volta83

25 ms·

151.
by volta83 5y ago
You have never built a super computer, have you? You have to connect thousands of cables, have hundreds of nodes, with thousands of components, everything interconnected, and if you connect one wrong, the computer outputs incorrect results.
152.
by volta83 5y ago
The fact that they didn't do this: > Their current training cluster would be the 5th largest supercomputer if Tesla stopped all real workloads, ran Linpack, and submitted it to the Top500 list. which is trivial to do, and pretty muc
153.
by volta83 5y ago
So how do banks verify identity in the US ?
154.
by volta83 5y ago
FAANG
155.
by volta83 5y ago
These are guidelines that the standard library aims to follow. It does not blindly follow them, and they only represent the values and trade-offs that make sense for the standard library. If your projects have different values and trade-off
156.
by volta83 5y ago
> If my 4K structure can be copied, should it derive `Copy`? This leads to users implicitly copying it around instead of using references more aggressively. Rust eliminates unnecessary memcpys, but this is a bit up to you. `Copy` as a sy
157.
by volta83 5y ago
The EU system is decentralized.
158.
by volta83 5y ago
> It is illegal for the US government to create a mandatory national identity system The system can be voluntary. If you don't need an SSN today, you wouldn't need to use that system. If some bank or health care provider only a
159.
by volta83 5y ago
There might be more data for levels.fyi in the US than EU. 600k$ is 480k EUR. I know SEs in Germany doing more than 400k EUR at FAANG.
160.
by volta83 5y ago
> In Europe, you go, where? Same place? All of FAANG and beyond has offices in Europe.
161.
by volta83 5y ago
I still fail to see the value in filtering out candidates by whether they know this rule that still nobody has given any actual arguments for. "If you Google this string, you'll find the arguments" is a logical fallacy. You a
162.
by volta83 5y ago
Good point! I guess the difference is whether it exists for the player that does the second move.
163.
by volta83 5y ago
Since you are openly advocating for people to blindly follow this rule… can you at least explain _why_ do you think this rule makes sense and people should follow it. When I hire, I couldn’t care less about whether the person uses I or We.
164.
by volta83 5y ago
I tried to make it win every time and I'm wondering whether given a Tic-tac-toe opponent that does random moves, it is always possible to play such that your opponent always wins.
165.
by volta83 5y ago
> How many GPUs would NVIDIA sell in the HPC market if it wasn't for CUDA and the various support libraries around it? We don't have to hypothesize about this. The HPC GPU market has multiple vendors, just look at how many HPC
166.
by volta83 5y ago
This is not really a strategy, but rather a consequence of LLVM having hundreds of developers, most of which use LLVM upstream, and therefore don't care about maintaining an API compatibility that they themselves do not need. It turns
167.
by volta83 5y ago
The making of is pretty interesting: https://www.youtube.com/watch?v=1qhqZ9ECm70 They apparently created all slides of the keynote with omniverse in a couple of weeks, and used a lot of technologies for no other reason than
168.
by volta83 5y ago
> but they needed to do something to address the CSAM issue lest governments come down on them hard. If Apple does not have decryption keys for iCloud content, they can't decrypt it, only the user can. For Apple, it's not techn
169.
by volta83 5y ago
We do this on all our Rust code as well (1.1 million LOC at this point). While we can benefit from #[no_std] crates on crates.io, unfortunately we can't use any crates that require standard because the standard library does not propaga
170.
by volta83 5y ago
I use btrfs snapshot with arch just for this, and they've saved me multiple times. Its like Apple's Time Machine, but also covers system files.
171.
by volta83 5y ago
> a formally verified spec A formally verified X is just a reliable X. If X is wrong, then you just have something that will reliably do the wrong thing.
172.
by volta83 5y ago
So? I don't see how focusing the available manpower on one implementation instead of splitting it over 10 different implementations would make this worse. I do see how it would make this much better. What one implementation lacks, anot
173.
by volta83 5y ago
+1 thanks for the correction, should have verified this. Last time I used ISPC it was using an extremely outdated LLVM version (5) or so, and it took years to move it to LLVM 7. I wasn't aware that they have been slowly migrating to ne
174.
by volta83 5y ago
> Don't want to end up with every C/C++ compiler being a clang skin in the same way as (almost) every browser is a chrome skin. Why not ? There is one open C++ spec, that's really hard to implement, so we have a dozen C++
175.
by volta83 5y ago
What that sentence actually means is "we don't want to help our competition (AMD, NVIDIA) but we don't want to re-implement the wheel". The API and fundamental part of LLVM is its IR, LLVM-IR, which is where most optimiz
176.
by volta83 5y ago
> Even if they pay the same tuition as masters or bachelors students this is usually around 800-900 CHF per semester, AFAIK, as long as you are registered as a PhD student, you have to pay the fees. That's easy to pay if you are an
177.
by volta83 5y ago
The student partner was living in Wuhan. The partner beg them to stop, and they didn’t. Its partner is lucky not to be sent to a “re-education” camp. We all agree that the fact that this can happen is wrong, but what the student did is not
178.
by volta83 5y ago
Probably this also saved the student some fees as well.
179.
by volta83 5y ago
This is funny because everybody makes mistakes , so of course people fail at semver. However, if you are the maintainer of a popular cargo package, and make a mistake, people _will let you know_. Because a lot of people in the Rust ecosyste
180.
by volta83 5y ago
What's the consultancy firm ? indra ?

More ›