Y
HN Search
Hacker News Search
new
|
comments
|
top
|
jobs
Panzer04
searching PlanetScale…
1.
▲
2.
▲
3.
▲
4.
▲
5.
▲
6.
▲
10 ms
·
121.
▲
by
Panzer04
1y ago
The problem with leading edge fabs is they require obscene amounts of capital to develop and are apparently (based on Intel's experience) extraordinarily risky to build. That's a hard thing to sell in any environment.
122.
▲
by
Panzer04
1y ago
This is just doomerism. Intel's current chips are "fine" and competitive with AMD chips. If anything, Intel is trying out more things than AMD is.
123.
▲
by
Panzer04
1y ago
Datacenter is not a monolith, and neither is it guaranteed to generate more profit or revenue. IIRC intel makes 2-3x in client sales what it does in datacenter.
124.
▲
by
Panzer04
1y ago
That seems like a huge burden, surely not? How often would a GC typically check for hanging references?
125.
▲
by
Panzer04
1y ago
That's not even the point they're really making here, IMO. The significant decrease they talk about is a side effect of their chosen language having a GC. This means the strings take more work to deal with than expected. This feel
126.
▲
by
Panzer04
1y ago
Because of a random anecdote on hackwrnews?
127.
▲
by
Panzer04
1y ago
Software. If you actually benchmark said chips in a computational workload I'd imagine the newer chip should handily beat the old M1. I find both windows and Linux have questionable power management by default. On top of that, snappine
128.
▲
by
Panzer04
1y ago
n^2 is probably the worst offender of these algorithms - it's fast enough to get into production and slow enough to blow up once you start using it. Rockstar infamously wasted five minutes loading GTAV because they had an n^2 algorithm
129.
▲
by
Panzer04
1y ago
100$/kwh on a battery that does 1000 cycles is 10c/kwh, 5000 cycles ("Claimed" lifepo4 these days), that's 2c per kwh. These aren't that unreasonable, albeit one would need to account for cost of capital and so
130.
▲
by
Panzer04
1y ago
The problem with Intel splitting is precisely that the fab is worth nothing. No one will give Intel real money for the advanced fabs, as they are worthless right now. I suspect that's probably the main thing holding Intel back from doi
131.
▲
by
Panzer04
1y ago
Intel, the CPU (and GPU) designer is doing fine. Intel the advanced IC manufacturer is worth effectively nothing right now. If they can prove a successful advanced manufacturing process and start making chips on it that will likely reverse
132.
▲
by
Panzer04
1y ago
The other things is steam doesn't constrain competition (afaik? Open to being wrong but this is how I'd understood it). Devs can sell their own games, games can be on other platforms, etc. Despite that gamers think it's worth
133.
▲
by
Panzer04
1y ago
Typically true. The more expensive the drive the more likely it'll be MLC or SLC, and if you're willing to go out of spec it's usually possible to reprogram a drive controller to operate in SLC mode.
134.
▲
by
Panzer04
1y ago
Assuming pcie SSDs are acceptable, they could have bought an optane drive with 60 drive write per day for five years endurance XD SSDs today have lower ratings because very few (consumer) use cases would prefer more write endurance over +
135.
▲
by
Panzer04
1y ago
Are you joking? You bought 30 64GB SSDs from 2010 in 2021? You realise 3k buys you much faster, much higher capacity modern SSDs with the same longevity characteristics? Hell, if it really matters there's still some optane floating aro
136.
▲
by
Panzer04
1y ago
It's interesting how you can practically tell how many people have a specific processor nowadays (12 core = 5900,7900,9900x, 14 core = i5 14600k, 245, 20 core = i7 etc) Pretty cool
137.
▲
by
Panzer04
1y ago
Dual rank was a lot faster in DDR4 days IIRC, because it allowed a memory channel to have more requests in-flight on the memory - it could be worth like +10% performance or more. Less impactful in DDR5 because they made it a design goal to
138.
▲
by
Panzer04
1y ago
SK Hynix is winning in DDR5 land - their memory performs significantly better in all respects compared to Micron and Samsung. Whether that matters much is debateable - maybe they get higher yields as a result (since more chips are sufficien
139.
▲
by
Panzer04
1y ago
Cheapest to the consumer, maybe? What's the actual financial standing of said companies? Plenty of government-owned utilities require lots of propping to cover their losses.
140.
▲
by
Panzer04
1y ago
It's not just about better numbers. Getting high clocks for a short period helps in a lot of use cases - say random things like a search. If I'm looking for some specific phrase in my codebase in vscode, everything spins up for th
141.
▲
by
Panzer04
1y ago
OP exaggerates the difficulty. You can trivially parallel the vast majority of lithium batteries so long as their voltage is reasonably close (I personally wouldn't fuss much over a 100mv difference, or even more in most cases, unless
142.
▲
by
Panzer04
1y ago
There's a difference between acceptable standards for a truly professional product that you'd have to certify and hobbyist stuff. Beyond short circuit/overcurrent, overtemp, over voltage, under voltage protections, what else
143.
▲
by
Panzer04
1y ago
Fundamental difference when it comes to performance. In terms of implementations, obviously arch matters a lot in terms of the "established" software ecosystems, but if you're writing some software and need to hit a performan
144.
▲
by
Panzer04
1y ago
There's no fundamental difference between risc, arm and x86. It looks like when all else is equal (fab, die size) the chips tend to have about the same performance characteristics - so why mess with what's working?
145.
▲
by
Panzer04
1y ago
Could not agree more. Have never used a chargeback and wouldn't want that ability in the vast majority of transactions. If I'm buying something on steam I know what I'm getting and they already have a robust refund capability
146.
▲
by
Panzer04
1y ago
Single point of failure? Yeah u can download the entirety of Wikipedia if you want to. What's the single point of failure?
147.
▲
by
Panzer04
1y ago
Just build a second Falcon Heavy. If there was a genuine possibility of SLS being competitive, that's one thing. It's another when it's worse in basically every way.
148.
▲
by
Panzer04
1y ago
If robotaxis were commensurately cheap and available I'd happily dump my car. I don't see it happening for another decade or more, though.
149.
▲
by
Panzer04
1y ago
Given cases like Elon moving to less protective jursidictions, those protections are not necessarily as protective as you might prefer, especially if you sign them away at some point in the past. It's going to be a lot harder to protec
150.
▲
by
Panzer04
1y ago
There are a whole lot of companies that are run to benefit their management and not their shareholders. At best, they might be making no money but keeping a bunch of people in work. There's a good reason public shareholders historicall
More ›