oivey

31 ms·

421.
by oivey 6y ago
I guess I don’t remember things the same. In my state, the most extreme restrictions (basically, everything is closed that isn’t remote or essential work) were lifted after a bit. I don’t recall ever being under the impression that restrict
422.
by oivey 6y ago
Most people on the internet are full of shit, and sometimes people in academia push unrealistic ideas. GP’s comment might be true (not my area of expertise), but HN definitely has a problem where software developers portray themselves as an
423.
by oivey 6y ago
In a few more months everyone will be vaccinated, significantly limiting the spread, slowing mutation. To be clear, we’re talking about the risk of the AZ vaccine specifically. It seems like there might be something specifically wrong with
424.
by oivey 6y ago
That’s not really the same thing since an individual who, for example, goes to large gatherings also endangers others by making essential activities like going to the grocery store more dangerous for everyone. Besides that, wider spread of
425.
by oivey 6y ago
Really, really depends on how you live. People in first world countries are worse for the environment than people in third world countries by a wide margin per capita.
426.
by oivey 6y ago
I was also confused. I agree that touching raw threads is usually not the right thing to do, and chains of parallel coroutines are one of the good abstractions. It’s crazy how few languages have easy access to that very basic abstraction.
427.
by oivey 6y ago
As pointed out in a sibling, processes have their own address spaces and aren’t as cheap to spawn as threads. I write code involving shared memory. It’s usable, but it also a pretty big pain to get right. It also significantly complicates t
428.
by oivey 6y ago
> As an observation, performance optimized code is almost always effectively single-threaded these days, even when using all the cores on a CPU to very efficiently process workloads. Not my experience at all. One big problem is that most
429.
by oivey 6y ago
Overpaid is really the wrong way to think about it. The wages in SV are driven by companies who are extremely profitable. Lower pay would mean more wealth for executives who are already some of the most wealthy people in the world (absent s
430.
by oivey 6y ago
HBO’s Chernobyl is more of a dramatization than a documentary. It certainly doesn’t contain interviews of people who were actually involved. There are quite a few good bits highlighting the dangers of nuclear energy, particularly in the con
431.
by oivey 6y ago
Referring to ML is so vague that it’s not particularly meaningful. The answer in general is no. There’s been some work showing neural networks can function with low precision floating point, but that’s still pretty different than fixed poin
432.
by oivey 6y ago
Yeah, but closer to the reals than fixed point or integers. That’s kind of the point of the standard.
433.
by oivey 6y ago
Lots of the math in games and ML assumes you’re operating in the reals. For example, the optimization in neural networks. Integers and fixed point won’t work.
434.
by oivey 6y ago
Ahh ok. I’ve had a similar conversation with a graphics guy that strongly paralleled this conversation, and I was confused then, too. Seemingly the asynchronous compute issues are a limitation when it only comes to shaders I guess? I write
435.
by oivey 6y ago
They’re leading right now by a bit, and when it comes to workloads involving tensor cores or ray tracing they’re burying AMD. I wish CUDA wasn’t vendor locked, too, but AMD and Intel need to start taking GPGPU software seriously. They’ve du
436.
by oivey 6y ago
Is that not what this is? https://docs.nvidia.com/cuda/cuda-c-programming-guide/index.... I thought this allowed kernels to run in parallel if both kernels could fit in the open resources. It doesn't have a p
437.
by oivey 6y ago
CUDA has had asynchronous command lists (multiple, actually, via streams) forever. What does AMD’s special hardware enable? Is it exposed in ROCm somehow? Edit: Maybe I should be more precise with my language. Using CUDA streams, you can qu
438.
by oivey 6y ago
This only seems to applicable to shaders? GPGPU in most contexts is referring to code divorced from any graphics pipeline, like CUDA or ROCm. CUDA has had asynchronous and concurrent kernel calls for a long time. How are asynchronous comput
439.
by oivey 6y ago
By async compute, do you mean something different than what CUDA streams expose? AMD’s compute cards have generally been worse than Nvidia’s as far as I know.
440.
by oivey 6y ago
Isn’t MAGMA more like a canned set of linear algebra routines rather than a general interface for executing arbitrary code on GPUs? That’s pretty different.
441.
by oivey 6y ago
In the small case here, there is no meaningful difference in speed between parsers. Using a quadratic algorithm has no advantage and is just an incorrect design.
442.
by oivey 6y ago
Rust is on LLVM for a reason: implementing new languages on top of GCC is hard. It seems much more reasonable for unusual architectures to contribute to LLVM. If companies want to sell their own architecture, they should provide the support
443.
by oivey 6y ago
At least part of the point was that C only appears to be portable. The abstract machine concept itself is leaky, lesser used compilers have serious bugs, build scripts aren’t portable across architectures, dependency management is poor, etc
444.
by oivey 6y ago
If you use “git add -p” it makes small commits pretty painless. I still like small commits in repos I work alone on because it makes reading the history during future debugging easier.
445.
by oivey 6y ago
Ultimately, I think the simultaneous complexity and lack of built in functionality will limit how often people end up using this new part of the standard. I can use coroutines to write high performance, parallel code in other languages with
446.
by oivey 6y ago
Eh, I think coroutines are a convenient way to achieve concurrency and parallelism. If you limit mutation and try to reduce the long reaching accessibility of variables then I think they’re generally very understandable, especially compared
447.
by oivey 6y ago
Part of this is that I’m tired, but it blows my mind how difficult C++ coroutines are as someone who considers themselves decent at C++ (although maybe I’m not) and uses coroutines in other languages. The amount of code needed to do almost
448.
by oivey 6y ago
For context, the cap is nearly 100x the normal price. In colder states power companies manage to be profitable at around the normal price of electric in Texas. The price of winterization isn’t that high, and a company that did it could have
449.
by oivey 6y ago
The point was often the plane would land somewhere that isn’t their destination, which would at least clue them into the fact that there was some significant mechanical issue.
450.
by oivey 6y ago
In the case of a contained engine failure, don’t planes immediately land at the closest safe location? Passengers would definitely know.

More ›