![[H]ard|Forum](/styles/hardforum/xenforo/logo_dark.png){kind=logo}
There's a story today on WCCFTech--so, you know, the usual caveats--that the USB-C ports have moisture sensors in them, with a reporting daemon, and, it being that site, a suggestion that tripping the sensor would cause Apple to deny warranty claims.
If you have the type of workload that could actually use the insane amount of GPU memory that these things offer - They pretty much stand on their own, and it's impressive given the laptop form factor.
No Apple has accepted warranty claims that had Apple Care attached where I have straight told them that the tech agent that the operator spilled a cup of coffee on it.
Geekbench isn't good at comparing Apple products to other products. It's fine for comparing M3's to M2's and M1's.
So you're saying that Apple is immune to benchmarks?
No yea, you think benchmarks don't matter. Not even sure if you're serious or joking now.
You can always compare Apples to Apples. You're excusing Apple for their performance.
I wouldn't call Final Cut Pro an industry standard. DaVinci Resolve is more industry standard.New apple laptops can perform better in certain use cases, but can it run crysis?. At the same time you aren't going to run some industry standard software like Final Cut Pro on windows...because you can't.
Everyone is definitely jumping to Resolve these days.
Apple has always bent people over backwards for RAM upgrades...heck, Dell does it in their server line charging $2k + for a 32GB stick you can get from anywhere else for a couple hundred...same model.
Yeah, when we rebuilt the cluster at my current job, we ordered the servers with the bare minimum RAM Dell would sell and then bought it from Crucial or Amazon, can't remember. Saved several thousand dollars on the final project cost. And that was buying a spare kit as well.
I feel like we're saying the same stuff but debating over semantics.
To be fair, AMD is using 4nm for their mobile 7000 chips.
The cost of going 3nm is probably why Apple is cutting cores and bandwidth in their M3 chips. I also wouldn't be shocked if the 3nm TSMC process is full of defects and the M3 Pro chips are probably cut down due to this. Apple seems to have made real performance improvements, but this allows them to cut down cores and bandwidth while also still being technically faster. The thing that's really telling is the M3 Max, where it seems Apple put a lot more effort into as they actually added more cores instead of taking them away. Another situation where Apple is telling it's consumers to buy the more expensive part or else. The M3 Max is a monolithic die, meaning that Apple is really up the nose in cost for this chip to be made. Unlike AMD who is saving money in manufacturing by using their chiplet design, as well as Intel who's Meteor Lake will have the CPU and GPU manufactured separately in what they call tiling. With AMD's chiplet and Intel's tiling, they are also increasing performance, because they can bin better. When Apple makes their M3 Max chips, they will have some defects on them and that will require them to either reduce clocks or cut off parts that maybe defective like cores or pipelines for bandwidth. What Apple supporters don't understand is that the reason Apple makes monolithic chips is not because it's better, but because they don't have the engineering capability. Just like how unified memory is not better but worse, and you can't upgrade the ram. Unified memory is always done to cut costs, and not increase performance. This is why the Xbox's and Playstation's have been doing it for over a decade. Also, the Apple fans need to put down the Geekbench crack pipe. They should stop with all the synthetic benchmarks while they're at it. We've seen comparisons with Geekbench that make it seem like the Apple Silicon is so much faster, but when tested on an actual application, it's so much slower. Good chance when reviewers actually run real world applications against the M2 and M1, they will find that the performance Apple was promising with the M3 will be very disappointing.
Apple paid just north of $1B USD for the TSMC 3nm tapeout.
Apparently for enterprise they only sell Apple Care Plus which covers accidental damage from spills and drops. Apple Care itself doesn’t cover that, so now I know.
So far, most of the real-world tests I've seen (Blender, video editing, games, a few others) show significant speed improvements, particularly for anything GPU- or video-heavy. Not enough to make someone ditch their M2 Pro/Max system, and it's not always an easy call if you have an M1 Pro/Max, either. But few people replace computers every year, and there are pros for whom Apple's performance advantages are particularly appealing.
I would like to see these reviews. Sadly, I'm waiting for Linus Tech Tips to do the M3, even though their reviews aren't good, but still better than most Apple only reviewers.
Doesn't sound very significant then. Apple's problem is their products aren't going beyond existing customers. They know an M2 owner won't upgrade, but they want M1 owners to upgrade and they want Windows laptop users to switch.
It's hard to find tests done on AMD's mobile 7000 series unplugged but from what I've seen it lasts just as long as Apples. Gotta remember the Steam Deck is x86 and lasts up to 2 hours on battery while playing games and won't slow down when unplugged, unlike the Switch.
There were prototypes of the pre-unibody Macbook Pro's that had cellular. But it seems to me the issue is they don't think that there is enough of a market segment to warrant the R&D as well as increased manufacturing costs to go after such a tiny portion of the market that can simply tether. It basically would require an entirely new production line just for that significantly more complex SKU, and I doubt more than 5% of MBP buyers would select it as an option. So unless it's going to be a $1000 option (or some other ridiculous number that most people wouldn't want to fork out the money for), the ROI just isn't there.
Yeah OEMs require this to be selected when using "build your own". Even though the motherboard often can accept the adapter, the antenna hardware isn't there.
From what I can tell the 5700X just edges out the M2 Max 12-core CPU in SMP in general, so the M3 would most likely be faster overall.
From what I can tell the 5700X just edges out the M2 Max 12-core CPU in SMP in general, so the M3 would most likely be faster overall.
From a cost perspective, the 5700X platform with similar specs would most likely be far less than that loaded M3 Max you referenced.
Upping an M2 Max Mac Studio with 128GB unified RAM increases the cost by $1200 USD alone, and the 8TB SSD increases the cost by $2400 USD by itself, so yeah, you are definitely going to pay the price for that "low power".
I think a better analysis would be a cost-for-cost comparison, assuming it can be done since Apple does make one pay for those specs.
It all depends on the user's requirements and tasks that the system will be used for.
My personal laptop is a Ryzen 7 7735HS with a Radeon 7600S GPU. What do you want me to test?
Keep in mind that my 5700X was bought used for $170 off Ebay, so it wasn't bought for performance but because there was a fire sale of people who over paid for computer parts in the past 3 years. If I wanted to build a PC that was fast, I would buy a Ryzen 7 7950X3D, along with a RX 7800 XTX because I use Linux and AMD is best on Linux. At best it would cost $3,700, and if it happens to be faster than a M3 Max then so much the better. While your tricked out M3 Max would cost over $7k.
Any laptop with a AMD Ryzen 9 7940HS can do what Apple can do, even unplugged it keeps it's performance. Something like the Asus Zephyrus G14 with the AMD Ryzen 9 7940HS. Not only that but most of them are equipped with a RTX GPU, so either way it beats Apple.
Can do assuming Microsoft has their shit together which is a bit of a stretch.
Or using it on an airplane! ;-)
His laptop will die after an hour of being unplugged.
You mean test my existing years old PC against a new M3 Max? Or a newer laptop based on the AMD Ryzen 9 7940HS? With what tests? Lets install our favorite Linux distro on these machines and find out. Oh sorry, Asahi Linux is still far from being a daily driver on Apple hardware. Lets run benchmarks on games like Starfield and Jedi Survivor. That's right, Apple is for productivity so lets fire up 3ds Max. You know what, lets run a VM instead. I'm sure that 8GB of ram that you're stuck with will magically turn into 16GB for a VM. You see the problem here?
There's a lot of what ifs going on there. What GPU ram, because there's the 780m and then there's the plethora of RTX GPU's that tend to be included? You don't have a dataset, you don't have anything. In GPU tests that aren't Geekbench, an RTX 3060 destroys the M2 Max. So I assume that any RTX-40 series GPU on a Ryzen 9 7940HS will continue to destroy it in performance. You don't even own an M3 Max. Not even sure what you're using now.
That and the M3 Pro is not really any better (and sometimes worse) than the M2 Pro due to the switch from 8P+4E to 6P+6E core configs. I can only speculate that was for battery life, but still kinda lame.
I do wish the M3 Pro was better... with that said, battery life is the name of the game for me, so I won't mind so much if I get one and it lasts an hour or two longer.
That’s half of it, but I bet Apple looked at their collected logs and metrics and saw the average user of the M1 platform wasn’t running it hard enough where they would notice the difference and those few who would could be upsold to an M2 Max or we’re going to be holding out for the M3 anyways.
I bet it's more for power efficiency. Either that or this is the result of 3nm manufacturing defects and this is the compromise Apple went with.
I think you missed the part where AMD's Dragon Range caught up. The Asus Zephyrus G14 laptop that I mentioned will performance nearly as well as Apple's when unplugged. According to Jarrod'sTech video, the Asus lasted for 560 minutes when playing a YouTube video while the Apple M2 Max lasted for 748 minutes. That's pretty good considering that AMD's idle power draw is higher. Also, these are gaming laptops with 165Hz and 240Hz displays and RGB lightening, so of course they eat more power.
This malaka here.
Maybe you missed the reviews the first few times I've linked it, so I'll link it again.
AMD is not the one Apple is worried about. Why you think Apple was quick to release their M3 news? I still can't even find a review of the new M3 chips. Intel is suppose to release their Meteor Lake chips soon.
That's really good. What do you think those M1 laptops get when playing games like World of Warcraft? It ain't 2-3 hours without turning down brightness and lowering graphic settings.
The problem that you and many Apple faithful seem to gloss over is that the magical efficiency only happens under very specific conditions. Fire up a multi-threaded application and the power consumption is no different between AMD and Apple. Fire up a game and suddenly 2 hours of battery would be an achievement. You keep comparing Apple GPU's to that of Nvidia, as if they're comparable. Everytime a GPU application is tested, even the M2 Max can only wish to perform as well as the RTX 3060. Apple is now getting features like Mesh Shaders and Ray-Tracing, 4 years after Nvidia, 3 years after AMD, and one year after Intel's A770.