Wow. This guy is just plain bad at this. Does heat mapping, puts the devices side by side.
The more I watch, the stupider he sounds when trying to get into technical information.
Good lord, “it beats the Mac!…now the 16GB version of the Mac did it 30 seconds faster than snapdragon but for some reason we chose to get the 8gb MacBook Air to run this entire video”.
People actually take this guy’s opinions as worth listening to?
It's worse than that ... he kept repeating that performance isn't affected by being on battery when it very clearly was. That might be true on ST test but while running CB R24 MT that's why he noticed that the chip was throttling while still cool. And his final on-battery MT CB R24 performance* was worse than scores from other outlets that ran their tests on-power.
Also, he went the other way too, claiming that Solar Bay didn't work because the Adreno GPU doesn't have hardware ray tracing. Adreno does have hardware accelerated ray tracing, Solar Bay runs just fine on Androids with Adreno GPUs, and Ryan Shout
reported scores for it. Not sure why it didn't work for Max Tech, that's odd. There have been reports of frequent crashes and bugs with some benchmarks, but my understanding was that was mostly for x86-Prism emulated programs which Solar Bay isn't (pretty sure).
So yeah lots of mistakes and errors.
*As an aside: it's possible that there is an issue with CB R24: note that in the far superior NotebookCheck article the Snapdragon's CB R24 single core score is much worse relative to the M2 Pro than it is under GB 6.2. I suspect the latter is closer to the truth given that it should be perform almost exactly like an upclocked M2. Still figuring out the ramifications of that if true.
First and foremost this is a Mac forum. Second, M3 got its ass whooped by Snapdragon X Extreme ultra whatever the name of that chip is in CPU Multi-core performance by almost 50%. For the price of M2.
That's the point though ... it did so in an actively cooled laptop using much more power. That's why the base M3 is a difficult thing to compare to and why NBC took issue with the comparison in their closing statements which
@Homy repeated. The base M3 can be fit, and perform well, in a passively cooled laptop/tablet and sip power even under load. The Snapdragon Elite isn't really in the same class of processor, even when binned. The Plus might be (more on that below). But bottom line if Qualcomm/Microsoft wants an Air competitor then they need a class of processor actually designed for those power envelopes, probably with a mix of performance and efficiency cores. They honestly need at least 2 SOCs to cover their use cases.
I think this gives ARM Windows fans a lot to look forward to. Not sure why would you want ARM PC but it's an option, I guess.
I want to be clear: Despite that their claims don't really hold up*, I don't think this is a bad first generation PC SOC at all. It's really quite good actually. I'm just worried for Qualcomm's sake that it's later than it ideally should've been. From Dell's leaked marketing slides, they're clearly 6-9 months behind at least. A year ago and they would've been in a much more comfortable position. Of course, we still have to see how AMD and Intel processors actually perform, no first party promises should be taken at face value, and Qualcomm's chips are reportedly much cheaper for OEMs. But, even so, Qualcomm's second generation SOCs will have to make substantial improvements (even more so to the GPU or get partners for dGPUs) and MS will have to grease the wheels to get more native apps and games.
*They were obviously playing word games, this is why I'm not disappointed or surprised. The marketing slides were pretty much accurate but presented in a way to give a misleading impression about what kind of device this would power. These are actively cooled devices and under load these Elite SOCs, especially the 80s and up, will use much more power than a base M3, especially in the passively cooled Air. Even the MaxTech video admitted that much and truthfully this was obvious even from their own presentations, they just hoped no one would notice. The Plus looks to be more similar to the base M3 under load and be a better comparison point. The few MT scores I've seen are slightly higher than the M3 as one would expect with probably more similar thermals and power envelope, obviously ST is a different story where the M3 is way ahead. And honestly for these classes of devices ST is probably the more important benchmark.