Not that crazy to beat a desktop CPU in specific category like this. This article went off the deep end when they made the comparison to the Huawei CPU. If you actually put one of these in a phone and spent any amount of time getting the kind of performance they’re claiming it would almost immediately thermal throttle or cook itself. These comparisons are pointless.
having a high peak performance means you can underclock it, giving it less energy to get the same performance as before with less heat, meaning at the thermal throttle point it is that much faster than before
Exactly what I wanted to say. On paper my pixel 1 was significantly faster then my pc at the time, an ancient core2quad q8400. In a lot of ways the phone did feel faster, then I tried to decompress a large 7z file on both of them and,despite using a mechanical hard drive, my pc took just over half as long.
Now overall, there were a lot of things that chip could manage better than my desktop but the point is that using a single benchmark or benchmark suite that leans into the strengths of the hardware is pointless.
Yep, synthetic benchmarks aren’t worth much outside of theoretical performance. Test a complete system as you can actually buy it and with an actual real-life workload - i’m sure the performance will still be good, but nowhere this outstanding.
Edit: that doesn’t even take into account the amount of times companies have been caught cheating on those tests, which is only possible because they are not a real life scenario
Eh, it’s 2 years newer than the AMD architecture, on a better node, and has better memory packaging. I mean it’s great they’re getting this performance, but wait until Zen 6 for a contemporary architecture and node comparison.
I feel like I’m taking crazy pills every time Apple releases a CPU. Their designs are good but most of it is the new TSMC node they use. When others get access to the node the difference is not that large but then the cycle starts again.
Geekbench is basically known for favoring Apple. It’s the consumer pc equivalent of “my dad can beat up your dad” because it is all made up and not real.
What does that stand for? I literally can’t look that up, all I get is t-shirts with phones on them, no matter how many apparel-related keywords I tell it to ignore
Actually insane they are beating desktop processors with a phone CPU.
Not that crazy to beat a desktop CPU in specific category like this. This article went off the deep end when they made the comparison to the Huawei CPU. If you actually put one of these in a phone and spent any amount of time getting the kind of performance they’re claiming it would almost immediately thermal throttle or cook itself. These comparisons are pointless.
having a high peak performance means you can underclock it, giving it less energy to get the same performance as before with less heat, meaning at the thermal throttle point it is that much faster than before
Exactly what I wanted to say. On paper my pixel 1 was significantly faster then my pc at the time, an ancient core2quad q8400. In a lot of ways the phone did feel faster, then I tried to decompress a large 7z file on both of them and,despite using a mechanical hard drive, my pc took just over half as long.
Now overall, there were a lot of things that chip could manage better than my desktop but the point is that using a single benchmark or benchmark suite that leans into the strengths of the hardware is pointless.
These “benchmarks” are great for marketing, so not pointless. Just pointless for users.
So it’s just pointless
Toms Hardware is a corporate propaganda op anyway.
Yep, synthetic benchmarks aren’t worth much outside of theoretical performance. Test a complete system as you can actually buy it and with an actual real-life workload - i’m sure the performance will still be good, but nowhere this outstanding.
Edit: that doesn’t even take into account the amount of times companies have been caught cheating on those tests, which is only possible because they are not a real life scenario
Core Ultra 7 355 is a mobile (laptop) CPU, with a nominal power draw of 25W and a base clock speed of 2.3GHz.
https://www.intel.com/content/www/us/en/products/sku/245722/intel-core-ultra-7-processor-355-12m-cache-up-to-4-70-ghz/specifications.html
Yeah that’s a different goalpost, but it’s still nuts.
Eh, it’s 2 years newer than the AMD architecture, on a better node, and has better memory packaging. I mean it’s great they’re getting this performance, but wait until Zen 6 for a contemporary architecture and node comparison.
I feel like I’m taking crazy pills every time Apple releases a CPU. Their designs are good but most of it is the new TSMC node they use. When others get access to the node the difference is not that large but then the cycle starts again.
Except they aren’t, outside of synthetic geekbench benchmark that is designed to produce this exact result…
Geekbench is basically known for favoring Apple. It’s the consumer pc equivalent of “my dad can beat up your dad” because it is all made up and not real.
Yes, geekbench numbers are completely meaningless and do not reflect reality
Unironically, yes. Especially when one is ARM and one is X86.
That would be crazy indeed and might be possible, but the CPU they compared against is a mobile CPU (Intel Core Ultra 355 that is).
Yeah they have to distribute all that compute for a reason.
What do you mean?
You don’t own your iPhone, and you def ass don’t own the baseband chipset.
Are you saying that there’s a conspiracy to build a giant distributed compute engine using our phones?
What are they using it for?
Middle out compression.
…and I can prove it!
Fam lookup TEE designs and all of the camera based processing that everyone is doing inherently. Then remove the TEE part.
Edit: https://en.wikipedia.org/wiki/Trusted_execution_environment
What does that stand for? I literally can’t look that up, all I get is t-shirts with phones on them, no matter how many apparel-related keywords I tell it to ignore
Edit: I see your reply to that other comment,
https://en.wikipedia.org/wiki/Trusted_execution_environment
I dont get your original comment though:
https://en.wikipedia.org/wiki/Trusted_execution_environment
I know what a TEE is, that wasn’t my question.