I think you missed an important point in the parent comment. You can override the allocation for C++ coroutines. You do have control over details like allocation.
C++ coroutines are so lightweight and customizable (for good and ill), that in 2018 Gor Nishanov did a presentation where he scheduled binary searches around cache prefetching using coroutines. And yes, he modified the allocation behavior, though he said it only resulted in a modest improvement on performance.
I only ran the benchmark on Quasar Alpha*; the rest of the scores come from the original paper [0] which was published before 3.7 was available. This is a pretty expensive benchmark to run if you're paying for API usage - I'd actually originally set out to run it on Llama 4 but abandoned that after estimating the cost.
* - I also reproduced the Llama 3.1 8B result to check my setup.
> Initially, the FSB was mainly interested in getting its hands on our equipment, presumably in the hopes of confirming its hypothesis regarding our links with the CIA.
> ...
> The FSB officers feigned surprise and promised to return the missing items immediately but succeeded in doing so only after 40 minutes had passed. Although they failed to bypass the pin codes on the phones or computer, the Kremlin’s agents did manage to install a tracker on Roman’s laptop, which he discovered within minutes.
> I had a weirdly similar experience shortly thereafter, not in Moscow, but — shockingly — in Berlin. Flying back from a screening of Navalny in New York and on the way to another one in the Hague, I was just passing via the German capital for a few hours to speak at a conference. The event was held at a pompous hotel in the city’s suburbs.
> ...
> During the event, I looked up the ownership of the hotel only to discover it was owned by a German, quite literally, “friend of Vladimir Putin”. I rushed out to get my suitcase, and the bellboy took a whopping twenty minutes to find it. On the way to the airport, I discovered a hard disk was missing from the suticase. I alerted the police who rushed to the hotel, only to be told that the security cameras had been down for maintenance.
Does this imply that the conference was held at this hotel purely to get access to his devices?
> The scheme was replete with cars bearing fake license plates, a route that avoided traffic surveillance cameras, and two speed boats that would need to be sunk at the end of the operation.
> Later on, a source in the Security Service of Ukraine (SBU) informed Roman that Kyiv’s intelligence services had gathered information showing that a Ukrainian criminal group had received an “order” from Moscow to kidnap him and take him to Russia. A reward of $50,000 was offered for his capture.
> At one point, one of them even booked a seat next to him on a flight from Budapest to Berlin, wearing a hidden camera to record his screen while he texted me. Their attempt to get his smartphone pin code was off by only one digit.
How many people were working full-time to get this guy?
It does sound crazy. If they really wanted to kidnap him, they would have succeeded instead leaving so many clues. The SBU told him? I am shocked, shocked.
Getting hands on electronic equipment is what any pro-Palestinian Western journalist is familiar with at airports etc.
My suspicion is the same as yours, that this may have been caused by local ISPs being overwhelmed, but it could be a million other things too. I had network issues. I live in a heavily populated suburban area. I have family who live 1000+ miles away in a slightly less populated suburban area, they had no issues at all.
This is how it was for me growing up blue collar in the northeastern USA in the 80s. My father fixed everything in the house and the vehicles. I inherited my older siblings clothes, and my younger siblings inherited mine. My mother would hem pant legs shorter when we were young, and then let them back out as we grew older. If you wore a knee or an elbow out of clothes, it was getting patched.
This instilled some good and bad tendencies in me. I do almost all of the repairs around the house myself. I work too much though, so I don't always have enough time or energy. Even though I can easily afford it, I have a hard time paying someone else to do them. This means I live with broken stuff longer than I should.
I'd probably have more money if I spent that time working on side projects instead of doing maintenance and repairs.
> This is how it was for me growing up blue collar in the northeastern USA in the 80s. My father fixed everything in the house and the vehicles. I inherited my older siblings clothes, and my younger siblings inherited mine. My mother would hem pant legs shorter when we were young, and then let them back out as we grew older. If you wore a knee or an elbow out of clothes, it was getting patched.
Thing is, they were able to in the first place.
Forget about fixing a modern car. The electronics side is a mixture of "a datacenter on wheels", DRM and anti-tamper technology (sometimes enforced or heavily suggested by law such as in emissions control, sometimes by reality, e.g. "Kia Boys") and high-speed protocols instead of early age wires and relays that you could troubleshoot with a decent multimeter. The physical side is a ton of plastics designed to absorb crash energy and finely tuned metal alloy stuff (with the form also having crash safety implication) that your average DIY person cannot reasonably weld instead of plain old steel sheets. You can't buy a "reasonably repairable" new car any more because of the legal mandates and because you don't want it to be stolen by some kid having watched a YouTube or Tiktok video showing how to bypass the locks.
And clothing... patching a 1980s piece was possible, the fabrics had weight and structural integrity of their own. Nowadays it's extremely thin fabric everywhere that shreds itself after a few washing machine cycles. Try to patch it and you'll more likely than not find out that your very act of pushing a needle through it to apply the patch just causes the next rip to appear. You are still able to purchase better quality clothing technically but you end up paying like 4x the amount and it's still made in some Bangladeshi or Chinese sweatshop under horrible safety and employee rights standards.
If one has the proclivity, then: One can get into rather far into troubleshooting and (and ultimately repairing) common modern automotive electronics with an Autel rig that, adjusted for inflation, costs less than an Atari 2600 did.
I feel like it’s giving good money away when you hire someone to do work for you that you know you can do. You look at the markups for things and it gives you pause. Things like a valve or whatever. You can go to the Home Depot and get it for cheaper even when you include the cost of whatever tools you need to get.
But at some point you have to say, let’s just get someone to to it (the deck, the fence, the gutters, etc.) still it’s like zen and the art of motorcycle maintenance. There is some personal satisfaction in being self reliant.
It would be absolutely devastating to me too. I only do my gaming on Linux, and my large game catalog is through Steam. Which means that even if they can't really shutdown Proton because it's open source, it still might amount to destroying my game library.
I guess it depends on what level you're generating the events at. On Linux, it would be completely reasonable to inject the input events at the input device level.
This is very straightforward (EV_REL) and requires a very small amount of code. There can be different problems to deal with when working at this level, but in my experience, everything works as expected with keyboards, mice, and gamepads.
That's the thing, it really depends on what you're trying to accomplish. If you're trying to move the mouse as if some remote program was a mouse attached to your computer, generating inputs makes sense. If you provide some kind of remote support application that just needs to make the mouse appear at the place the remote tech indicates, changing the raw cursor makes sense.
Both approaches are reasonable and both are implemented in desktop operating systems for this reason.
This is really interesting for me to read. I encountered a DMA lockup in the hardware by an Ethernet MAC implementation on an ARM chip. It was a Synopsys Designware MAC implementation. It would specifically lockup when PTP was enabled. From my testing, it seemed like it would specifically lockup if some internal queue was overrun. This was speculation on my part, because it would only lockup if I tried to enable timestamping on all packets. It seemed to work alright if the hardware filter was used to only timestamp PTP packets. This can be a significant limitation though, as it can prevent PTP from working with VLANs or DSA switch tags, since the hardware can't identify PTP packets with those extra prefixes.
The PTP timestamps would arrive as a separate DMA transaction after the packet DMA transaction. It very possibly could have been poor integration into the ARM SOC, but your PTP-specific issue on x86 makes me wonder.
Group policy settings will keep it off. Can be set if you have a pro version of windows or a business/enterprise version. It's the only way I will trust using MS Edge.
Has anyone here has used FreeNX? I had read good things about its performance over the network (and even cell networks), but I've never encountered anyone who's actually used it. I guess I'm going to have to be the person to try it so I can be the "person who's used it" to others. And to those wondering, I think this is relevant to this link, as I believe FreeNX is more-or-less the X11 protocol, but does much more caching and compression.
I used to use it back in the day. It had amazing performance — the first Remote Desktop that showed me what was possible, and far and away the best way to do remote X.
C++ coroutines are so lightweight and customizable (for good and ill), that in 2018 Gor Nishanov did a presentation where he scheduled binary searches around cache prefetching using coroutines. And yes, he modified the allocation behavior, though he said it only resulted in a modest improvement on performance.