loeg

21 ms·

571.
by loeg 7mo ago
I brought a graph in my earlier comment! Even over the past year, they're highly correlated. (And the S&P500 is ahead over the period -- not the other 25%.)
572.
by loeg 7mo ago
I don't disagree with any of that.
573.
by loeg 7mo ago
Market cap. If it had the P/E of MSFT, it would still be 140x bigger than the smallest S&P500 component.
574.
by loeg 7mo ago
The broader US market is only about 25% bigger than the S&P500, FWIW. (Or put another way, S&P500 is about 80% of all US equity.) They also trade in almost lockstep: https://totalrealreturns.com/n/VFIAX,VEXAX?
575.
by loeg 7mo ago
It is a crazy meme stock, but in terms of the S&P500, it's 2000x the size of the smallest S&P500 component.
576.
by loeg 7mo ago
Again, this is just a weird workflow, and you're assuming copy/edit don't bump mtime. That usually isn't the case. If you're doing this weird thing, you can just run `touch` when you move files over existing files
577.
by loeg 7mo ago
Copy (1) and edit (2) both bump mtime, usually. It's not obvious that in the workflow you describe ninja is problematic, rather than the workflow itself (which is atypical).
578.
by loeg 7mo ago
The ratio is that bad anyway.
579.
by loeg 7mo ago
The F14/Abseil maps are included and use cute SIMD tricks, FWIW (discussed by the blogspot author).
580.
by loeg 7mo ago
CPUs and GPUs sit in pretty distinct niches; they don't substitute like you're implying.
581.
by loeg 7mo ago
> Under a new decree approved by President Lee Jae-myung during a Cabinet meeting on March 11, mid-to-large-sized public parking lots with 80 or more spaces must install solar power generation South Korea is going to get a lot of 79-spac
582.
by loeg 7mo ago
I'm not sure C++ provides a more satisfying answer here than "don't use this with a T that throws in copy." (And also, why would you want that?)
583.
by loeg 7mo ago
Men play volleyball with a higher net to ameliorate this problem.
584.
by loeg 7mo ago
Boys are stronger than girls even pre-puberty (~10%). https://onlinelibrary.wiley.com/cms/asset/bd2689e4-7094-4e0b... https://onlinelibrary.wiley.com/doi/full/10.1002/ejsc.12268
585.
by loeg 7mo ago
I think he's complaining that because the head_ and tail_ loads in push/pop are relaxed, rather than also being acquire, they can be reordered relative to the acquire tail_ and head_ loads respectively. I don't believe this
586.
by loeg 7mo ago
I don’t understand what you’re getting at.
587.
by loeg 7mo ago
I believe ConcurrencyKit's impl does this. https://github.com/concurrencykit/ck/blob/master/include/ck_...
588.
by loeg 7mo ago
There are real practical implications of both the producer and consumer mutating the same cache line to take a lock that is fundamentally avoided by this "lock-free" design. It isn't meaningless.
589.
by loeg 7mo ago
ConcurrencyKit ck_ring. The SPSC macros are the most similar to this article: https://github.com/concurrencykit/ck/blob/master/include/ck_...
590.
by loeg 7mo ago
> In reality multiple threads for single core doesn't make much sense right? Not necessarily, I think -- depends what you're doing.
591.
by loeg 7mo ago
> write with release semantic cannot be reordered with any other writes, dependent or not. To quibble a little bit: later program-order writes CAN be reordered before release writes. But earlier program-order writes may not be reo
592.
by loeg 7mo ago
> I think the key difference is that they only need to coordinate when the reader and writer are close together. This was already the case with the cached index design at the end of the article, though. (Which doesn't require extra
593.
by loeg 7mo ago
You ultimately need a mask to access the correct slot in the ring. But it's true that you can leave unmasked values in your reader/writer indices.
594.
by loeg 7mo ago
This was, in fact, mentioned in the article.
595.
by loeg 7mo ago
For ultra low-power embedded, wouldn't a mutex approach work just fine? You're running on a single core anyway.
596.
by loeg 7mo ago
This is a SPSC queue -- there aren't multiple writers to coordinate, nor readers. It simplifies the design.
597.
by loeg 7mo ago
Yeah, or you could put a generation number in each slot adjacent to T and a read will only be valid if the slot's generation number == the last one observed + 1, for example. But ultimately the reader and writer still need to coordina
598.
by loeg 7mo ago
> There's no relationship between the two written variables. Stores to the two are independent and can be reordered. The aq/rel applies to the index, not to the unrelated non-atomic buffer located near the index. No, this is in
599.
by loeg 7mo ago
LL/SC is still hinted at in the C++11 model with std::atomic<T>::compare_exchange_weak: https://en.cppreference.com/w/cpp/atomic/atomic/compare_exch...
600.
by loeg 7mo ago
Or just eat an extra few minutes of charging time once or twice a year; it's simply not a big deal. Charging at home saves me so much time relative to getting gas that the occasional road trip wait is already overcompensated for. ICE

More ›