Considering we see no tangible improvements in GPU graphics or performance as well as CPU performance, this isn't so interesting in the smartphone space. Pretty much any high end Android phone can run any app or game and run it very well.
In this case I don't mind, the current x gen y naming nonsense didn't make it much sense when you start looking at the individual products. And the elite lineup is at least in line with their notebook series (thoughts those chips have terrible naming too).
Looks like a power hog space heater... The 9400 is way superior... I am saying this as a snapdragon fan who was planning to buy the one plus 13... The fact they didn't even mention CPU power consumption itself is enough proof it's really bad except on power unlimited benchmark scores.. ( I have the same one)
@@soulsbournewe will have to wait geekerwan’s video, because I saw a post with a caption saying that the 8 elite is far more efficient than both the 9400 and a18 pro and there is a screenshot as well showing the power consumption from the 8 elite; it consumed only 11.9w when it achieved a multi core score of 9600, whereas the a18 pro only achieved 9200 when it also consumed the same wattage as the 8 elite.
How long have all the chip manufacturers been telling us that multi-cpu\ multi-thread is the way forward, yet so many things seem to be single core execution.
Because of battery life. It would probably half the hours of battery and degrade the batteries faster. With how fast CPU's are now multi threaded operations are not as necessary
We should do a comparison of chips by not just normalizing frequency but also power draw, fabrication node etc so we can see whether the uplift really coming from architecture change.
Perf/MHz alone doesn't tell you much, perf/MHz/Watt is a better metric. And according to Qualcomm's power efficiency claims, it looks like they have brought a working Netburst/Bulldozer-esque high clock design to mobile with more pipeline stages while minimizing static power (lower fan-out depth combinatorial logic) and leveraging savings in dynamic power (sequential switching logic). Well done.
@@LogioTek That early benchmark is not far off from the actual chipset. Apple do this too with their bionic chipset before their iphone was officially released at the market. Scores will be slightly different depending on the hardware optimization. A common issues in every smartphones especially android.
@@byssmal Do you not understand that I'm talking about power consumption during benchmarks? Until it's measured and replicated by 3rd parties we simply can't say for sure.
@@LogioTek Qualcomm reducing the cores to 6 of their 8 elite proved that 8 cores will consume more power and overheating. A bold move to mitigate some issues.
Hats off to meditek and arm. They seemed so low key but managed to complete well against so hyped up oryon cores and apple, all this with being significantly efficient as well
Competition is great for us consumers… Qualcomm and mediatek multicore domination will force Apple to innovate and offer better performance. And also force x86 vendors to align and offer better efficiency. Or x86 will die.
As long as Apple has full control of the hardware and software it will still outperform the rest, what you see there are benchmarks where companies typically increase the wattage for these tests; actual performance is very different.
@@jorvinlaguna1932 Apple is the worst phone device you can go with, not even worth a penny. They overheat like a toaster, poor performance and battery drains very fast during gaming. Also the camera performance isn't as good as android is
i think you shouldn't jump ship so fast brother let's wait and see how the heat up 😂 anyways 870 is still in my opinion the best processor by far without any issues
thx for the video, the single core performance of 8 elite looks great , but I'm a little worry about not having custom design efficiency cores and how this will affect battery life in idle and at lite use.
The clock frequencies listed are the maximum achievable frequencies of those cores. Clock frequencies vary and hit peak values very infrequently during any benchmark tests. So, your Single core score per GHz segment 6:16 doesn’t convey anything meaningful as you are simply dividing max frequencies.
@@GaryExplains first off I love your videos! And thanks for your response. I see your point. But, the benchmark scores were got from dynamically changing clock levels (All 3 CPUs rarely hit Peak frequencies). So, I don’t think considering only max frequency gives us an accurate picture of IPC here, since we’re punishing the ones that offer higher operating frequencies but rarely operate at those levels. The correct way to do this would be to ensure all CPUs operate at their max frequency for the entirety of the benchmark and then calculate max IPC. These are just my thoughts. Edit (cos, I can’t seem to reply to your chats): Geekbench 6 uses real-world tests so it’s not gonna load up the CPU to be at Peak frequencies for most tests in their benchmark. “Geekbench 6 measures your processor's single-core and multi-core power, for everything from checking your email to taking a picture to playing music, or all of it at once. Geekbench 6's CPU benchmark measures performance in new application areas including Augmented Reality and Machine Learning, so you'll know how close your system is to the cutting-edge.” “Geekbench uses practical, everyday scenarios and datasets to measure performance. Each test is based on tasks found in popular real-world apps and uses realistic data sets, ensuring that your results are relevant and applicable.” Source: www.geekbench.com
I think for single threaded it is safe to assume that a) the CPU is at its max freq as much as is humanly possible to ensure as they are CPU intensive. b) Any non peak usage (like when one test ends and another starts) will be replicated across all systems as they all run the same types of tests.
I would albo think that single core should hit max frequency, but lower than mediatek result is strange - so reference design might not hit it for entire test duration
Hey, Mr. Gary! Qualcomm is clocking this even higher than their laptop version X Elite. Do you think it will be efficient enough or can sustain performance in phones like S25 as Samsung like to keep the thermal limit lower?
Leaks have stated that the thermal condition on these chips are really bad, even reaching around 100 C if the leaks were correct. There is no way they can sustain this. Though they have increased the clock speeds significantly, their IPC is not that great compared to Apple and even somehow even to MediaTek as well. MTK can be a serious competitor this time.
40%-50% improvement is a bit underwhelming comparatively to hype I had but still great. Should be really great in emulation (can't wait to emulate Bluestacks on winlator lol). Also really impressed by dimensity 9400
As usual, the primary determinant of the _snappiness_ of a device is the single core performance - multicore is good for slogfests like transcoding or rendering, but not everything can be multithreaded. Good single core performance nowadays requires multiple decoders, a deep pipeline, good branch prediction, a big reorder buffer, redundant arithmetic engines, and a whole lotta engineering.
at 7:28 processors don't work linearly it may perform higher at a certain optimum frequency. ex if you operate all these cpus at 3ghz all of them will score differently unlike just dividing it by clock speed.
But you can calculate what that particular core does per ghz. So, it tells exactly how well it performance running on 1 ghz. To put it differently: how efficient the core is.
I say go hard and make the best snapdragon you can. I own a tab s9+ and use it for vacation and emulation. I also own a iPhone 16 pro max. Competition is important and keeps these companies innovating. I have always been an android and iOS user. I also own a Odin 2 max for my primary portable game system.
Is there a way to compare performance per watt. Cuz performance per clock speed is a meaningless chart in a mobile, as the wattage is the limiting factor.
I think the next generation of smartphones is the first that will need to be way more efficient and powerful due to the demands of the AI models which will only grow and demand more power, battery life will be interesting for heavy AI phone users.
The situation is quite challenging because of the ongoing court case with ARM. It seems that ARM is not allowing Qualcomm to use the latest technology, and Qualcomm is counter suing as a result. If the court decides against Qualcomm, there's a possibility that Oryon could be banned while appeals are underway.
What is that dual channel lpddr5x ram running at 5.3 ghz only? Does those channels have twice the pipeline capacity of normal ram? Cuz mediatek got a 10 MT/s ram out there. Edit: 10 Ghz to 10 MT/s
Waiting for the comment: "What do we need this power in Android or smartphone?? Play Candy crush??..." LOL for those users who are like this quoted comment, well smartphones are no longer just your old jurassic touchscreen device for basic things like before. 1. They're now mini computing and multimedia devices capable of running the same resolutions, video capabilities seen on computers and highend cameras so this is one area its needed. 2. Gaming and evolving features on every apps- AAA games only found way back on PCs are now making there way on mobile too. Games like Genshin Impact or Undawn are openworld games same with their PC counterparts with raytracing and large file sizes and are constantly evolving and will require a much powerful CPU as they age. 3. AI, Desktop and Console mode- mobile devices now can be your fullfledge desktop device and console for gaming. Samsung Dex, RedMagic Studio, and Motorola are some that has this.
Wonder what to do with all the performance, but still good to see the competition alive. Apple had their comfortable lead for a while, and they have not been overly innovative since the M series was introduced.
Is the Snapdragon 8 Elite still based on ARMv8.x? Is that because of Qualcomm's dispute with ARM? Meanwhile the A18 Pro and the Dimensity 9400 are both based on ARMv9.x.
The only problem with benchmarks is that they have little effect on the ISP and Ai stuff. Yes it will have an effect but the machine learning algorithms will determine how good the cameras are and the Ai. In other words performance alone cannot tell us how capable the device will be in actual day to day use.
The A18 pro chip is expected to get a boost in performance later on In terms of multicore performance it is expected to reach above 9000........ It is quite nice to see Qualcomm able to beat an iPhone in terms of multicore performance after over a decade but in single core iPhone still leads
@@GaryExplains Apple is bad at low-level optimization, it's not the first time at all that with iOS updates, their SoC achieve better performance, and even both better performance and better efficiency.
Looking at samsung current business strategy, I don't think they want to waste another billions of money to r&d new generation of exynos chips at least for 3-5 years from now. Unless samsung leadership changed.
@@anb4351 everything that utilizes arm new cortex x925 will gain massive single core performance eventually. They just need to find out how to make it more efficient thru process node and combination of core.
@@GaryExplains yes, I'm agree as Oryon on X Elite is Arm v8...a step back compare to 8 Gen 3...as Apple switched to v9, we will probably see very soon a new Geekbench 7 with v9 support (GB 6 doesn't), scores may differ...thanks for the answer and the nice video (as Always)
ok, so the Dimensity's standard ARM cpu is about as powerful as SD's special Oryon-cores. The question is, how does this translate into power-draw? Because, from a pure performance point of view, everything's the same.
9:54 3 million points for GPU perfomance in Antutu...which sums up other points (CPU+GPU+MEM) and doesn't provide the objective results (like RAM Mb/s copy) 3DMark on the other hand provides the average FPS (and looks like the score is actually the FPS) alongside with the overall score points
@@GarySims not only that. Antutu is simply...not a good benchmark. Because what are these points actually represent? CPU - what did they used? Linpack? Dhtrystone? GPU - How complex the scene is? And what it actually tests - avg FPS? Median FPS? min FPS? Does it take into account the stutters? MEM - RAM copy? Latency? Storage - Sequential Write/Read? Random? Which block size of the random write/speed UX - UI FPS? Plus it sums up the points which can result in this - 2 mil phone with great GPU, but the second 2 mil phone has better CPU and storage chip. And the manufacturers can cheat in Antutu to rig the results.
@@deliternity I agree and I haven't included AnTuTu in my videos for years, for the reasons you state, however I opted to include it this time and prefixed with with something, "for those of you who follow the AnTuTu scores" or something like that.
While frequenty is a factor in power consumption so is voltage and the process node. Also more efficient doesn't mean less power in an absolute sense, it means it can achieve higher performance at the same power than the previous generation, or the same performance at a lower power.
That MTK 9400 ST score is anomalously high. The median score is much lower. Whereas that A18P score is at or below media. The real difference in pts/ghz is 7-8% much different than what is shown here.
I have in the past, I have videos about this very subject. It is a myth created by Apple that somehow other tech companies can't create or integrate or optimize software.
Hoffe der Chip wird nicht zu heiß werden. Selbst mein A17 Pro wird warm. Ich vermute je stärker die Chips werden umso schwieriger wird es sie zu kühlen da nicht viel Platz für eine passende Kühlung ist. Aber was weiß ich schon, bin ja kein Ingenieur.
AMD's C-cores follow the same idea. Same design and feature-set as the big cores, just less cache and more density (as you can pack circuits closer togeather when aiming at lower clock-speeds). And AMD's notebook CPU is pretty competitive regarding efficiency.
May finally upgrade my Pixel 6 (thought it still runs fine) next year after 3 years. Hopefully this new SOC is in new OnePlus phones as I'm considering switching to their next R tier phone. I have the OP Watch 2 and love it.
*All these CPU cores and GPU boost just to watch RUclips videos 😂😂 I'm all in for it though.. can't wait to send text messages and WhatsApp messages with my 4.32Ghz CPU cores.*
![Felix "Unfair" | [Stray Kids : SKZ-PLAYER]](http://i.ytimg.com/vi/Oswujxm2Ag0/mqdefault.jpg)
All these cpu for watching youtube 😂
Tik Toc
@@KenjiEspresso that too
No
Considering we see no tangible improvements in GPU graphics or performance as well as CPU performance, this isn't so interesting in the smartphone space. Pretty much any high end Android phone can run any app or game and run it very well.
Nah bro Imma be playing Destiny Rising, Waframe Mobile, Warzone Mobile, Assassin’s Creed games, The Division Mobile
Will Qualcomm ever keep the same naming System for more than 4 years ?
Marketing strategy. Or maybe because snapdragon 8 elite will finally use armv9x
In this case I don't mind, the current x gen y naming nonsense didn't make it much sense when you start looking at the individual products. And the elite lineup is at least in line with their notebook series (thoughts those chips have terrible naming too).
Marketing people calls it "innovation"
No!!! Change it even more often.
Might be finally worth upgrading from SnapDragon 865 to 8 Elite. My multi-core speed isn't as good as the 8 Elite's single-core speed 😢.
Looks like a power hog space heater... The 9400 is way superior... I am saying this as a snapdragon fan who was planning to buy the one plus 13... The fact they didn't even mention CPU power consumption itself is enough proof it's really bad except on power unlimited benchmark scores.. ( I have the same one)
@@soulsbourne it will only use the high wattage during benchmarks. but it will likely be very efficient in gaming and regular use. let's wait and see.
They were repeating 44% more efficiency in the event@@soulsbourne
@@soulsbournewe will have to wait geekerwan’s video, because I saw a post with a caption saying that the 8 elite is far more efficient than both the 9400 and a18 pro and there is a screenshot as well showing the power consumption from the 8 elite; it consumed only 11.9w when it achieved a multi core score of 9600, whereas the a18 pro only achieved 9200 when it also consumed the same wattage as the 8 elite.
@@TheVisualizer2001 Only one source? Have you check other sources? I highly would say the Snapdragon 8 Elite is the equal to Nvidia GTX 690.
Nothing beats Gary Explains!
How long have all the chip manufacturers been telling us that multi-cpu\ multi-thread is the way forward, yet so many things seem to be single core execution.
For example, the Browser (Javascript is single threaded), memory throughput, and emulation - they can benefit from higher single core performance
Because multi threaded programming is incredibly hard
Some things can only run single core can not be broken up into parallel parts to be processed.
Because of battery life. It would probably half the hours of battery and degrade the batteries faster. With how fast CPU's are now multi threaded operations are not as necessary
Javascript : Hi
Thanks Gary for explaining, you are the man!
We should do a comparison of chips by not just normalizing frequency but also power draw, fabrication node etc so we can see whether the uplift really coming from architecture change.
Perf/MHz alone doesn't tell you much, perf/MHz/Watt is a better metric. And according to Qualcomm's power efficiency claims, it looks like they have brought a working Netburst/Bulldozer-esque high clock design to mobile with more pipeline stages while minimizing static power (lower fan-out depth combinatorial logic) and leveraging savings in dynamic power (sequential switching logic). Well done.
There are already antutu and geekbench scores. We can only cross reference them right now.
@@byssmal Only by empirically measuring power draws during these so called "test" scores, conclusions can be made.
@@LogioTek That early benchmark is not far off from the actual chipset. Apple do this too with their bionic chipset before their iphone was officially released at the market. Scores will be slightly different depending on the hardware optimization. A common issues in every smartphones especially android.
@@byssmal Do you not understand that I'm talking about power consumption during benchmarks? Until it's measured and replicated by 3rd parties we simply can't say for sure.
@@LogioTek Qualcomm reducing the cores to 6 of their 8 elite proved that 8 cores will consume more power and overheating. A bold move to mitigate some issues.
These companies have the worst naming schemes, they are making Apple look good comparatively. LOL
Haha wait till you see AMD"s Laptop CPU naming scheme
@@kentakanashi7284 Oh, I have... the AI stuff, 3D... they even make Intel's names seem simple!
Not worse than AMD's upcoming "Ryzen MAX AI HX 395" 🤣
I'm guessing the next one will just be called Snapdragon Elite 2. No need to keep the "8" in the title.
@@elimalinsky7069 Based on their “rules” it will continue to be crazy
Hats off to meditek and arm. They seemed so low key but managed to complete well against so hyped up oryon cores and apple, all this with being significantly efficient as well
The ones Apple uses in the A18 Pro are the Donan P-Core at 4.05 Ghz and the Donan E-Core at 2.2 Ghz.
How do you conclude it's more efficient? More ipc ≠ more power efficient.
Throttling through the roof. LFG
Competition is great for us consumers… Qualcomm and mediatek multicore domination will force Apple to innovate and offer better performance. And also force x86 vendors to align and offer better efficiency. Or x86 will die.
As long as Apple has full control of the hardware and software it will still outperform the rest, what you see there are benchmarks where companies typically increase the wattage for these tests; actual performance is very different.
@@jorvinlaguna1932 Apple is the worst phone device you can go with, not even worth a penny. They overheat like a toaster, poor performance and battery drains very fast during gaming. Also the camera performance isn't as good as android is
@@gezimislamaj5309 bla, bla, bla.
Apple fan boys are worried omg 😂
@@gezimislamaj5309 uh, ok
Hopefully there is no repetition of sd820 and sd888 devastating overheat and battery issue
The 820s heat was waaaaaay better that the 810s heat. The 810 ruined a lot of phones that year.
870 to 8 Elite that would be the best upgrade
Same here but they said that this is nothing they are preparing something too big for the next year so I'm waiting for that you too should wait
i think you shouldn't jump ship so fast brother let's wait and see how the heat up 😂 anyways 870 is still in my opinion the best processor by far without any issues
@@gamerlucky For heat up buy a cooler if you play on 120 fps its gonna heat...buy a good cooler.
@@hotoke666 or buy redmagic with built-in sonic jet fan 🥺
*Always exciting when there is a new architecture, let's see how well it performs in temps, stability and efficiency!*
thx for the video, the single core performance of 8 elite looks great , but I'm a little worry about not having custom design efficiency cores and how this will affect battery life in idle and at lite use.
It'll not affect battery life negatively
Afaik qualcomm designed orion to be scalable. So they can focus for speed or power efficient.
Interesting to see they followed suit with Mediatek and now use a similar concept.
The clock frequencies listed are the maximum achievable frequencies of those cores. Clock frequencies vary and hit peak values very infrequently during any benchmark tests. So, your Single core score per GHz segment 6:16 doesn’t convey anything meaningful as you are simply dividing max frequencies.
It tells you the peak IPC at the same clock freq.
@@GaryExplains first off I love your videos! And thanks for your response. I see your point. But, the benchmark scores were got from dynamically changing clock levels (All 3 CPUs rarely hit Peak frequencies). So, I don’t think considering only max frequency gives us an accurate picture of IPC here, since we’re punishing the ones that offer higher operating frequencies but rarely operate at those levels. The correct way to do this would be to ensure all CPUs operate at their max frequency for the entirety of the benchmark and then calculate max IPC. These are just my thoughts.
Edit (cos, I can’t seem to reply to your chats):
Geekbench 6 uses real-world tests so it’s not gonna load up the CPU to be at Peak frequencies for most tests in their benchmark.
“Geekbench 6 measures your processor's single-core and multi-core power, for everything from checking your email to taking a picture to playing music, or all of it at once. Geekbench 6's CPU benchmark measures performance in new application areas including Augmented Reality and Machine Learning, so you'll know how close your system is to the cutting-edge.”
“Geekbench uses practical, everyday scenarios and datasets to measure performance. Each test is based on tasks found in popular real-world apps and uses realistic data sets, ensuring that your results are relevant and applicable.”
Source: www.geekbench.com
I think for single threaded it is safe to assume that a) the CPU is at its max freq as much as is humanly possible to ensure as they are CPU intensive. b) Any non peak usage (like when one test ends and another starts) will be replicated across all systems as they all run the same types of tests.
I would albo think that single core should hit max frequency, but lower than mediatek result is strange - so reference design might not hit it for entire test duration
Is there any speed comparison of L2 cache of 8 elite and A18 pro?
Is there any chance it could have performed better if the Qualcomm SoC was built on ARMV9 instead of ARMV8?
samsung better be releasing their next tab line up with this elite soc.
Shit, this made me regret buying the OnePlus 12 this year😢
S25 Ultra will be faster due to overclock special Made for Galaxy Snapdragon
@SolarErazer idk how much more they can overclock 😆. It's already in the 4ghz range
@@WayN87 Yeah, it's 4.3G. Why not get to 5Ghz 😂
Hey, Mr. Gary!
Qualcomm is clocking this even higher than their laptop version X Elite. Do you think it will be efficient enough or can sustain performance in phones like S25 as Samsung like to keep the thermal limit lower?
Time will tell!
Leaks have stated that the thermal condition on these chips are really bad, even reaching around 100 C if the leaks were correct. There is no way they can sustain this. Though they have increased the clock speeds significantly, their IPC is not that great compared to Apple and even somehow even to MediaTek as well. MTK can be a serious competitor this time.
This is a better binned chip hence I am sure they are running at lower voltages
Also note X Elite is TSMC N4 and 8 Elite is N3E - this may help a lot with the clockspeeds.
@@TuxikCE so now you are believing in some leaks ?? first wait for actual device than do yapping in comment section
Probably means that all region will get the Snapdragon for the S25 as there is no way Exynos will be able to match the Snapdragon.
Imagine this on speed test g 😂
Qualcom is following Nvidia to include Vram on those GPU? Whats next? TGP for mobile processor?
40%-50% improvement is a bit underwhelming comparatively to hype I had but still great. Should be really great in emulation (can't wait to emulate Bluestacks on winlator lol). Also really impressed by dimensity 9400
Does anybody knows what he use for presentation?
As usual, the primary determinant of the _snappiness_ of a device is the single core performance - multicore is good for slogfests like transcoding or rendering, but not everything can be multithreaded.
Good single core performance nowadays requires multiple decoders, a deep pipeline, good branch prediction, a big reorder buffer, redundant arithmetic engines, and a whole lotta engineering.
at 7:28 processors don't work linearly it may perform higher at a certain optimum frequency. ex if you operate all these cpus at 3ghz all of them will score differently unlike just dividing it by clock speed.
But you can calculate what that particular core does per ghz. So, it tells exactly how well it performance running on 1 ghz. To put it differently: how efficient the core is.
I say go hard and make the best snapdragon you can. I own a tab s9+ and use it for vacation and emulation. I also own a iPhone 16 pro max. Competition is important and keeps these companies innovating.
I have always been an android and iOS user. I also own a Odin 2 max for my primary portable game system.
How about if they run at the same power consumption and not frequency?
Is there a way to compare performance per watt. Cuz performance per clock speed is a meaningless chart in a mobile, as the wattage is the limiting factor.
There is a way yes, but not until actual devices come out.
@@GaryExplains Ahh alright.
I didn't find TDP of 8 Elite, only efficiency difference in percentage is published. Does anyone know its exact power consumption numbers?
I think the next generation of smartphones is the first that will need to be way more efficient and powerful due to the demands of the AI models which will only grow and demand more power, battery life will be interesting for heavy AI phone users.
Hey Gary, do you think the exclusive “For Galaxy” variant single core performance will close the gap with A18Pro SoC?
where is wattage and thermals?
Scores higher than m2 nice
But what apps on android can leverage these performance bump?
I think they may have an issue- their license to manufacture it is being rescinded by ARM in 60 days...unless something dramatic happens at trial.
Nice background image!
I couldn´t find if this oryons are armv9. Does anybody know?
Armv9.2a
I haven't checked yet, but I am assuming Armv8.
The situation is quite challenging because of the ongoing court case with ARM. It seems that ARM is not allowing Qualcomm to use the latest technology, and Qualcomm is counter suing as a result. If the court decides against Qualcomm, there's a possibility that Oryon could be banned while appeals are underway.
the tab s10 should have waited for these, but i guess the 9300+ is good too
can you differentiate between both 8 Elite and D9400 in all areas....
First will be 8 elite & 2nd will be Dimensity :
8 elite / Dimensity 9400
CPU :
Power - 100/86
Efficiency - 100/95
GPU :
Power - 100/100
Efficiency - 97/100
Modem - 100/90
ISP - 100/80
Gaming performance:
100/90
Gaming efficiency:
100/97
Emulation :
100/ almost '0'
Battery backup @ same capacity :
100/100
Let's see how it plays with power consumption and heat.
Indeed, this is just the launch data, much more testing is coming.
The Mediatek is impressive given the clock speed, as is the A18pro with its single core score
Can anyone clarify my doubt, if the 45% cpu boost is on top of 44% power efficient or its 45% cpu boost at the same power as 8 gen 3 and vice versa .
11:57 hey I know that sound
gyatt dayum that titan's level of threat dat snapdragon is packing mother of all performance
Is it the same CPU or SoC architecture as those SD laptops?
80% same. Think of it like zen to zen+
I was wrong Gary. SD on 3nm owns the competition!
Really REALLY Interested in Dimensity 9400 and this
Also I'm wondering how big are those Orion cores compared to the 925 Core on 9400
They promote that the GPU power saving is to 40% but they never mention how hungry for power that CPU is ?
So how energy efficient that CPU is?
Finally, here it is
This channel is the best
Will this chip be in the next s25?
The 12 MB Graphics memory looks something Similar to Amd Infinity Cache on Radeon Graphics Cards. Which actually heps ALOT is high fps gaming
Hi Gary, great video! Quick question: What are your thoughts on the rumored Royal Core project that Intel canceled, and their rentable unit concept?
Yes!! but can it compete with my Samsung Note 10 from 2020 ???
What is that dual channel lpddr5x ram running at 5.3 ghz only? Does those channels have twice the pipeline capacity of normal ram? Cuz mediatek got a 10 MT/s ram out there. Edit: 10 Ghz to 10 MT/s
DDR=double data rate. 5300Mhz=10600MT/s
If nothing, at least android emulation is gonna be great.
What are you doing with android emulation? Just curious
Waiting for the comment: "What do we need this power in Android or smartphone?? Play Candy crush??..."
LOL for those users who are like this quoted comment, well smartphones are no longer just your old jurassic touchscreen device for basic things like before.
1. They're now mini computing and multimedia devices capable of running the same resolutions, video capabilities seen on computers and highend cameras so this is one area its needed.
2. Gaming and evolving features on every apps- AAA games only found way back on PCs are now making there way on mobile too. Games like Genshin Impact or Undawn are openworld games same with their PC counterparts with raytracing and large file sizes and are constantly evolving and will require a much powerful CPU as they age.
3. AI, Desktop and Console mode- mobile devices now can be your fullfledge desktop device and console for gaming. Samsung Dex, RedMagic Studio, and Motorola are some that has this.
There have been several of those types of comment already.
Speed test g, speed test g, speed test g
Can't wait for optimized Turnip drivers for the Adreno 830 🔥
Emulation monster!!!
just hope this didn’t give the reason for manufacturer to cut battery down
Whatever happened to Speed Test G
It ended because it didn’t gain enough popularity to justify the time and effort it required.
Wonder what to do with all the performance, but still good to see the competition alive. Apple had their comfortable lead for a while, and they have not been overly innovative since the M series was introduced.
Is the Snapdragon 8 Elite still based on ARMv8.x? Is that because of Qualcomm's dispute with ARM? Meanwhile the A18 Pro and the Dimensity 9400 are both based on ARMv9.x.
Yes. Working on it in the court.
Also phoenix should launch 2.5 years ago, but because of dispute they can't. So they just increase the clock
Now the S 8 Elite just needs to be equipped with proper cooling, a tandem OLED display, a microSD slot, and long-term software support.
Pretty please.
Why insted of cores You says coooos?
The only problem with benchmarks is that they have little effect on the ISP and Ai stuff. Yes it will have an effect but the machine learning algorithms will determine how good the cameras are and the Ai. In other words performance alone cannot tell us how capable the device will be in actual day to day use.
Obviously.
The A18 pro chip is expected to get a boost in performance later on
In terms of multicore performance it is expected to reach above 9000........ It is quite nice to see Qualcomm able to beat an iPhone in terms of multicore performance after over a decade but in single core iPhone still leads
"The A18 pro chip is expected to get a boost in performance later on" Really? Says who? How? Why later?
@@GaryExplainsthat boost already happened .
GB 6 scores :
iOS 18- 9100 @ 12.05 w
iOS 18.1 - 9350 @ 11.8 w
Further snapdragon 8 elite on GB 6 :
9350 @ 10.6w
10500 @ 16w
@@GaryExplains Apple is bad at low-level optimization, it's not the first time at all that with iOS updates, their SoC achieve better performance, and even both better performance and better efficiency.
I am on exynos, hopefully Samsung will also do something to compete
Looking at samsung current business strategy, I don't think they want to waste another billions of money to r&d new generation of exynos chips at least for 3-5 years from now. Unless samsung leadership changed.
@@anb4351 everything that utilizes arm new cortex x925 will gain massive single core performance eventually. They just need to find out how to make it more efficient thru process node and combination of core.
Still no 10bit 4:2:2 Hardware ProRes or DnXHR Video Encoder/Decoder :(
Isn't ProRes an Apple codec? Why would you expect support in a Qualcomm chip for that? If Qualcomm did support it, what would be the use cases?
All mobile gpus have a local memory. Usually way smaller for tile memory. All mobile gpus also had l2 cache so the 8.20 comment was kinda odd
Great! Now my commodore 64 emulation can run smooth
But how it will compete with newest Mediatek?
Eh?
Is it Arm v8 or v9 ?
I haven't double checked by I am assuming Armv8.
@@GaryExplains yes, I'm agree as Oryon on X Elite is Arm v8...a step back compare to 8 Gen 3...as Apple switched to v9, we will probably see very soon a new Geekbench 7 with v9 support (GB 6 doesn't), scores may differ...thanks for the answer and the nice video (as Always)
Geekbench 6 has Armv9 support.
@@GaryExplains are you sure ? On Geekbench 6's results, my Dimensity 9200+ is considered as ArmV8 not V9
Have you tried 6.3?
GARY !!! GOOD MORNING PROFESSOR
Good morning!
Very impressed by mediatek. Standing among the top players out of nowhere.
Elite, i believe, this naming was created years ago
It isn't the first time that Qualcomm is using the word Elite, no. There is Snapdragon Gaming Elite, the Snapdragon X Elite, etc.
So they are calling efficiency cores performance cores? That's genius marketing
ok, so the Dimensity's standard ARM cpu is about as powerful as SD's special Oryon-cores. The question is, how does this translate into power-draw? Because, from a pure performance point of view, everything's the same.
9:54 3 million points for GPU perfomance in Antutu...which sums up other points (CPU+GPU+MEM) and doesn't provide the objective results (like RAM Mb/s copy)
3DMark on the other hand provides the average FPS (and looks like the score is actually the FPS) alongside with the overall score points
I should have labelled that slide Other Benchmarks, not just GPU Performance. Sorry about that.
@@GarySims not only that. Antutu is simply...not a good benchmark. Because what are these points actually represent?
CPU - what did they used? Linpack? Dhtrystone?
GPU - How complex the scene is? And what it actually tests - avg FPS? Median FPS? min FPS? Does it take into account the stutters?
MEM - RAM copy? Latency?
Storage - Sequential Write/Read? Random? Which block size of the random write/speed
UX - UI FPS?
Plus it sums up the points which can result in this - 2 mil phone with great GPU, but the second 2 mil phone has better CPU and storage chip.
And the manufacturers can cheat in Antutu to rig the results.
@@deliternity I agree and I haven't included AnTuTu in my videos for years, for the reasons you state, however I opted to include it this time and prefixed with with something, "for those of you who follow the AnTuTu scores" or something like that.
@@GarySims Thanks. Much appreciated.
(but I still don't get why people are still falling to the big numbers...)
Despite having a higher frequency it will consume less power. How?
While frequenty is a factor in power consumption so is voltage and the process node. Also more efficient doesn't mean less power in an absolute sense, it means it can achieve higher performance at the same power than the previous generation, or the same performance at a lower power.
lets see what will Google offer in next year ...... can we expect at least 8gen 2 performance in tensor G5
That MTK 9400 ST score is anomalously high. The median score is much lower. Whereas that A18P score is at or below media. The real difference in pts/ghz is 7-8% much different than what is shown here.
The Dimensity numbers are from MediaTek itself, I mentioned that in the video.
@@GaryExplains fair enough.
It looks like the AMD of ARM are finally having their Zen moment.
Will it integrate well with the software?
Yes, why wouldn't it?
@@GaryExplains Since you are the expert why don’t you tell us how?
I have in the past, I have videos about this very subject. It is a myth created by Apple that somehow other tech companies can't create or integrate or optimize software.
ARM says hold my beer....
Snapdragon 8 elite already has beaten my potato PC
These will be very Expensive.
I hope its energy efficiency can match D 9400
Gary, PLEASE explain!
For all the new faster cores I was expecting to see huge jumps over the others CPUs. Not so much.
Hoffe der Chip wird nicht zu heiß werden.
Selbst mein A17 Pro wird warm.
Ich vermute je stärker die Chips werden umso schwieriger wird es sie zu kühlen da nicht viel Platz für eine passende Kühlung ist.
Aber was weiß ich schon, bin ja kein Ingenieur.
Snapdragon 8 elite seems more promising on tablets where they have better thermal capacity than smartphone for higher sustained performance.
No efficiency cores hurt my eyes the efficiency must be awful compared to apple
Nope, "effcicency" cores are cost effective to padd out the specsheet to an octa core but don't add much value otherwise
They help with low power efficiency that is the most used in phones then it definitely helps a lot @@floppa9415
AMD's C-cores follow the same idea. Same design and feature-set as the big cores, just less cache and more density (as you can pack circuits closer togeather when aiming at lower clock-speeds). And AMD's notebook CPU is pretty competitive regarding efficiency.
@@ThorDyrden
AMD's CPUs are less efficient than Intel's ones at lower power stages, we still need efficiency cores.
@@aymanelzain1271 No. Even Intel is moving away from e cores in a few generations. They are moving to something closer to AMD c cores.
May finally upgrade my Pixel 6 (thought it still runs fine) next year after 3 years. Hopefully this new SOC is in new OnePlus phones as I'm considering switching to their next R tier phone. I have the OP Watch 2 and love it.
8 elite 8 epic 8 legend 😁
*All these CPU cores and GPU boost just to watch RUclips videos 😂😂 I'm all in for it though.. can't wait to send text messages and WhatsApp messages with my 4.32Ghz CPU cores.*