Hacker Newsnew | past | comments | ask | show | jobs | submit | XelNika's commentslogin

> production 1.7GHz quad-core Intel Core i7-based 13-inch MacBook Pro systems with Intel Iris Plus Graphics 645, 16GB of RAM, and 2TB SSD

https://www.apple.com/macbook-pro/#footnote-4

So yes, that is compared to a very old 14 nm design, presumably the i7-8557U per Wikipedia.


Your comment implies that it’s obviously not this spec that they compare against. Could you spell it out for the ignorant like me? What about that config makes it definitely not the thing that is 86x slower?


I don't see anything in the GP that implies that. It's simply a CPU that was released before an entire AI economic bubble was a twinkle in Jensen Huang's eye. Of course it has piss-poor AI performance vs something with hardware dedicated to accelerating that workflow.

It's not that the comparison is incorrect, just that it's a silly and unenlightening statement, bordering on completely devoid of meaning if it weren't for the x86 pun.


They'll be the only people running this thing in 2030 so they can produce 286x and 386x and maybe 80286x performance gains by then.


They have price limits, but I think GP means volume alert limits. Someone unexpectedly bids a third of the Finnish production capacity, you call them and ask if they are sure.


There are many market participants, some of whom start trading on this information immediately. The Finnish grid is connected to the rest of Europe so a market participant could in principle flow this cheap energy from Finland all the way to Spain if the cables had capacity. I would guess it was too late the moment it happened.


You don't even need to worry about what someone can physically provide. All bids by market participants can (arguably should) be checked for extreme outliers. A bid of multiple GW from a participant that usually only trades hundreds of MW should immediately set off alarms. It only takes a phone call like "hey, are you sure about this?"


With Zig and Rust you have to explicitly opt-out with `ReleaseFast` and `unsafe` respectively, that makes a big difference. Rust has the added safety that you cannot (to my knowledge at least) gain performance by opting out with a flag at compile-time, it has to be done with optimized `unsafe` blocks directly in the code.

Lazy C++ is unsafe, lazy Zig is safe-ish, lazy Rust is safe. Given how lazy most programmers are, I consider that a strong argument against C++.


[dead]


>It has nothing to do with opting out.

It does. The original code compiled because the borrow is computed using `unsafe`. That `unsafe` is the opt-out.

>Zig, Rust and no language saves you when you write incorrect unsafe code. My original point is disqualifying c tools is misleading and everything suffers from incorrect unsafe code

And the other people's point is that if one language defaults to writing unsafe code and the other language requires opting out of safety to write unsafe code, then the second language has merit over the first.


Sure, but unlike C/C++, in Rust my entire codebase is not enclosed in a gigantic unsafe block


> Performance topped out at 111 MiB/s (931 Mbps), which is suspiciously close to 1 Gbps.

That's because of overhead in TCP over IPv4. You're testing the payload throughput, not the physical throughput. The theoretical maximum performance without jumbo frames is around 95%.

https://en.wikipedia.org/wiki/Jumbo_frame#Bandwidth_efficien...


Ah, good to know. Thanks!


Intel Alder Lake delivered around a 15% improvement to single-core performance gen-on-gen so it's just you.


It's not the 40% year of Moore's law...


Moore's law is about transistor density, not single-core performance. Even if a new generation of chips obeys the "law", there is no requirement that the designers dedicate the improvement to single-core performance. Alternatives include multi-core performance and miniaturisation.


Yes, it's not good.

> That way any other people on his network wouldn’t have a problem, either.

Given that he's setting DNS servers statically on his own clients and has working DNS resolution on the DHCP-provided DNS server(s), that shouldn't really be an issue for other people.


While the project is super interesting, this seems like network horror to me. Not only is something "wrong" with the network (guessing dns.msftncsi.com is blackholed), the author is setting DNS servers statically on his clients instead of using DHCP. If I'm right that the domain is simply blocked, I think it counts as yet another horror that the author did not realise that before buying a replacement WiFi card. It pains me that the blog post never explains the root cause of the issue nor whether it was intentional.


There are also plenty of other wifi connection test domains [1][2] around that he didn't have to buy one.

Apple returns "success" 200, opensuse returns 304 (no-content) and may close the browser tab instantly so you have to fiddle in console to see it working.

  await fetch('http://conncheck.opensuse.org',{mode:'no-cors'})
[1] http://conncheck.opensuse.org

[2] http://captive.apple.com/hotspot-detect.html


> There are also plenty of other wifi connection test domains [1][2] around that he didn't have to buy one.

Strictly speaking, I have enough garbage domains that I could have used any one of them, but when I found a cute short name I grabbed it. ... Hosting it myself means my computer doesn't tell someone additional where I am all the time ... Although since the requests are plain http, that doesn't really mean much.

Also note:

    C:\> curl http://captive.apple.com/hotspot-detect.html | xxd
      % Total    % Received % Xferd  Average Speed   Time    Time     Time  Current
                                     Dload  Upload   Total   Spent    Left  Speed
    100    69  100    69    0     0     69      0  0:00:01 --:--:--  0:00:01   489
    00000000: 3c48 544d 4c3e 3c48 4541 443e 3c54 4954  <HTML><HEAD><TIT
    00000010: 4c45 3e53 7563 6365 7373 3c2f 5449 544c  LE>Success</TITL
    00000020: 453e 3c2f 4845 4144 3e3c 424f 4459 3e53  E></HEAD><BODY>S
    00000030: 7563 6365 7373 3c2f 424f 4459 3e3c 2f48  uccess</BODY></H
    00000040: 544d 4c3e 0a                             TML>.
That LF at the end means now I need to figure out how to get that into the registry (maybe it's trivial but it takes an additional ΔT and ΔW.

As for the opensuse one, I am not sure if the check works with empty content. Again, not hard to figure out, but additional ΔT and ΔW.

Plus, these sites can change whatever they do at any time and if I customize to match them, then I am at their beck and call as opposed to controlling both sides of the equation.


The Wi-Fi upgrade story linked in the beginning is also a slightly weird solution. Rather than buy an mPCIe to m.2 adapter and 1st party Intel (or any modern) card he bought a half mini adapter and rebranded half mini Intel card for twice the price.

Not to say there is anything wrong with a stream of consciousness blog about what one has had fun doing doing lately rather a heads up since this is being shared these fixes are just that, not high effort recommendations or whatnot normally posted on HN.


> for twice the price.

Prices vary over time. Upon seeing this, I checked the current price and it is almost twice as much as I paid at the time and the item is listed as out of stock by this seller.

Indeed, there is a 8265 listed at $21[1] and a converter listed at $8[2]. At the time, I did not know such converters existed.

[1]: https://www.amazon.com/Intel-Dual-Band-Wireless-Ac-8265/dp/B...

[2]: https://www.amazon.com/M-2-PCIe-Converter-Bluetooth-2010-201...


Fair point, maybe some differences due to location as well as it shows in stock for me.

FWIW the mpe-AX3000H is a repackaged ax200 not an 8265 (ac only chip). Doesn't change the price just noting it's a different chip than the 7260/mpe-AH3000H https://www.amazon.com/Compatible-AX200NGW-9260NGW-Upgrade-W...


> setting DNS servers statically on his clients instead of using DHCP.

Yes, because these personal laptops end up being used in coffee shops or hotels on occasion. So, whatever I do just on my home router doesn't help.


"On his clients" meaning his eg laptop. Personally my laptop uses its own preconfigured DNS servers, even (especially!) at coffee shops/etc because the coffee shop's DNS isn't to be trusted but more than that, they are frequently extremely slow. (It does take some fussing if there's a captive portal, but that's easy enough to handle.)


> some fussing if there's a captive portal

`example.com` comes in handy.


That is fair, I suppose, but why use a different DNS server than the default for your home network? I still think there's something fundamentally wrong with a DNS configuration that breaks Windows connectivity tests.

I do hope you're using something encrypted. Plain DNS can be redirected and manipulated quite trivially.


Yes, it says so in footnote 1 in the article. It's also not the benchmarked code so it doesn't really matter.


I missed the footnote. My point was more that one could have compared this with the C implementation as heapq.merge is a different algorithm and so there is no actual comparison of the same algorithm in C and Python.


Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: