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

.. which confirms all of my stereotypes. Looks like the AWS engineer who reported it used a m8g.24xlarge instance with 384 GB of RAM, but somehow didn't know or care to enable huge pages. And once enabling them, the performance regression disappears.

Because such settings aren’t obvious to those not familiar with them. LLMs should make discoverability easier though

Honest question: what's the value of running the benchmark and reporting a performance regression if the author is not familiar with basic operation of the software? I'd argue that not understanding those settings disqualifies you from making statements about it.

The performance was reduced without a settings change. That is still a regression even if huge pages mitigates the problem.

I'd be curious to know if there's still a regression with hugepages turned on in older kernels.

If you are benchmarking something and the only changed variable between benchmarks is the kernel, that is useful information. Even if your environment isn't correctly setup.


Some software clearly wants hugepages disabled, so it's not always the slam dunk people seem to be making it out to be.

ie Redis:

https://redis.io/docs/latest/operate/oss_and_stack/managemen...


It’s more of a hostage situation.

Those were transmitted offline so they didn't have authentic NVENC H264 compression artifacts. Never before have you seen it with 260 Mbps ;)

/s


To me, the whole thing sounds like cheating in benchmarks.

Intel built a tool that will only activate for a specific benchmark - but not for real-world software which accomplishes similar things - and then the tool will replace generic bytecode with a (most likely) handcrafted and optimized variant for running this specific benchmark on this specific CPU. That means BOT will only boost the benchmark score, but not help at all with the end-user workflows that the benchmark is trying to emulate. Thereby, Intel's BOT makes the benchmark score misleading, which is why Geekbench is flagging them.


This is what I thought, too. Watch them dig a hole for themselves while you take a vacation. And then when they fail to debug the mess, charge them by the day for your expertise, because it’ll be necessary for untangling that mess.

Fully agree. The low-res makes this unnecessarily hard to navigate. Which is a downside if your core gameplay is to teach players to navigate in a challenging environment.

I guess this implies high quality WebRips for years to come, because blocking Intel Laptops isn’t feasible for any streaming service. And from what I heard, control over SGX would let you extract the WideWine etc. key store without!!! the OS knowing about it. So this is like a secret debug access starting at OS&TPM level.

Widevine L1 has never been used on Windows machines anyways. PlayReady SL3000 works on Windows machines, but that’s broken already anyways.

Use AI! It'll make you a 10x engineer! (cost-wise) /s

I've recently had the displeasure of Opus 4.6 hallucinating an API. It would have been great if that API had existed, but it did not. Still, it then looped until I manually terminated it while trying to make tests pass. In my case, I used up about $12 of usage in 30 minutes. My guess would be mostly through the (pretty verbose) thinking tokens.

But it's not just Anthropic. I had the same issue with Gemini 3.1 Pro.


While I believe this article to be correct, Germany is an interesting counter-example:

Germany first introduced mandatory child car seat laws on April 1, 1993. [1]

That year, fertility was at 1.28 kids per woman. Since then, it has increased to 1.62.

[1] https://de.wikipedia.org/wiki/Kindersitz

https://www.macrotrends.net/global-metrics/countries/deu/ger...


You can’t compare those numbers because the population in 1993 and today comprises different groups who are materially different in terms of fertility rate. Last year, the fertility rate for women with German citizenship was 1.23.

The other major change is that in 1990, you had a reunification of east and west germany. Fertility rates in East Germany were low before reunification and collapsed right after reunification. But they recovered from the early 1990s to the late 2000s. So the 1993 aggregate average is artificially low. In neighboring France, the fertility rate in 1993 was 1.7.


Could be as if you live in a city in Germany you don't necessarily need a car?

But I would be surprised if car seats are the main driver here in any case.


Certainly not as much as the U.S or Canada, but not as little as Tokyo or Netherlands. I found various cities easy to navigate without one, but they seemed oddly car-centric as well. Berlin is very car and very public transportation somehow

Public transit and kid/stroller public transit are very different beasts.

Once you’ve had to lift a stroller into a “high” bus you’ll not want to do it anymore.


> Once you’ve had to lift a stroller into a “high” bus you’ll not want to do it anymore.

I imagine I'd agree, but is that related? You might have more niche knowledge that I'm not aware of in this context. In my city in Canada, all buses lower upon request


Those who have children can’t afford a car.

"Wouldn't the real cause of the depressed birthrates be the requirement to own a car in order to have children?"

Yes. The one-time setup costs for "properly" raising kids are probably around $30k. All the kids stuff is extra expensive (in the west) and for the kids seats you need a large car (in the west) and there's social stigma against kids sharing a room (in the west), so you also need a larger apartment.


Kid stuff is free if your time is worth nothing. There’s always another child outgrowing their car seats and toys and clothes and cribs and games.

Can confirm, the "stuff" costs basically nothing in the scheme of things and nearly all of it can be had used. A bunch of it's also not really all that necessary. Clothes and toys can all be had for very little, without even that much time investment, folks are drowning in this stuff and lots of it just gets thrown away.

The real money goes to:

1) Healthcare (in the US).

2) Childcare or foregone wages.

3) School/housing location (same thing; either tuition, or spending 20+% more for the same amount & quality of house in a nicer school district [and the ongoing cost of servicing the extra mortgage on that]; you can skip this, but if you can at all afford it, you'll not feel like it's optional)

4) Space. Larger housing and larger cars. You can skip this kind of (larger car is less-optional if you have more than three kids) but at significant cost to QOL.


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

Search: