The A18 Pro chip from the iPhone 16 Pro when compared to the M1 is:> ~40% faster in single-core performance> ~3% slower in multi-core performance> ~10% faster in GPU performancehow long until Apple brings back the cheap base "MacBook" for $699 with an A-series chip?
>>102453358Based
>>102453358>btfos itoddlers in your path
>>102453371finally, 1 extra hour in my daily netflix loop
>>102453358Who cares its a phone
Aren't the M series chips based off the A series ones? Guess this gives us a hint of what M5 will be like. Looks like it'll have great performance.
>>102453464Many people
>>102453898Name 8
>>102454060NIGGER
>>102455113does this count as six people or one
>>1024558078
>>102453886A18 uses the same internal architecture and node as the M4, so we already know what it looks like
>>102453464a phone faster than your laptop
>>102456741Didnt answer the question
>>102453358Why does apple refuse to be competitive in the gpu market? A 6900xt from FOUR years ago that's basically considered a mid range gpu is still faster than apple's flagship $4000 chip. And if there's a technical reason why they can't do it, why not allow external gpus again?
who the fuck are the people that obsess over phone benchmarks? like i seriously could not ever give a fucking shit about this stuff. what practical benefit am i going to get from knowing a phone is "beating" a 3 year old laptop?
>>102456941The 6900XT by itself uses like 350W, the m2 ultra uses 300W for the entire mac studio system.>muh computeAMD's openCL support is STILL a joke, meanwhile Metal Compute actually just works fine. Apple is unironically the only alternative to nvidia for amateur GPGPU and AI.
>>102457020Those are metal benchmark scores.
>>102453371>comparing against last year's competitor>with a bigger battery>barely winslmao, intel and qualcomm are so funny
>>102456741impossible. a laptop has an order of magnitude higher power and head dissipation budget. they have not come that far in a few years.
>102453358>102453359>102453371>102453406>102453464>102453886>102453898>102454060>102455113>102455807>102455975>102455992>102456741>102456779>102456941>102456964>102457020>102457030>102457035The Intel i3 8100 from nearly a decade ago and selling for $10 on eBay destroys the latest crapple currynigger CPooU in real performance.
>>102456941Apple is by far the leader in GPU performance per watt, the M3 Max is very close to the mobile 4080 while using 1/4 of the power, so yeah, they're competing in the high end tier with Nvidia already>A 6900xt from FOUR years ago that's basically considered a mid range gpubullshit, the 6900XT is on the same performance category as the 4070 ti, which is a high end GPU (it was even meant to be called the "12GB 4080")
>>102457130get a toilet currynigger
>>102457130Who cares about performance per watt when your flagship "pro" computer can't even match a mid range $1200 gaming pc and you don't even have the option to use an external gpu? They might as well just admit that they've abandoned the high end computer market and only make low and mid range computers now.
And it's completely useless unless apple implements a feature like Samsung's DeX, which guess what, they never will because they want to sell MacBooks.
>>102457174> 4090> 24GB of VRAMvs> M3 Max MacBook> up to 128GB of VRAMthe best way to locally run LLMs with lots of parameters are MacBooks
>>102453371>bigger battery>worse screen>newer product>barely "wins"
>>102457174>who carespeople with brains
>>102457505>my computer takes twice as long to finish the same task but at least is uses less energy!It's fine if you want to admit that macs are the computer equivalent of an eco friendly compact car, but then price them like one and stop trying to market them like a luxury supercar.
>>102457030Support isn't about scores in a microbench, it's about how reliable the platform is. AMD opencl drivers still crash in monolithic kernels. Even some benchmarks crash on AMD GPUs
>tfw too poor to buy the best M4 macbook pro
>>102453358how long until you get a toilet currynigger?
>>102458285> M4 macbook prono such thing, only iPad slop has the best chip.
>>102457165> out of curiosity check your screenshot> M3 Max faster than 7900 XTX in Blenderthanks jeet, you actually made me realize Apple Silicon is even betterhttps://opendata.blender.org/benchmarks/query/?compute_type=OPTIX&compute_type=CUDA&compute_type=HIP&compute_type=METAL&compute_type=ONEAPI&group_by=device_name&blender_version=4.2.0
>>102453358according to this meaningless benchmark.how long does it take an iphone to compile chromium? oh wait, we'll never know because apple keeps their ecosystem locked tight.
>>102453371>>102453406finally, 1 extra hour in my daily gooning routine
>>102458908no one compiles Chrome on their phones though. That performance is great for gaming though, you can't play RE 4 remake and other AAA titles on Android phones because they're just too slow for that kind of workload
>>102459001>no one shits in a toilet though though. That performance is great for street shitting though, you toilet shit on Android phones because they're just too clean for that kind of workloaddo iJeets really?
>>102459068but anon... street shitting countries use Androidin fact India is known to LOVE Java and Android, if you read the chat in Google events it's full of filthy jeets
>>102459116> Nigeria> ~10% is "Unknown"nani??? what are they using, Ubuntu mobile?
>>102453371funny because Netflix could just optimise their shit software and probably triple the battery life of all those devices
>>102459411how? By creating a new magic video decoder that runs efficiently on 100 different chip architectures?
>>102457176following your logic, Samsung doesn't want to sell their notebooks?
>>102459155Likely some old phones no major OS from Google, Apple or Microsoft. Simple dumb phones or maybe feature phones using proprietary OS.
>>102457102>impossible>a laptop has an order of magnitude higher power and head dissipation budgetyou're retarded. The M1 uses more power than any A-series chip and runs fine on macbooks and ipads.The A18 consumes way less power while being considerably faster, so yeah, it's very possible.
>>102457104> M1> 35W> i7-8565U> 15Wthese poojeet screenshots are hilarious, Apple really did something rightalso> linux kernel compilation> real performance
>>102457020Apple is no alternative to NVIDIA at all, even a top of the line M2 Ultra is 10 times slower at matrix multiplication than a 4 year old 3090.
>>102461408>top of the line M2 Ultrathat's not top of the line, the M3 Max outperforms the M2 Ultra on many workloads, see: >>102458828
>>102461486It doesn't at matrix multiplication, it has almost twice the GPU cores and memory bandwidth. I choose the M2 Ultra because it is the one that scores the best, an M3 Max is much, much worse.
>>102458465>gameshow about some shit I care about in a laptop?
>>102461558>because it is the one that scores the bestin what benchmarks?
>>102461637llama.cpp
>>102461657looks like this workload only cares about MUH CORES. Even the M1 series is scaling linearly with the amount of cores.
>>102461657>ai shit who cares
>>102461725>>102461743Here comes the cope. It's the same workload as for any other AI app, and I was responding to someone who said that "apple is the only alternative to nvidia for ai".
>>102457165good morning sir!
>>102461783huh.... I wonder why this relation exists. WEIRD!
>>102460519Comedy option: Windows phone.
>>102453358what is geekbench
>>102453358The Quality Communication Antirrhinum Eight Generation Four is also faster than the Year Two Thousand Twenty M1. What a coincidence.
