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

You can game just fine with a modern iGPU these days, depending on what you play. You don't always need VRAM.

More channels or soldered memory? Channels are basically RAID 0 so it depends what you're measuring. Soldering memory down was the only way to use LPDDR5X so if you wanted the best memory you had to solder it down. LPCAMM2 exists though so newer devices can use that instead of soldering them down, but not all devices would be able to fit the required LPCAMM2 slots.

SOCAMM2 allows for removable ram in nearly 0 added space.

That's tiny! But it still depends a lot on form factor. LPDDR5X is used in phones and making memory removable would have its compromises. You may even have compromises in laptops. Look at how tiny a MacBook Air's mainboard is and you'll see the RAM modules on the same package as the SoC. SOCAMM2 is too large for that but a variant with only two modules could possibly work.

I bet it could fit. It just do a combo of soldered ram and non soldered like many laptops used to.

Yup, the only reason Macs have higher memory bandwidth is because they use more memory channels, which gives them a wider bus. Both Intel and AMD only allow more than dual channel memory on server class processors these days.

The “Apple only does X better because they do Y” thing has been a meme for ages. I remember dismissals like “PowerPC is only faster at math because it has more integer units” or something along those lines, and thinking, uh, isn’t that a good thing?

There's a qualitative difference between "they're doing a different thing" and "they're doing the same thing, tuned differently". GP is saying that this is a case of "they just tuned it differently".

This distinction doesn't change what the performance numbers look like today, but it does inform what changes would be necessary for those numbers to look different tomorrow. E.g. Apple Silicon isn't fundamentally orders of magnitude more efficient than x86, they just used smaller features. Newer Intel and AMD chips made on equivalent processes _also_ get similar efficiency gains.


There are AMD and Intel devices on similar process (not talking about A20Pro or M6 which are set to ship later this week), and they do not get the same gains.

And honestly, they have historically had different markets.

When the design is for only one customer, you don't need to generalize things, and those things you generalize to give different customers different options has costs.

AMD will soon be a larger customer for TSMC than Apple (NVIDIA is already there) so Apple's pre-booking new processes is likely to be gone in the near future.


Depends on the expense trade off.

If I get 10% more performance for 50% more cost it really depends on one's needs, for example.


The point is that "unified architecture" is a buzzword unrelated to what actually makes things fast.

It does make some things faster because you can share memory between CPU, GPU, etc. without copying. But not everything.

"You're only better than me at sports because you practice more and try harder!"

I am just saying it's not magic and x86 is capable of doing the same. Quad channel memory used to be more common in consumer hardware but now it looks like you don't even have the option anymore for desktops. AMD's Strix Halo was the first sign to reversing that (it has quad channel memory!) and hopefully we see more of that in the future.

AMD does market segmentation and limits consumer chips to dual-channel. You need to cough up the dough and get Threadripper for quad-channel. Or even more for Threadripper Pro to get octa-channel.

As I mentioned above AMD's Strix Halo has quad channel memory and is consumer level. But yes, other than that everything is segmented away.

Afaik steam deck is quad channel, despite it being a pretty low end chip (with a decent GPU though)

Kind of but not really. DDR5 splits your typical 64-bit channel into two 32-bit subchannels meaning the bus width is not increased. These subchannels are not always advertised because it's a just a feature of DDR5. Actually adding more channels increases bus width, which is what meaningfully improves memory bandwidth.

Threadripper have that extra bandwidth and M5 still is faster

Can you link a specific benchmark?

Keep in mind that non-Pro threadripper is still only 256 bits wide and Pro is 512. And the memory is 30% slower than with an M5. So an M5 Ultra has 3x the memory bandwidth of the best threadripper.


Don't some AMD full server CPUs have 12 memory channels ?

Yes, some do. So similar bandwidth to a Max but with lots of slower cores. Half as much as an Ultra.

Except strix halo.

The memory bandwidth is a small thing compared to the massive win you get by not having to move data between two memory pools at all.

Depends on your workload. And AMD has supported unified memory long before Apple Silicon existed anyway.

Often you don't move memory around as a programmer, but that's exactly what happens in the background.

It's the address space that's unified, not always the physical hardware.

The data movement (when needed) is handled transparently in the background by page faults and other tricks.


Moving memory around is a bottleneck. Non-unified memory usually means copying over the PCIe bus which is way slower than RAM (and way way slower than VRAM). Actually unified memory means you don't need to copy anything at all though which is the absolute best case for performance.

It's so much easier to switch browsers than platform/OS. Unless you have an iPhone or iPad.

This is for a $1 microcontroller. I'm assuming you're talking about the Raspberry Pi computers based on the $50 cost and root.

It's a $1 MCU, which will get embedded into a $50 device. If you're a company selling a cloud-plus-device product, then that 250k tooling cost and not-exactly-trivial attack process will be quite effective at stopping Chinese clones: with a unique per-device key there's no way they'll sell enough units to make a profit before you will inevitably ban their cloned key.

Well not chinese clones, those are well funded factories with state su subsidies, but it will definitely stop rogue individual hackers.

It's not even 250k cost, just 250k capital, it can be rented.


To be fair Apple of all companies have the best shot at pulling it off. They've been perfecting their hardware security for years for other reasons and this is just another way to take advantage of that work. But yes, if someone breaks it then the trust is gone and it casts doubt on all of the photos that were ever captured using the broken system.

I'm curious what specific capabilities you're referring to here.

I have an outback with AT tires and I can fit my surfboards inside. I don’t want a compact. Rivian R2 is tempting but esspensive.

I replaced my old Forester with the new Trailseeker. Stsrting price (and interest rate) is much better than the R2.

I can slide in an 8’ board no problem when I’m too lazy to use the roof rack.

I haven’t upgraded the wheels or tires yet but the stock ones have had no problem going up and down the old logging and forest service roads of NW Oregon and SW Washington.


Thank you for the rec I’ll take a look!

If the Toyota sienna had a plugin ev option that would be really compelling for me. The electric van market is surprisingly sad. I currently only boogie board because it fits in my car.

I’m really tempted by current sienna hybrids, my current gas one gets 20mpg on a good day. Plug-in for the daily school commutes and after school activities would be great

The r2 is right on the median price for new cars in the US, believe it or not. Crazy. But yeah, an Outback is still cheaper

Subaru dealers near me have the Trailseeker, an EV comparable to the Outback, listed around $35k

Ooo nice I wrote Subie off after the Solerra flop but I’ll take a look

The Solterra got a lot better in the 2026 model year. It’s not a bottom-tier EV anymore. (Disclosure: I have one)

Lightnings are awesome trucks

Used Tesla Model Y with AT tires? You can get a small lift kit for them if you want more ground clearance.

Fuck Elon.

They don't specifically say you can install your own OS but they have similar wording on the Steam Frame page as their other systems:

> Just like any SteamOS device, install your own apps, open a browser, do what you want: It's your PC.


But as gp correctly points out, with ARM there tend to be additional technical challenges for that compared to your normal PC even if Valve does nothing to intentionally lock it down.

No Man's Sky is doing it now. It's also useful for foveated streaming which works for all games.

https://www.nomanssky.com/cosmos-update/#:~:text=Added%20sup...


Curious if there's any info about how much Foveated rendering helps performance roughly. Is it like 50% more frames?

It'll be highly dependent on the game. Unlike foveated streaming the implementation of foveated rendering could mean many things. Actually rendering at a lower resolution in areas you aren't actively looking is probably the most complicated approach because it seems like you'd have to render two times (low res + high res) to get the desired effect.

> Actually rendering at a lower resolution in areas you aren't actively looking is probably the most complicated approach because it seems like you'd have to render two times (low res + high res) to get the desired effect

In the newer implementations, you can split the display into tiles, and render each tile at a different resolution, so it's not that bad. A bunch of the intermediate rendering targets are already rendered at reduced resolution, so you won't see a full 4x speed up in the low-res regions, but you should be able to achieve a ~30% overall speed up with modern foveated rendering.


That's really cool to know, thanks for sharing!

The framebuffer that you are filling out is still at native resolution, but while computing that frame buffer you can do many of the steps at reduced resolution. For example, instead of evaluating the pixel shader code for every pixel, it could evaluate once for a block of four pixels in the reduced resolution region and use the result for all four.

No, memory is not limited to facts. There's just no visualization to go along with the recollection of faces etc. You can absolutely still recognize people correctly when looking at them.

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

Search: