NSA publish Suite A algorithms? The math on whether you find the vulnerability before somebody else does is very different when you employ as many cryptographers as the NSA. They also have
775.
by zrm 8y ago
> I consider it worse, since it's too easy for people to become content with it. It's not just that. For a given vulnerability, there is an amount of time before the good guys discover it and fix it, and an amount of time befor
776.
by zrm 8y ago
The angle and layout of the 4000 is different. A lot of people prefer the original.
777.
by zrm 8y ago
How about bringing back the original version of this: https://en.wikipedia.org/wiki/Microsoft_Natural_Keyboard Just take the 1994 version and remake it in USB, with no other changes whatsoever. Thanks.
778.
by zrm 8y ago
> Seen that way, the defense has to be against remote code execution, not against cross-process memory reads. This doesn't always look like what it is. You have games that allow e.g. custom levels with map editors, and then the user
779.
by zrm 8y ago
What you're describing is equivalent to a dead drop, which is an old school anonymity method. It's not that it doesn't work, it's that it's completely independent of Tor and it doesn't work in all the contexts
780.
by zrm 8y ago
I think you're referring to traffic analysis for protocol identification, which is something else entirely, and Tor handles that as well. Resisting protocol identification can be done very efficiently -- you basically get the same info

More ›