yeah but you can do a lot if you don’t have to allow for supporting extremely slow storage (the downside of which, as we’ve seen in recent years, is that a lot of PC stuff isn’t even gonna be QA’ed below sata 3)
Nvidia should be able to get below 10w/tflop with these cards if they continue to focus on fp32 in their gaming lineup which is definitely very cool but personally I think I will wait until there’s another non-fp32 feature I’m excited about like fixed function av1
The nVME drive is exactly what’s making me excited about the PS5
I kinda feel like super optimized io is the only place where they can cost effectively offer a performance advantage over a pc at this point. Building a specialized cpu or gpu isn’t feasible anymore, so what’s left?
This is Digital Foundry in 2019 analyzing Microsoft discussing using SSDs as virtual ram which is what Mark Cerny was saying was going to revolutionize PS5 graphics.
Tbh I think that if you can raise the bar on a piece of hardware that has been holding back software design on PCs (supporting disc drives) for a decade, then I think consoles can actually do something great. Fingers crossed
I wish every emulator had an auto updater because trying to get anything off github always makes me feel like an idiot
Very happy with how easy upgrading the wifi on my motherboard was.
Almost like, PowerMac nice on an Asrock Z370.
I am getting a rock solid connection except when I walk 5 meters away on Bluetooth so I’m looking at longer antenna leads to place them somewhere besides against the wall behind a bunch of metal
There it is, Macs are switching to ARM processors.
I guess from our POV the main impact is that the historical Steam library is probably toast. We can’t expect games with like one Mac purchase a month to recompile and resubmit.
EDIT: well, now they’re showing an install/startup-time autotranslator, and specifically highlighting a game demo for it.
Safari is also getting support for extensions made for other browsers, and a dedicated extension store.
hmm
the interesting thing is that both apple and microsoft have been demoing first-order virtualization support for not just linux CLI stuff but also GUI apps in recent months too so I wouldn’t be surprised if that’s their compromise for wanting to maintain such an arbitrarily high standard for “actively maintained software” on their platform.
I’m really excited about this by and large, they already frontloaded a lot of the potential downsides in recent years by deprecating OpenGL and dgafing Vulkan, deprecating 32-bit x86 stuff and requiring onerous signing, etc. I really dislike x86 essentialism and seeing as Apple has already made it impractical for many of their users to not own a second, non-Mac computer in recent years, I think it’s mostly upside from here.
(I lost maybe 10% of my macOS game library from the 32-bit deprecating, not more)
Virtualization performance and feature parity has been poor enough in the past that its main use on the client-side so far has been companies forcefeeding their bad software to employees. That said, Apple has way more resources, incentive and OS/hardware control to do it well than anybody ever has previously, so I’m cautiously optimistic that the Rosetta thing won’t be too noticeable except in increased startup or install times.
I’m not sure I understand what would be beneficial about this. it seems like their goal is to make the mac desktop experience more like ios, which doesn’t make sense to me at all. like… the ios versions of flagship creativity apps like photoshop are dogshit compared to the mac os versions. am I misreading something about this situation? who is this change for?
yeah you’re misreading it completely, they’re just changing the desktop architecture so all the existing apps compile on arm
I understand that but I’m seeing it spun as basically “ios for macbooks”:
and that seems… dumb
they are merging platforms very very gradually (rumouring xcode for iPad Pro which would finally let you actually compile and ship code from there albeit in a limited capacity, allowing people to ship iOS codebases on Mac) which I think is very low on the list of value adds from this but for whatever reason people are sensitive to it because microsoft farted around with it for years and also because people don’t like having to own multiple devices in theory
I’m trying to squeeze every drop of life I can out of my Pixel XL and it’s holding up admirably thanks to OS updates but I am oddly compelled by the Surface Duo
It’s been a while since a novelty smartphone has caught my eye at all 
I don’t know if you know this or felix is assuming you do but arm processors run circles around intel per watt. They’ll have a form factor and battery life advantage because their hardware is integrated. Their current notebooks are all thermally constrained because they refuse to design around intel’s cooling needs. This is the main reason.
Microsoft tried to make an arm-powered surface device with qualcomm and qualcomm doesn’t have the design advantage apple does.
microsoft’s platform value is also much more dependent on permanent compatibility with decades of x86 software, whereas apple actually can plausibly say “we’re going to make everyone update and implement a decent virtualization solution for the rest”
I mean I hope so
I mean I totally get the technical reasons why this could be advantageous, but I find the overall strategy that it seems to reflect to be kind of alarming. I feel like apple has been taking macbooks and mac pros in a consistently boneheaded direction since like 2016 in terms of product and interface design and making them “more like ipads” or whatever seems to be consistent with that trend, regardless of the undeniable performance advantages of this architecture shift. it just seems to be kind of insidious! but we’ll see I guess.
The truly baffling thing is their refusal to put a touch screen on any of these computers when they expect iOS apps to come over unchanged
yeah apple has been doing a lot dumb shit pretty much for as long as intel has been stagnant but as long as they have the best mobile platform by far this is a step in the right direction (given how much they’ve already been blowing up to lay groundwork)
