Poor lighting is no laughing matter.

Not sure how well that without work even if you could bridge them properly to share their vram, as the latency from the other gpu will be pretty high compared to the local vram. Frame times won’t be that great is my prediction. It works ok for LLMs because they aren’t a realtime compute task like gaming graphics is. If an LLM takes an extra 0.12s to compute its result, you don’t notice, but if a gpu misses a frame deadline by 0.12s, that’s a stutter that represents less than 10 fps.

I believe that’s why earlier attempts at dual gpu systems mostly fizzled out (plus cost concerns). You can get some pure acceleration if you can fit the entire working memory onto both GPUs’ vram (so your total vram is effectively min( GPUA_VRAM, GPUB_VRAM ) rather than GPUA_VRAM + GPUB_VRAM, though that also requires all pixels to be independent of anything calculated on the other GPU, other than maybe post processing effects that could be handled on whichever GPU is handling the display.

That’s not to say that you can’t get extra performance out of multi-gpu setups that don’t just mirror their RAM, but it’s more complicated than “sum of the capabilities of each GPU”.

It’s also the main reason I haven’t touched the epic store even for the no strings attached (I think?) free games. They started out wanting (and getting) exclusive games before pivoting to free games, showing they were all too happy using business bullshit rather than good competition to take market share, which also implied things about what they’d do if they got to a dominant position.

Real dumb ones.

I spent more time “tinkering” on windows after some updates to figure out how to set it back to how I want it after some update silently resets things to default or moves some setting for no good reason. And then there’s the “hey please use our shit software that you went out of your way to install an alternative to… no? Ok, I’ll ask later” full screen ads, ads in the start menu, ignoring default search engine or browser settings when it could. It was a hostile user experience.

Linux has been such a better experience that windows shills like you are a complete joke. Though interesting that you used hedging language here and talk about “people saying they do more tinkering than actual gaming” rather than “people doing more tinkering than actual gaming”, almost as if you know you’re full of shit but are trying to avoid outright lying for some reason.

Personally a big thing that holds me back from even seriously considering mobile gaming as an extension to normal gaming instead of “the best you can do when you don’t have access to better devices” is the input side. I hate touchscreen controls for anything with real time action because touching the screen obscures it from view, plus you don’t get the tactile feedback of buttons that makes it trivial to play without needing to readjust your finger positions constantly.

And I know there are external controllers you can use with phones, but if I’m making space (and remembering to bring) one of those, I might as well just bring my steam deck or even the switch instead.

It’s funniest when they act like their rage is justified. Or destroy their own things.

If you read the negative reviews on steam, you can often filter these shitty ones out. Also, be suspicious by default of free games. There are some that were made by some game dev just having fun but most want to make money some way or another and f2p monetization is often pretty toxic, like adding a long grind just to sell ways to avoid it or p2w mechanics.

I haven’t felt like I’ve gone wrong by generally just buying the cheapest well-reviewed games on my wishlist, occasionally stretching the budget to grab a mid-priced game I’m interested in and rarely going for a AAA priced game. Means that I get to some great games later (eg, Hollow Knight is one of the more recent ones, highly recommended btw) but it’s been a great way to experience a wide variety of games from today and the last 15 years for not that much money.

Oh crap I forgot my testing hammer, I’ll be right back.

That’s assuming it’s being banned to stop it from happening rather than being banned to give an excuse to punish those who do it.

My not very generous guess is someone wanted to give the impression of something being really old without a clue of what really old is.

Like it wouldn’t surprise me if it turned out bread was accidentally invented when someone screwed up an attempt to make beer.

Maybe about the same sway but they are purchasing memory on a different market that is about future goods rather than goods currently sitting on shelves. So they will have more day to day visibility on RAM futures than your average consumer will. Price increases will happen there before they affect store shelves.

Doh, no such luck for me. Oh well, I still have my PS2 lol.

Python code:

results = {}
for ip0 in range(256):
  ip0_str = to_str( ip0 ) + '.'
  for ip1 in range(256):
    ip1_str = ip0_str + to_str( ip1 ) + '.'
    for ip2 in range(256):
      ip2_str = ip1_str + to_str( ip2 ) + '.'
      for ip3 in range(256):
        ip_vector = [ ip0, ip1, ip2, ip3 ]
        ip_str = ip2_str + to_str( ip3 )
        results[ ip_vector ] = ping( ip_str )

Might look a bit nicer using format strings instead. That map will contain on the order of 4 billion entries (one for each value of 2³²), and the actual size will depend on what format the ping function returns, 4 bytes for each key (optimized from my initial version that used the IPs as keys), plus all the internal structures for the map like key hashes and the hash table itself. Ie, this takes more memory to run than viewing OP’s full sized image and I wouldn’t suggest running it with less than 32GB of RAM. Though it would take less memory if it generated the image directly or at least making the keys implicit (which the image does, as they are encoded into the x, y coordinates rather than stored).

Edit: Let’s look at runtime, too, because why not. Assuming every single IP responds in 0.01 seconds (they’ll take longer, especially the ones that time out instead of respond), rounding the total to that nice 4 billion number get us 40 million seconds. An hour is less than 4000 seconds, so it would take over 100 hours to run this script.

Though it could be parallelized, since you can ping many targets at once. Not sure what the maximum number of pings you could have in flight is, but whatever it is, you’d be much better off using a script that did like 80% of that (to leave some margin for the rest of the system to use, also ISPs might not be happy with you maxing out your ICMP traffic).

Or more specifically, what is the sum of all latencies divided by the number of responses times the total numer of requests sent (to scale up for the ones that didn’t reply, assuming their average latency would have been similar to the total average)?

I mean, it’s a binary: either you like it or you don’t (plus the neutral option). The ratings in the meme were portion of people who liked it and didn’t. So you will either always fall with the majority or minority (or won’t care if you feel neutral about it). There is no wildly inaccurate for that, there’s just how many others you agree or disagree with.

Assuming their sample population is representative, at least, because what it says about the population can be inaccurate (but “my friend group all agrees with the minority side” doesn’t necessarily mean it’s inaccurate for the pop since your friend group probably doesn’t represent the entire pop).

Don’t feel too bad, Emu, Elmo made a similar mistake last week.

You have too many negatives there.

Should be “I didn’t say she did say they didn’t.”

Though… You also didn’t say that she didn’t say that they didn’t (that was the other commenter) so I guess you’re still accurate.

Remember when they started adding that shit and you could get kids who were running it kicked off the internet any time you said “fuck”?

Or standardize agreements that can be agreed to with just clicks so you can tell what’s up from just seeing which agreement modules are included and what parameters they have (along with a bunch of sites that explain them in easier terms for the less literate, which will be useful because they are standards used by many agreements rather than needing a unique one for each version of each document).

Hit the reset button then the power button, as the bios usually handles that interrupt by just shutting down immediately (and CPUs reset into low performance modes, so it reduces the power on your CPU immediately).