It doesn’t even “seem” white spread. Literally a single guy making unsubstantiated claims.
Got a tip for us?
Let us know
Become a MacRumors Supporter for $50/year with no ads, ability to filter front page stories, and private forums.
Qualcomm Says New Snapdragon 8 Elite SoC Has 'World's Fastest Mobile CPU'
- Thread starter MacRumors
- Start date
- Sort by reaction score
It doesn’t even “seem” white spread. Literally a single guy making unsubstantiated claims.
It has the fastest single-core performance and runs at a higher frequency, with it being eight cores vs. Apple 6 cores. But single-core Apple is still faster.
The Media Tek chip is the one Apple has to worry about runs almost at fast as Apple's in single core and will be releasing a desktop chip next year with the help pf Nvidia.
The Media Tek chip is the one Apple has to worry about runs almost at fast as Apple's in single core and will be releasing a desktop chip next year with the help pf Nvidia.
Most people buy the Pro phones as fashion accessories, and the silicon in their phone sits around doing nothing until they upgrade to the latest fashion accessory. People who actually use their Pro phones for "pro" features (at least the ones that Apple allows) understand how inadequate the thermal solution for the iPhone 15 Pro is.
Well, they engineered themselves out of a plastic bag to become the most valued and rich company in the world, so I guess your point is moot.
Thanks for your contribution, whatever point you might have had you lost it when you said such bs, including the comparison of chip design to ordering pizza from dominos, lol
I suppose we'll have to see how Zen 5, the architecture that had minimal performance gains and focused on efficiency, ages.
That being said, most software hasn't outrun the decade-old i5 in my desktop at work.
I had severe overheating on my 15 Pro Max and had to replace the battery in a year. I purchased a 16 Pro and it runs warm at times but it much better.
Have a look at Samsung Dex. It’s something in that direction for Android.
About the new Snapdragon chip, they need 6 P-Cores to call it ‘the fastest’…
The Samsung S24 Ultra comes in with Geekbench scores of 2139 (single core) and 6684 (multi core).
The 12 Pro Max with the A14 Bionic comes in at 2119 (single core) and 14 Pro with the A16 Bionic hits 6676 (multi core).
For reference, the 16 Pro/Pro Max A18 Pro is getting 3409 (single core) and 8492 (multi core).
Whatever “world‘s fastest chip” is being touted has zero reflection of the capabilities of the chips in latest and greatest Android phones. This happens every year and this 2-4 year benchmark lag for Snapdragon chips is not a new phenomenon either.
Firefox literally sucks on Android. I don’t even know or understand why it exists. It still has no support for add-ons (i.e. adblockers), very slow browsing speed and no theming capabilities (the main reason I use it as my main browser on Mac). Chrome on the other hand eats battery in minutes. I absolutely love Samsung browser on new Galaxy devices, I’ve heard it uses webkit (correct me if I am wrong) thus it offers almost Safari-like speed, no adblock support either unfortunately (at least those adblocks that they offer do not work properly).
Well, as a long time Apple user I gotta unveil the “myth” of 7 years+ iOS update support: there is no other way.
Android has one major advantage over iOS – it doesn’t require OS updates for the apps to function as expected, i.e. you can pretty much install most apps from Google Play onto devices running as old as version 4.4 Kitkat. Cannot say same about iOS 6, many apps require version 12 or 15 nowadays.
On iOS it is not possible to update Safari without updating OS itself. If phone is not updated for 3 years or more, sites become unresponsive or often fail to load. Rarely a problem even on old Android phones.
Exynos proved its uselessness, Chinese mobile chipsets have rather poor app support and sometimes applications fail to load. Qualcomm is just sort of universal choice for now
What's really sad about this arrant nonsense is that some gullible readers may actually be taken in by it. The poster is a master of irony, but not engineering.
He sure has stated it repeatedly. First time I remember seeing this insanity was in April. Back then he was spanked hard and didn't show his face for a while, probably in sheer embarrassment. Since it's almost all relevant to his current postings, I'll repeat it here verbatim. Alternatively, click the link in the quote attribution to see the original thread. There's a bunch of good info in the last few pages.
He sure has stated it repeatedly. First time I remember seeing this insanity was in April. Back then he was spanked hard and didn't show his face for a while, probably in sheer embarrassment. Since it's almost all relevant to his current postings, I'll repeat it here verbatim. Alternatively, click the link in the quote attribution to see the original thread. There's a bunch of good info in the last few pages.
For the TL;DR crowd, here's the most important part of the response to the high (hahahaha) IQ person:
[...] The notion that the chip design is meaningless is ... more wrong than mustard on ice cream. It's laughable. For a simple proof of the sheer stupidity of it, consider two different core designs on the SAME process: Apple's P and E cores. There's roughly a factor of 3 difference in performance. Or look at Intel's P & E cores - the difference is even larger. [Note that this was written before Lunar Lake, and it refers to Intel's non-chiplet processors] Naturally, in both cases, the P cores are a LOT larger. Design with more transistors, you can get a faster core. Pretty basic.
You could also compare Apple's older N7 cores (A12 or A13) with another vendor's N7 core. The differences are stark.
Lastly, as I mentioned in a previous post, design will determine the highest clock you can run a chip at. In the language of the P-E curve, the curve doesn't extend forever. It cuts off at a certain point, beyond which more power won't get you any more performance, because the design is literally not capable of it.
[...] The notion that the chip design is meaningless is ... more wrong than mustard on ice cream. It's laughable. For a simple proof of the sheer stupidity of it, consider two different core designs on the SAME process: Apple's P and E cores. There's roughly a factor of 3 difference in performance. Or look at Intel's P & E cores - the difference is even larger. [Note that this was written before Lunar Lake, and it refers to Intel's non-chiplet processors] Naturally, in both cases, the P cores are a LOT larger. Design with more transistors, you can get a faster core. Pretty basic.
You could also compare Apple's older N7 cores (A12 or A13) with another vendor's N7 core. The differences are stark.
Lastly, as I mentioned in a previous post, design will determine the highest clock you can run a chip at. In the language of the P-E curve, the curve doesn't extend forever. It cuts off at a certain point, beyond which more power won't get you any more performance, because the design is literally not capable of it.
Hence my point about CL being a crap compiler.
There are also some problems with NT's architecture on system call latency as well which is a damage multiplier on windows x86 as it is already (look at how slow small atomic things against the filesystem and fudging kernel objects is). I think if I remember, reading the call conventions years ago, the SWI semantics are terrible as well on ARM. Not sure if the latter applies now but on 32-bit ARM years ago (windows phone) it was awful.
Android probably uses (I don't know) LLVC/clang or GCC or something far superior.
Actually Firefox does support add blockers in Android. It's the only reason I installed it. I use it with uBlock Origin.
And lastly, about the chip itself...
Do we know if this is the same core as in the recent SXE? (I'm pretty sure it is - they're using the exact same name, Oryon, rather than "Oryon 2" or something else.) Because if it is, any notion that it's as fast as the A18 in single-core (which is by far what matters most in a phone) is hilariously wrong. The A18 crushes the SXE. A18 raw GB6 score is superior, but more importantly, it uses less power. That in turn means that you can actually run your core at top speed for a meaningful amount of time without throttling or draining your battery. EDIT: Yes, It's actually two new cores, about which I and many others posted a lot. It's a huge improvement but it's still a generation or more back from the A18.
For multicore, the situation is a little different. I'd expect an 8-P-core SD8G4 to be faster than the A18P in GB6 multicore. BUT... how often does this matter in a phone? I think the only significant place you're going to see a phone max out multicore for any length of time will be in some games. (There are doubtless less common other cases - but the situation should be the same for them.) And in that case... the phone with *actually* better performance will be the one that can *sustain* peak MC performance. That will again be the A18, because it's way more efficient.
The NPU stuff, AFAIK, is an open question not likely to be resolved soon as I have yet to see any broad cross-platform measurements that are meaningful. There are things that try, but there's no way to know if they're even remotely optimized properly for any particular platform.
GPU is interesting. QC definitely beats Apple by significant margins in some benchmarks. However the Adreno in the SXE (and presumably in the SD4g4) is much less capable in some ways than the Apple GPU. How this will work out for gamers, I don't know - this is not something I know all that much about. @leman can comment more on this. I do think that over longer timeframes than benchmarks take, heat and power draw will dominate the discussion, and there again Apple will win even for most games - but I don't know that.
This has every sign of being a replay of the SXE rollout. QC talks big, but in the end they're not even close to Apple. However, that doesn't matter so much because they're not really competing with Apple, whatever they're fantasizing about. They're competing with other vendors, and in the market they're actually in, they likely have a quite competitive product.
For multicore, the situation is a little different. I'd expect an 8-P-core SD8G4 to be faster than the A18P in GB6 multicore. BUT... how often does this matter in a phone? I think the only significant place you're going to see a phone max out multicore for any length of time will be in some games. (There are doubtless less common other cases - but the situation should be the same for them.) And in that case... the phone with *actually* better performance will be the one that can *sustain* peak MC performance. That will again be the A18, because it's way more efficient.
The NPU stuff, AFAIK, is an open question not likely to be resolved soon as I have yet to see any broad cross-platform measurements that are meaningful. There are things that try, but there's no way to know if they're even remotely optimized properly for any particular platform.
GPU is interesting. QC definitely beats Apple by significant margins in some benchmarks. However the Adreno in the SXE (and presumably in the SD4g4) is much less capable in some ways than the Apple GPU. How this will work out for gamers, I don't know - this is not something I know all that much about. @leman can comment more on this. I do think that over longer timeframes than benchmarks take, heat and power draw will dominate the discussion, and there again Apple will win even for most games - but I don't know that.
This has every sign of being a replay of the SXE rollout. QC talks big, but in the end they're not even close to Apple. However, that doesn't matter so much because they're not really competing with Apple, whatever they're fantasizing about. They're competing with other vendors, and in the market they're actually in, they likely have a quite competitive product.
Last edited:
They sucked. They made an inferior product compared to Apple's M chips. And they have more or less figured it out, and are catching up. intel and AMD will figure it out as well. And make inroads for sure. But, in the end. Microsoft will have to make a choice. Stick with 2 different architectures (ARM and x86-64). Which I can't see happening forever. Or one of these chips will win out. But how long it takes for that to play out is anyones guess.
When will we see Nvidia or AMD GPU support for ARM? When can you build a tower (or any kind of desktop) with Qualcomm chips and any 3rd-Party add-on? It's nice in a thin and light laptop, but will there be a desktop variant that can push the limits even further? Then of course the server end. When is that coming? As a consumer, what should you buy? As a gamer, there is only one choice and that is X86 for now.
Not really... they are catching up, much like Zeno's Achilles is "catching up" to the tortoise.
As I wrote a while back, it's not so hard to get to 80% of Apple's performance (and, more importantly, performance/power). Getting to 90% is really difficult. Getting to 100% is, so far, completely impossible for anyone except Apple.
Things haven't changed much since I wrote that. The M4 shipped (before the SXE, to everyone's surprise!), as did the SXE, Lunar Lake, and Zen 5. Intel especially has improved their position, but not enough. In the larger picture, Apple is still the performance leader, and especially crushing it on perf/power. It doesn't seem likely that anyone's going to catch up to them any time soon, on a tech level. And my 80% (above) turns out to be pretty optimistic for non-Apple players, in fact, if you're looking at perf/power.
On the product marketing level it's a different story. Apple won't keep up with raw MC performance at certain price points, because they choose not to. Obviously, given the size of their E cores and their chips, they could put 12-18 E cores in the M5 if they felt like it. Even 8-12 E cores in the A19. That seems very unlikely though. They obviously believe that certain levels of MC performance are all the market calls for at certain price points. If you're outside that envelope, then Apple chips won't be your best option. But I think that they're generally right, and that most people won't need/want that.
Last edited:
Resolved via software, so not likely the chips.
This is interesting. In the recent past, Samsung processors, even after the new processor would be announced, were comparable to iPhones which were 2 Generations behind. So when the iPhone 14 Pro came out, the Samsung which came out after the iPhone has comparable specs to the iPhone 12 Pro Max. Or I guess you can compare it to the iPhone 13 non Pro (as that was basically an iPhone 12 Pro), but Samsung processors were not comparable to the current generation top end iPhone.
Exynos (Samsung) tends to be further behind that Snapdragon (Qualcomm). Especially on process node.
Samsung only uses Exynos in some markets; they use Snapdragon elsewhere.
I think in the end, while the Snapdragon 8 Elite is definitely fast, I do worry about how much battery is needed for decent life per battery charge and also the amount of heat this SoC could generate. Let's hope that Samsung has upgraded their vapor chamber cooling system so the SoC does not run overly hot in high CPU demand situations on the Galaxy S25 models.
