The 10 core Helio X20 might be in a little trouble according to a rumour on Weibo suggesting Xiaomi, HTC and Lenovo have scrapped plans to use the chip.
With 10 cores running in a tri-cluster, and being the successor of the hugely successful Helio X10, there is a lot of hot anticipation around the Mediatek Helio X20 chipset, but perhaps a little too hot!
According to rumours on Weibo this morning, Xioami, HTC and Levnovo (three of the big smartphone players) have had to scrap plans for a Helio X20 powered phone due to optimisation issues resulting in overheating.
Apparently the companies don’t want to go through a similar overheating story so many other brands did in 2015 with overheating Snapdragon 810 SoC’s.
Gizchina News of the week
The rumour is of course unconfirmed, but it would explain why the first phone makers trumped to launched a Helio X20 phone are the likes of Zopo and Doogee.
Both Zopo and Doogee plan to launch Mediatek Helio X20 phones at MWC this month. We’ll be at the show and will see if we can confirm or deny the issues.
Oh Damn, I purposely don’t buy Redmi note 2 because been waiting for X20
Truly, i do not believe those rumours. Xiaomi and Lenovo (not sure about HTC) will release at least a phone with X20 soc (the when however is a question). So they are not scrapping the x20 soc. But i do believe the soc it’s gonna have some heating issues at heavy loading and that may delay Xiaomi and other developers showing a phone with that soc soon. And i also do not like that’s is presented to be a flagship soc but so bellow the competition (SD 820, Exynos 8890, even Kirin 950) in Benchmarks (i know, it’s not right to only look at benchmarks but why THAT much difference with competition in, let’s say Antutu?).
Why, o , why you good mediatek stay behind the competition? Plz speed up the game or you gonna loose it!
Me by you should. They clocked it far to hi, two quad A53 clusters just waist more energy & 20 nm bulk process is not better than 28nm HPL.
I am still claiming that S652 will be the SoC of the year (based on user needs & experience).
Well, all the data so far (unfortunately) agree with your opinion that the SD652 will be the soc of the year. Mediatek falled short to competition this time around and they really should do something about it soon.
However i do not agree with the nm. The lowest the nm the better a soc can be at energy consumption and heating issues. Except if they do not do something right. But maybe the 20nm is not enough for the Helio X20 to have all 10 cores cool.
You people think that it’s all about nm!
It’s much more to it than that from used process, used libs, tools and additional optimization blocks.
Now read:
http://www.anandtech.com/show/8718/the-samsung-galaxy-note-4-exynos-review/2
As it is their is no PHLL (or how TSMC like to call it HPM) 20nm process only the bulk one & even it will be more power efficient than Samsung ones it still won’t be much compared to 28 nm HPLL. Certainly not enough to support such a big frequencies MTK is pushing.
What is all the data? Supposed leaked reports that can’t be verified? Spec sheets of chips? Until we have actual phones in hands with these chips we can do nothing but speculate. Last year at this time everyone speculated that Qualcomm was going to blow away Mediatek.
So Snapdragon 650s all round?
I have been singing this since the Xiaomi Redmi Note 3 Pro was announced…
I have been saying this since October.
Of course not… There’s plenty of room for sd652 too! ?
I wanted to mention it but forgot what number it was, was thinking 657 for some reason lol
Probably because they changed the name after some time, and it never helps to remember names ?
Hopefully it can be sorted out and more SoC for our reasonable priced phones.
LOL, so much for the MediaTek fan boys on this site taking shots at the SD 810, which by the way, is doing brilliantly in a number of well optimized phones.
MediaTek’s reputation is already questionable due to its historically bad chips – this is deserved insult to injury.
Never understood that the heating allegations against the SD 810 was all about. I mean yes, it does warm up like ANY other SoC during intense gaming and all, but the issues are hyperbole – room heater and all – never had problems to the extent that I’d be irritated. No throttles, no surface-of-the-sun temps.
Happy Xiaomi Note Pro user here, the phone which has been reported a lot for overheating etc.
One Plus Two user here – never had any issues, except throttling during some badly developed games. None of the A-list titles like MC or Asphalt posed issues. However, badly optimized games DID throttle the SoC a number of times, predictably.
Never understood the pattern really, as I was expected the most demanding games to actually throttle the SoC, not some sidekick games from unknown developers.
It could be that games from Top Developers are more optimized in spite of being more “visually” demanding, and therefore they manage to keep the load on the chip fairly low? I’ve read a number of times on XDA that smaller game firms don’t bother too much with optimizing.
Just guessing – could be wrong.
There are games that are only graphic intense (like for instance driving simulations) & ones that are graphic & CPU intense (like for instance real time strategies). Same goes for GPU benchmarks. That should be a answer to your question.
That’s because OnePlus used a heatsink, graphite and thermal paste to deal with the heat. Most phones don’t have any of that. I used a OnePlus Two for a week, and it wasn’t nearly as bad as the HTC, but I still felt it heat up beyond normal at times.
Also, oneplus used a sw trick: they allowed the phone to use just 2 of the big cores each time, and it helped a lot… I’ve read many reports of this behavior on oneplus forum
Several phones warmed up a lot more than normal SoC’s do. The HTC I used I had to put down after 20 mins of gaming, I have never had to do that on another phone. There were even detailed tests done that measured the heat levels of SD810 devices vs SD801 devices and the difference was extreme. Qualcomm and manufacturers came up with fixes for the issues with different cooling options (thermal paste, heatsinks, underclocking etc). For the Mi Note Pro Xiaomi applied for 5 different thermal cooling patents, which is why you might not feel as much heat.
There were also numerous users who reported issues with components inside 810 phones failing because of the heat. There was definitely something to the allegations.
I like competition and i want Mediatek to be up there with good socs and better prices. I have owned MTK based soc phones and been very happy with them.
So it’s not about being a fan boy (you shouldn’t be either), competition is the best thing that can happen to us users…
I am not a fanboy of either company, I like them both and I like Samsung the best (for SoC’s not for their terrible phones). But this is an unconfirmed report from Weibo, we have no devices to actually test it out. The 810 we have actual devices that heat up.
Use logic. If S810 overheated @ 2GHz & even on 1.8GHz wile S808 run good on 1.8 GHz using the same manufacturing process as X20 wile difference between A57 & A72 power consumption is about 20% the X20 will be able to safely reach some 2 GHz, 2.2GHz in the best case (that I doubt).
Clock by clock A72 is much more efficient than A57 rather than mere 20%.
Kirin 950 vs Exynos 7420 power consumption charts shows that A72 draws almost half of the power of A57 at 2.1GHz, assuming Samsung 14nm and TSMC 16nm has been roughly comparable.
Well ARM says 20~25%.
Its Samsungs last year’s 14nm LPE against STMC-s this year FF+ & difference is about 8-9%.
It looks like HiSilicon used a very advanced libs for rooting optimizing it for size & power consumption what we can see & on big cluster die size & this can caunt for up to additional 22%. Good news for HiSilicon is that at least they listen to critics so they licensed up new Tensilica DSP cores that will find its place on Kirin 950 successors.
Lol…
Not surprise at all! The 20nm process is notorious for being hard to keep heat under control. If qualcomm can’t do it with octa core, then it would be quite amazing if Mediatek can do it with 10 cores and high frequency! Well, guess the old adage still right “if it’s too good to be true then…”!
Samsung did it with the Octacore Exnyos 5433 so it is entirely possible.
well, they didn’t
http://www.anandtech.com/show/8718/the-samsung-galaxy-note-4-exynos-review/6
http://www.anandtech.com/show/9878/the-huawei-mate-8-review/3
The Exnyos wasn’t an Octacore 20nm chip? Samsung seems to disagree with you.
http://www.samsung.com/global/business/semiconductor/minisite/Exynos/w/solution.html#?v=7octa
the point was that it was overheating, quite badly. look at power in those links. Kirin goes to bellow 4W while this Samsung went totally nuts. The 7420 on 14nm was much better but still too high http://www.anandtech.com/show/9330/exynos-7420-deep-dive/5
http://images.anandtech.com/doci/8718/A57-power-curve_575px.png
Those links are the first I ever heard of it overheating. The Galaxy Note 4 I used had no issues with it.
the fact that you don’t notice doesn’t mean it’s not there.
It has nothing to do with the temp of the phone’s case. A chip throttles at a certain temp to protect itself. It can be at 100C and you wouldn’t know it by touching the phone.
If you want to see how it behaves in actual usage , monitor clocks and actual temps.
If you want a quick test , use something like CPUBurn to load a number of cores and let it run for a some time while monitoring clocks. For the GPU GFXbench long term perf is a good indicator.
I don’t have the Note 4 anymore, my wife had one for a few months before we traded it in and got her an S6.
Maybe you are curious to test on your current devices, if you do it would be nice to see the results in video editing. You should monitor GPU clocks too, see if the GPU offloads some of the work. I’ve actually tried to find a site that tests video editing on Android but could not find any. Would be a relevant thing to bench.
Just for fun how much more more energy Exunos 7420 drives at is maximum frequency with all cores of the big cluster running against same situation when same cores run at 1.6GHz?
http://images.anandtech.com/doci/9330/a57-power-curve.png
Now difference between max clock’s on S652 & S650 parts & X20 are even bigger some 700MHz. Looks like I whose wrong it will be even more than 75% but we can settle it down to some 65~70% when we put into the calculations different manufacturing node’s.
Conclusion the X20 will be one hot chip.
Still it’s easy to resolve this just lowering the clock’s to let’s say 2GHz & problem is gone.
You ignore the number of cores and the fact that it’s a diff core. How hot is the SD652 with all 4 big cores loaded even by your math if x20 would be that hot?
Maybe try going backwards from 16ff+ ( not that these graphs are that great for in illustrating per core power since they don’t quite isolate just the cores).
http://images.anandtech.com/doci/9878/power-big.png
Smart-ass as you see adding more ticks on quarts crystal than 1.6GHz have a significant jump on power consumption. 950~1000 is on green side, 1000 to 1300 is tolerable, 1300 to 1600 is high & over that it’s insane. This doesn’t change much with process as it’s physically associated to silicone. It’s not my math look at the diagram I posted.
It’s a same cores A72s.
Why would I go back to the 16nm FF+ when X20 uses 20nm bulk lithography?
The S652 will have a rather much larger die size (34% from lithography & at least 35% more on actual components) over the X20 & that will give him a more space to lose heat. Even if it throttle down (witch won’t be a case in real usage scenarios) it will be a little to 1.6GHz & only in extreme cases still 4×1.6 is more than 2×2.1.
Now consider all innovations that S652 is bringing to none that X20 is bringing & they will make a difference in real usage along with better user experience.
lol, i get that this rumor feeds your Mediatek hate but ….
You just blast anything MTK related with hate. Anything and everything.
But you are just going all over the place and this conversation is pointless since you have certain fixations and you do not address in any way any objections i might raise.
You would keep going even if you had actual results and always find some new ways to hate.
You really think there is nothing wrong with you if you have feelings for a corporation?
And please do not engage me anymore,take your hate speech elsewhere.
I want to fixate you to a wall.
I don’t hate MTK but I am beginning to hate you.
I am perfectly fine thanks.
Would really like to engage you in real life and wipe the floor with you.
I perfectly normally told that STMC’s 20 nm bulk process is better than Samsungs one but it’s still not significantly better than 28nm HPM so you are the one that whosent been reading. Still it’s more normal to compare those two 20 nm processes than any FinFET to them. So your objections are rather imaginary ones. By the way I did touch the Kirin power projections & explained them just couple of comments under. Next time do the math and come with facts if you want to say something to me & you will have a big up vote from me.
Now if you insults me one more time I will track you down!
Just to ruin your sleep, MTK just had a talk on the x20 at ISSCC and they still claim 2.5GHz. If you are there, go talk to them, if not, tough shit.
As for your threats , thank you for that ,nothing has made be lol that hard in a long time. And btw, saying those things only reinforces the idea that you have certain anger issues. Could be a matter of education or age but something is off.
It might be a matter I had with your mother long time ago.
I toth shit how you eat one. Now I need to stop with that.
just for fun check out this graph with SD810 allowed to go crazy instead of throttling with 1 and 2 cores loaded, with 2 cores it hits 105C and shuts down in seconds http://img1.mydrivers.com/img/20150616/e962772bff78480b8936089e41d53ec0.jpg
The main problem with overheating is not the overheating itself but the lack of disclosure. Temperature based throttling is actually an acceptable practice and works well if managed properly , Intel introduced Turbo a long time ago and what they are doing now is a lot more evolved. Nvidia in GPU has Nvidia Boost and so on. So as long as the consumer knows what is going on and the chip maker promptly discloses what is being done , it would be fine.Sadly in mobile what folks are doing is not ok, it’s misleading , it’s fraud and you can call their products defective. Plus they don’t seem to be employing the technique on purpose, they mess up and pretend the problem is not there. If they would do it on purpose and provide sufficient info on what they are doing, it would be beneficial.You provide a short term boost ,when needed and the end result is a better experience and better perf. As for x20, we’ll see soon. If it blows, we have other options.
They did it by scaling down to 32 bit memory controllers.
Would be funny, anything is possible on 20nm but remains to be seen. It seems to have been delayed a bit so maybe they did a respin and fixed it but nothing is certain before we have devices to test. With just 2 cores they should reach high enough clocks evenif they miss 2.5Ghz.
News coming from the trustworthy weibo account @totallynotqualcomm ?
I’m going to believe it when I’ll see it… I wouldn’t like it that much, but I’ve got to admit that it would be fun to have another year of puns about heating (“helio” in the name is not that reassuring anyway ?)
Anyway, let’s talk about strategies:
-xiaomi could scrap x20 because of their problems of selling mediatek based devices to India, and they already did their move with note 3 pro and sd650… Most likely, there will be a mi5c with sd652, but the line up is already crowded enough: there’s hardly other room for the x20.
-lenovo could use the x20 only on the vibe series, but now that they’re planning to use the “moto” brand for the higher end devices, there’s chances we won’t see any good mediatek soc on the vibe brand, sticking it to entry level and lower mid ranger (so the top would be p10, if this scenario is real). So, lenovo could scrap x20 too
-htc is a famous brand, has a partnership with mediatek, it’s trying to make a profit in every way possible cutting costs here and there, and has no problem to sell problematic phones with overheating issues. So, even if those problems are true, htc will definitely sell a variant of his m10 line up with the x20 on board, no matter what.
Here you have it: the rumor could be partially true, but this doesn’t mean x20’s problem are true! ?
LOL @ HTC!
I would hit that
I will disbelieve it when an X20 phone gets in your hand.
Oh Damn, I purposely don’t buy Redmi note 2 because been waiting for X20
Truly, i do not believe those rumours. Xiaomi and Lenovo (not sure about HTC) will release at least a phone with X20 soc (the when however is a question). So they are not scrapping the x20 soc. But i do believe the soc is gonna have some heating issues at heavy loading and that may delay Xiaomi and other brands showing a phone with that soc soon. And i also do not like that it’s presented to be a flagship soc but so bellow the competition (SD 820, Exynos 8890, even Kirin 950) in Benchmarks (i know, it’s not right to only look at benchmarks but why THAT much difference with competition in, let’s say Antutu?).
Why, o , why you good mediatek stay behind the competition? Plz speed up the game or you gonna loose it!
Me by you should. They clocked it far to hi, two quad A53 clusters just waist more energy & 20 nm bulk process is not better than 28nm HPLL.
I am still claiming that S652 will be the SoC of the year (based on user needs & experience).
Well, all the data so far (unfortunately) agree with your opinion that the SD652 will be the soc of the year. Mediatek falled short to competition this time around and they really should do something about it soon.
However i do not agree with the nm. The lowest the nm the better a soc can be at energy consumption and heating issues. Except if they do not do something right. But maybe the 20nm is not enough for the Helio X20 to have all 10 cores cool.
You people think that it’s all about nm!
It’s much more to it than that from used process, used libs, tools and additional optimization blocks.
Now read:
http://www.anandtech.com/show/8718/the-samsung-galaxy-note-4-exynos-review/2
As it is their is no PHLL (or how TSMC like to call it HPM) 20nm process only the bulk one & even it will be more power efficient than Samsung ones it still won’t be much more power efficient compared to 28 nm HPLL. Certainly not enough to support such a big frequencies MTK is pushing.
What is all the data? Supposed leaked reports that can’t be verified? Spec sheets of chips? Until we have actual phones in hands with these chips we can do nothing but speculate. Last year at this time everyone speculated that Qualcomm was going to blow away Mediatek.
This is what happens when you try and do ‘flagship’ on the cheap. Qualcomm couldn’t quite manage it on a 20nm node and it seem MT have hit the same problems. If they wan’t to compete with the big boys they have to use the more expensive 14/16nm nodes that Qualcomm and Samsung are using. There was never any need for 2 quad A53 cores when they could’ve used dual cores instead and this would at least have allowed them to use more GPU cores since this is always their weakness. The fact the Huawaei built the Kirin 950 at 16nm with lower clocked A72 cores tells you all you need to know.
So Snapdragon 650s all round?
I have been singing this since the Xiaomi Redmi Note 3 Pro was announced…
I have been saying this since October.
Of course not… There’s plenty of room for sd652 too! π
I wanted to mention it but forgot what number it was, was thinking 657 for some reason lol
Probably because they changed the name after some time, and it never helps to remember names π
UGH…I should had known, no wonder only few companies will had that chipset
Perhaps a new revision will solve out (X22) ….
But then again i hardly doubt it.
Still…I hope this rumor is not true, so we’ll just wait
it looks like it needs to delay just a bit more…and revamp it just a little
you know uh…16nm, LPDDR4, and perhaps T880MP6…Just Saying, BTW! PHEW!
Hopefully it can be sorted out and more SoC for our reasonable priced phones.
LOL, so much for the MediaTek fan boys on this site taking shots at the SD 810, which by the way, is doing brilliantly in a number of well optimized phones.
MediaTek’s reputation is already questionable due to its historically bad chips – this is deserved insult to injury.
Never understood that the heating allegations against the SD 810 was all about. I mean yes, it does warm up like ANY other SoC during intense gaming and all, but the issues are hyperbole – room heater and all – never had problems to the extent that I’d be irritated. No throttles, no surface-of-the-sun temps.
Happy Xiaomi Note Pro user here, the phone which has been reported a lot for overheating etc.
One Plus Two user here – never had any issues, except throttling during some badly developed games. None of the A-list titles like MC or Asphalt posed issues. However, badly optimized games DID throttle the SoC a number of times, predictably.
Never understood the pattern really, as I was expected the most demanding games to actually throttle the SoC, not some sidekick games from unknown developers.
I like competition and i want Mediatek to be up there with good socs and better prices. I have owned MTK based soc phones and been very happy with them.
So it’s not about being a fan boy (you shouldn’t be either), competition is the best thing that can happen to us users…
It could be that games from Top Developers are more optimized in spite of being more “visually” demanding, and therefore they manage to keep the load on the chip fairly low? I’ve read a number of times on XDA that smaller game firms don’t bother too much with optimizing.
Just guessing – could be wrong.
There are games that are only graphic intense (like for instance driving simulations) & ones that are graphic & CPU intense (like for instance real time strategies). Same goes for GPU benchmarks. That should be a answer to your question.
I am not a fanboy of either company, I like them both and I like Samsung the best (for SoC’s not for their terrible phones). But this is an unconfirmed report from Weibo, we have no devices to actually test it out. The 810 we have actual devices that heat up.
Reputation in the present day favors Mediatek. All of their 64 bit chips were excellent. Qualcomm on the other hand suffered from problem after problem with their 64 bit chips. Save your Qualcomm fanboyism until we have actual devices to test out.
That’s because OnePlus used a heatsink, graphite and thermal paste to deal with the heat. Most phones don’t have any of that. I used a OnePlus Two for a week, and it wasn’t nearly as bad as the HTC, but I still felt it heat up beyond normal at times.
Several phones warmed up a lot more than normal SoC’s do. The HTC I used I had to put down after 20 mins of gaming, I have never had to do that on another phone. There were even detailed tests done that measured the heat levels of SD810 devices vs SD801 devices and the difference was extreme. Qualcomm and manufacturers came up with fixes for the issues with different cooling options (thermal paste, heatsinks, underclocking etc). For the Mi Note Pro Xiaomi applied for 5 different thermal cooling patents, which is why you might not feel as much heat.
There were also numerous users who reported issues with components inside 810 phones failing because of the heat. There was definitely something to the allegations.
Use logic. If S810 overheated @ 2GHz & even on 1.8GHz wile S808 run good on 1.8 GHz using the same manufacturing process as X20 wile difference between A57 & A72 power consumption is about 20% the X20 will be able to safely reach some 2 GHz, 2.2GHz in the best case (that I doubt).
Also, oneplus used a sw trick: they allowed the phone to use just 2 of the big cores each time, and it helped a lot… I’ve read many reports of this behavior on oneplus forum
Clock by clock A72 is much more efficient than A57 rather than mere 20%.
Kirin 950 vs Exynos 7420 power consumption charts shows that A72 draws almost half of the power of A57 at 2.1GHz, assuming Samsung 14nm and TSMC 16nm has been roughly comparable.
Well ARM says 20~25%.
Its Samsungs last year’s 14nm LPE against STMC-s this year FF+ & difference is about 8-9%.
It looks like HiSilicon used a very advanced libs for rooting optimizing it for size & power consumption what we can see & on big cluster die size & this can caunt for up to additional 22%. Good news for HiSilicon is that at least they listen to critics so they licensed up new Tensilica DSP cores that will find its place on Kirin 950 successors.
Lol…
Not surprise at all! The 20nm process is notorious for being hard to keep heat under control. If qualcomm can’t do it with octa core, then it would be quite amazing if Mediatek can do it with 10 cores and high frequency! Well, guess the old adage still right “if it’s too good to be true then…”!
Samsung did it with the Octacore Exnyos 5433 so it is entirely possible.
well, they didn’t, at 1.5Ghz it would have been fine, at 1.9GHz, as it shipped,not so much
http://www.anandtech.com/show/8718/the-samsung-galaxy-note-4-exynos-review/6
http://www.anandtech.com/show/9878/the-huawei-mate-8-review/3
The Exnyos wasn’t an Octacore 20nm chip? Samsung seems to disagree with you.
http://www.samsung.com/global/business/semiconductor/minisite/Exynos/w/solution.html#?v=7octa
the point was that it was overheating, quite badly. look at power in those links. Kirin goes to bellow 4W while this Samsung went totally nuts. The 7420 on 14nm was much better but still too high http://www.anandtech.com/show/9330/exynos-7420-deep-dive/5
Those links are the first I ever heard of it overheating. The Galaxy Note 4 I used had no issues with it.
The main problem with overheating is not the overheating itself but the lack of disclosure. Temperature based throttling is actually an acceptable practice and works well if managed properly , Intel introduced Turbo a long time ago and what they are doing now is a lot more evolved. Nvidia in GPU has Nvidia Boost and so on. So as long as the consumer knows what is going on and the chip maker promptly discloses what is being done , it would be fine.Sadly in mobile what folks are doing is not ok, it’s misleading , it’s fraud and you can call their products defective. Plus they don’t seem to be employing the technique on purpose, they mess up and pretend the problem is not there. If they would do it on purpose and provide sufficient info on what they are doing, it would be beneficial.You provide a short term boost ,when needed and the end result is a better experience and better perf. As for x20, we’ll see soon. If it blows, we have other options.
the fact that you don’t notice doesn’t mean it’s not there.
It has nothing to do with the temp of the phone’s case. A chip throttles at a certain temp to protect itself. It can be at 100C and you wouldn’t know it by touching the phone.
If you want to see how it behaves in actual usage , monitor clocks and actual temps.
If you want a quick test , use something like CPUBurn to load a number of cores and let it run for a some time while monitoring clocks. For the GPU GFXbench long term perf is a good indicator.
You said that you do video editing, try to monitor clocks while doing that, would be interesting to know what you get so if you do it, do share the results.
I don’t have the Note 4 anymore, my wife had one for a few months before we traded it in and got her an S6.
Maybe you are curious to test on your current devices, if you do it would be nice to see the results in video editing. You should monitor GPU clocks too, see if the GPU offloads some of the work. I’ve actually tried to find a site that tests video editing on Android but could not find any. Would be a relevant thing to bench.
just for fun check out this graph with SD810 allowed to go crazy instead of throttling with 1 and 2 cores loaded, with 2 cores it hits 105C and shuts down in seconds
Just for fun how much more more energy Exunos 7420 drives at is maximum frequency with all cores of the big cluster running against same situation when same cores run at 1.6GHz?
Now difference between max clock’s on S652 & S650 parts & X20 are even bigger some 700MHz. Looks like I whose wrong it will be even more than 75% but we can settle it down to some 65~70% when we put into the calculations different manufacturing node’s.
Conclusion the X20 will be one hot chip.
Still it’s easy to resolve this just lowering the clock’s to let’s say 2GHz & problem is gone.
You ignore the number of cores and the fact that it’s a diff core. How hot is the SD652 with all 4 big cores loaded even by your math if x20 would be that hot? And you are adding MHz not in %, that’s ridiculous. Plus, you know very well that SD650 could actually be clocked higher than 1.8GHz, they don’t do it to not hurt SD652.So if you can do better than 1.8GHz on 28nm with 2 cores , hard to argue that on 20nm you can only match it.If you want to argue that 2.5GHz on 20nm could be problematic, sure it could be ,remains to be seen if they get to it or not. But to argue that 2GHz is all you can get with 2 cores on 20nm, is a bit much and they would need to mess up the memory controller or something else in a big way.
Maybe try going backwards from 16ff+ ( not that these graphs are that great for illustrating per core power since they don’t quite isolate just the cores).
Smart-ass as you see adding more ticks on quarts crystal than 1.6GHz have a significant jump on power consumption. 950~1000 is on green side, 1000 to 1300 is tolerable, 1300 to 1600 is high & over that it’s insane. This doesn’t change much with process as it’s physically associated to silicone. It’s not my math look at the diagram I posted.
It’s a same cores A72s.
Why would I go back to the 16nm FF+ when X20 uses 20nm bulk lithography?
The S652 will have a rather much larger die size (34% from lithography & at least 35% more on actual components) over the X20 & that will give him a more space to lose heat. Even if it throttle down (witch won’t be a case in real usage scenarios) it will be a little to 1.6GHz & only in extreme cases still 4×1.6 is more than 2×2.1.
Now consider all innovations that S652 is bringing to none that X20 is bringing & they will make a difference in real usage along with better user experience.
lol, i get that this rumor feeds your Mediatek hate but ….
You just blast anything MTK related with hate. Anything and everything.
But you are just going all over the place and this conversation is pointless since you have certain fixations and you do not address in any way any objections i might raise.
You would keep going even if you had actual results and always find some new ways to hate.
You really think there is nothing wrong with you if you have feelings for a corporation?
And please do not engage me anymore,take your hate speech elsewhere.
I want to fixate you to a wall.
I don’t hate MTK but I am beginning to hate you.
I am perfectly fine thanks.
Would really like to engage you in real life and wipe the floor with you.
I perfectly normally told that STMC’s 20 nm bulk process is better than Samsungs one but it’s still not significantly better than 28nm HPM so you are the one that whosent been reading. Still it’s more normal to compare those two 20 nm processes than any FinFET to them. So your objections are rather imaginary ones. By the way I did touch the Kirin power projections & explained them just couple of comments under. Next time do the math and come with facts if you want to say something to me & you will have a big up vote from me.
Now if you insults me one more time I will track you down!
Just to ruin your sleep, MTK just had a talk on the x20 at ISSCC and they still claim 2.5GHz. If you are there, go talk to them, if not, tough shit.
As for your threats , thank you for that ,nothing has made be lol that hard in a long time. And btw, saying those things only reinforces the idea that you have certain anger issues. Could be a matter of education or age but something is off.
It might be a matter I had with your mother long time ago.
I toth shit how you eat one. Now I need to stop with that.
P’s. ZnaΕ‘ mogao bi te ja i potraΕΎiti Ε umadinac!
They did it by scaling down to 32 bit memory controllers.
I take Weibo leaks with a very tiny grain of salt. Let’s wait till MWC and see what really happens. There were Weibo rumors of the SD820 overheating too.
I remember it now,…
“”Good Qualcomm, bad Qualcomm
Good Mediatek, bad Mediatek””
I couldn’t care less as long as the other 2 pairs of 4 cores can take the shift and allow the 2 brute-force ones to cool off, I will just hope for those phones not to melt my hands. I’d even wear gloves if hot hahahaha
This is just the PR war before MWC I would say.
Yea it take very little stock in rumors i hear before MWC. There will all sorts of crap coming out the next few weeks.
Oh god, another not confirmed (neither serious) article about mtk xDDDD
Really, what is happening with bullshitting mtk on this site lately? Please, write articles about facts or at least leaked pictures or whatever… A little bit of (objective) true will be appreciated π
Truth, sorry, keyboard π
Keyboard? Who is still using keyboard dude?? Especially when you comment about mobile technology. haha Just kidding. π
Xiaomi hardly used mediatek chips anyway.
Not now. Now they started using mediatek in their redmi series. But if its overheating maybe they will switch all their phones to snapdragon like they did in 2015 Q1 and Q2
They are already doing it especially after the India ban.
Would be funny, anything is possible on 20nm but remains to be seen. It seems to have been delayed a bit so maybe they did a respin and fixed it but nothing is certain before we have devices to test. With just 2 cores they should reach high enough clocks evenif they miss 2.5Ghz.
News coming from the trustworthy weibo account @totallynotqualcomm π
I’m going to believe it when I’ll see it… I wouldn’t like it that much, but I’ve got to admit that it would be fun to have another year of puns about heating (“helio” in the name is not that reassuring anyway π)
Anyway, let’s talk about strategies:
-xiaomi could scrap x20 because of their problems of selling mediatek based devices to India, and they already did their move with note 3 pro and sd650… Most likely, there will be a mi5c with sd652, but the line up is already crowded enough: there’s hardly other room for the x20.
-lenovo could use the x20 only on the vibe series, but now that they’re planning to use the “moto” brand for the higher end devices, there’s chances we won’t see any good mediatek soc on the vibe brand, sticking it to entry level and lower mid ranger (so the top would be p10, if this scenario is real). So, lenovo could scrap x20 too
-htc is a famous brand, has a partnership with mediatek, it’s trying to make a profit in every way possible cutting costs here and there, and has no problem to sell problematic phones with overheating issues. So, even if those problems are true, htc will definitely sell a variant of his m10 line up with the x20 on board, no matter what.
Here you have it: the rumor could be partially true, but this doesn’t mean x20’s problem are true! π
LOL @ HTC!
I would hit that
I will disbelieve it when an X20 phone gets in your hand.
I hope that Meizu uses Samsung SOC again for the MX 6 rather than the Helio X20.
Look at the past two models and we can see what they will most likely do:
MX4 – MT6595
MX4 Pro – Exnyos
MX5 – Helios X10
Pro 5 – Exnyos
Judging by this the MX6 will have X20, the Pro 6 will have Exnyos.
Pro 6 will have exynos 8870
I also think they will use 8870. A 8870 phone is expected from Lenovo too.
Funny this is that MX4 Pro used the mid range exynos 5430 (of Galaxy Alpha’s fame) instead of the top end Exynos 5433 (of Note 4’s fame), leading to the strangeness of a “pro” model having a weaker CPU than the non-pro (in both multi and single core).
Of course it also had 1GB more Ram ultimately making it faster, but yeah it was a weird decision…
Sometimes you gotta look beyond the SoC and look at the other features. The Pro had a 2K display, more RAM, fingerprint scanner, better camera setup (same rear sensor though), more built in storgae (64 vs 16), better LTE coverage and more.
Yeah, but my post was very specific (I was not talking generally/about other things, just speed): It was (and still is) the first (and probably the last) time that I saw a phone using the Pro moniker to be using a weaker CPU.
My pet theory is that they wanted to use the Exynos 5433 (which is indeed quite faster than MT6595) but the stock was not enough and Samsung only gave them access to Exynos 5430.
BTW it’s not as if Exynos 5430 is bad, it’s quite fast in my Galaxy Alpha, it’s just not “Pro” material…
Depends how you look at it. You are only looking at the raw power (which there really isn’t that big of a difference between the two, in fact in Geekbench they are near even). For me where the 5430 is a much better chip is in the other features of a SoC like GPS, LTE coverage and efficiency.
Yeah, it’s a 20nm chip which makes it very efficient. Like I said, I have it in one of my phones and I love it…. Still it’s weird for a pro line to choose the the midrange chip of a line, instead of the flagship SoC (5433 has far better CPU, GPU, memory bandwidth and probably GPS and LTE).
Also Meizu mostly chooses the high end Samsung Soc, I don’t know what happened with MX4 Pro, probably Samsung told them “No”, thus went with the mid-range one.
It was most likely an availability factor, at the time they were designing the 4 Pro, the 5433 still hadn’t been launched yet, the 5430 was their top chip at the time. The MX4 Pro and the Galaxy Note 4 were both launched around the same time October 2014. The Note 4 was Samsungs newest flagship and they used that to show off the 5433. Samsung and Meizu have been partners since Meizu made it’s first cell phone, if the 5433 was available they would have used it.
Actually all 3 were released October 2014 (Galaxy Alpha was though to be released in September, but actual stock of the Exynos 5430 variant came in October, I remember I was waiting for that phone).
So they had to choose between two SoCs not yet released when making MX4 Pro, they chose the mid-range one because Samsung Told them so (probably low stock of Exynos 5433). I’m *sure* they wanted Exynos 5433 instead.
Still it’s kind of weird. Meizu mostly chooses high end chips.
No one will be using the X20 until its thermals drop below egg cooking temperature.
We’ll see when actual phones launch, we could be saying the same about the 820, there are reports it suffers from overheating too.
http://www.phonearena.com/news/Snapdragon-820-chipset-reportedly-overheats-in-testing-Samsung-to-the-rescue_id75133
This is what happens when you try and do ‘flagship’ on the cheap. Qualcomm couldn’t quite manage it on a 20nm node and it seem MT have hit the same problems. If they wan’t to compete with the big boys they have to use the more expensive 14/16nm nodes that Qualcomm and Samsung are using. There was never any need for 2 quad A53 cores when they could’ve used dual cores instead and this would at least have allowed them to use more GPU cores since this is always their weakness. The fact the Huawaei built the Kirin 950 at 16nm with lower clocked A72 cores tells you all you need to know.
UGH…I should had known, no wonder only few companies will had that chipset
Perhaps a new revision will solve out (X22) ….
But then again i hardly doubt it.
Still…I hope this rumor is not true, so we’ll just wait
it looks like it needs to delay just a bit more…and revamp it just a little
you know uh…16nm, LPDDR4, and perhaps T880MP6…Just Saying, BTW! PHEW!
I take Weibo leaks with a very tiny grain of salt. Let’s wait till MWC and see what really happens. There were Weibo rumors of the SD820 overheating too.
I remember it now,…
“”Good Qualcomm, bad Qualcomm
Good Mediatek, bad Mediatek””
I couldn’t care less as long as the other 2 pairs of 4 cores can take the shift and allow the 2 brute-force ones to cool off, I will just hope for those phones not to melt my hands. I’d even wear gloves if hot hahahaha
This is just the PR war before MWC I would say.
Yea it take very little stock in rumors i hear before MWC. There will all sorts of crap coming out the next few weeks.
Oh god, another not confirmed (neither serious) article about mtk xDDDD
Really, what is happening with bullshitting mtk on this site lately? Please, write articles about facts or at least leaked pictures or whatever… A little bit of (objective) true will be appreciated π
Truth, sorry, keyboard π
Keyboard? Who is still using keyboard dude?? Especially when you comment about mobile technology. haha Just kidding. π
Xiaomi hardly used mediatek chips anyway.
Not now. Now they started using mediatek in their redmi series. But if its overheating maybe they will switch all their phones to snapdragon like they did in 2015 Q1 and Q2
They are already doing it especially after the India ban.
I hope that Meizu uses Samsung SOC again for the MX 6 rather than the Helio X20.
Look at the past two models and we can see what they will most likely do:
MX4 – MT6595
MX4 Pro – Exnyos
MX5 – Helios X10
Pro 5 – Exnyos
Judging by this the MX6 will have X20, the Pro 6 will have Exnyos.
Pro 6 will have exynos 8870
Funny this is that MX4 Pro used the mid range exynos 5430 (of Galaxy Alpha’s fame) instead of the top end Exynos 5433 (of Note 4’s fame), leading to the strangeness of a “pro” model having a weaker CPU than the non-pro (in both multi and single core).
Of course it also had 1GB more Ram ultimately making it faster, but yeah it was a weird decision…
Sometimes you gotta look beyond the SoC and look at the other features. The Pro had a 2K display, more RAM, fingerprint scanner, better camera setup (same rear sensor though), more built in storgae (64 vs 16), better LTE coverage and more.
Yeah, but my post was very specific (I was not talking generally/about other things, just speed): It was (and still is) the first (and probably the last) time that I saw a phone using the Pro moniker to be using a weaker CPU.
My pet theory is that they wanted to use the Exynos 5433 (which is indeed quite faster than MT6595) but the stock was not enough and Samsung only gave them access to Exynos 5430.
BTW it’s not as if Exynos 5430 is bad, it’s quite fast in my Galaxy Alpha, it’s just not “Pro” material…
Depends how you look at it. You are only looking at the raw power (which there really isn’t that big of a difference between the two, in fact in Geekbench they are near even). For me where the 5430 is a much better chip is in the other features of a SoC like GPS, LTE coverage and efficiency.
No one will be using the X20 until its thermals drop below egg cooking temperature.
Yeah, it’s a 20nm chip which makes it very efficient. Like I said, I have it in one of my phones and I love it…. Still it’s weird for a pro line to choose the the midrange chip of a line, instead of the flagship SoC (5433 has far better CPU, GPU, memory bandwidth and probably GPS and LTE).
Also Meizu mostly chooses the high end Samsung Soc, I don’t know what happened with MX4 Pro, probably Samsung told them “No”, thus went with the mid-range one.
I also think they will use 8870. A 8870 phone is expected from Lenovo too.
It was most likely an availability factor, at the time they were designing the 4 Pro, the 5433 still hadn’t been launched yet, the 5430 was their top chip at the time. The MX4 Pro and the Galaxy Note 4 were both launched around the same time October 2014. The Note 4 was Samsungs newest flagship and they used that to show off the 5433. Samsung and Meizu have been partners since Meizu made it’s first cell phone, if the 5433 was available they would have used it.
We’ll see when actual phones launch, we could be saying the same about the 820, there are reports it suffers from overheating too.
http://www.phonearena.com/news/Snapdragon-820-chipset-reportedly-overheats-in-testing-Samsung-to-the-rescue_id75133
Actually all 3 were released October 2014 (Galaxy Alpha was though to be released in September, but actual stock of the Exynos 5430 variant came in October, I remember I was waiting for that phone).
So they had to choose between two SoCs not yet released when making MX4 Pro, they chose the mid-range one because Samsung Told them so (probably low stock of Exynos 5433). I’m *sure* they wanted Exynos 5433 instead.
Still it’s kind of weird. Meizu mostly chooses high end chips.
Anyone notice that the rumor have iPhone 6 plus? How I can believe something coming from one use IOS?
Really? I don’t know what’s more pathetic you or your comment. Saying someone is a liar because he uses an iPhone.
ohh brother, it was just a joke , so it seems you are using IOS ? π
HAHAHA, I WOULD BUY IT ANYWAYS!!!
Anyone notice that the rumor have iPhone 6 plus? How I can believe something coming from one use IOS?
ohh brother, it was just a joke , so it seems you are using IOS ? π
HAHAHA, I WOULD BUY IT ANYWAYS!!!
The fact MediaTek blatantly violate the GPL is enough of a reason not to use them.
A Chinese publication got a comment from MTK, through Google Translate seems just the standard denial , all is ok and so on http://mobile.it168.com/a2016/0203/2048/000002048788.shtml
Ofc that’s as unreliable as the rumor. MTK should react strongly, with actual proof,if all is ok. They had a week full of negative news and that might not be coincidental. If some PR team is going negative, hitting back hard would be preferable.
If the chip is borked, they better not try to mislead and just deal with it.
The fact MediaTek blatantly violate the GPL is enough of a reason not to use them.
A Chinese publication got a comment from MTK, through Google Translate seems just the standard denial , all is ok and so on http://mobile.it168.com/a2016/0203/2048/000002048788.shtml
Ofc that’s as unreliable as the rumor. MTK should react strongly, with actual proof,if all is ok. They had a week full of negative news and that might not be coincidental. If some PR team is going negative, hitting back hard would be preferable.
If the chip is borked, they better not try to mislead and just deal with it.