seunfly:
It perform less in terms of raw processing power but it is ahead interm of image processing and AI. The only andriod phone close to it is vivo x70 pro plus and it is because of the super image processor V1 chips vivo specifically design for that.

atheistandproud:


It's actually kind of weird. Why did they put two X1 cores if they won't use them

ORIJIN201:
it will take less time doing tasks thereby saving power.

atheistandproud:


Yes I agree.

So what's holding it back?

Abi na the 2 A76 when them use?

Because the others use A78 which is far more powerful than the A76.

atheistandproud:


It's actually kind of weird. Why did they put two X1 cores if they won't use them

seunfly:


They wanted the phone to be powerfull when u are doing single but complex tast such as machine learning, image processing, artificial intelligence processing, speech recognition, google assistance etc but google doesn't want to over burden the phone with too much power hungry chips since X1 is very, very power hungry so they step down to Arm cortex A76 to reduce the power load on the battery, this decision saves tensor from being a battery killer chips and it also makes it less powerful because SD 888 and exynos 2100 used Cortex A78 while Huawei Kirin 1000 makes use of Cortex A77, this place tensor behind those chip in raw power.

What google simply does was to focus on their intention for the phone, they are not here for gaming or multitasking, they are here for Artifical intelligence, using ur phone to control ur house, cars, electronics, pertern recognition etc, it is extremely powerful in this particular aspect, it will be very fast in such tast while it will only be capable in other aspect.

Google is going in another direction entirely from modern phones, they wants all our day to day activities to be socs based instead of normal cloud base. This simply means that all the google earth, google map, speech recognition, android assistance capability will be handled by the chips instead of usual google server.
We are seeing the end of cloud based processing and more of phone based service processing, as soon as they get things right, u won't need to be connected to the internet or Google server to be able to use services like Google map and co.

emae009:


Google map uses every android device available in that vicinity to you give Real Time map updates, accident reports, and time saving routes. It can't do that while being offline. It has to connect to Google server except it's pre-determined route while home to office which atm can be saved offline in normal devices

seunfly:


I agree with u but Google knows best, that is their plan and that is the direction they wants to move.

We don't know what they know.

I remember some years when they started the cloud thing, many analysts think they are not realistic, they want to go cloudless people are saying it is not possible.

Samething they told Apple it is not possible to operate the internet of today on phone but Apple figured it out. Never say never because this guys are constantly researching and pushing boundaries.
Google made the decision because of privacy issue.

Today's machine learning and AI algorithm relied too much on server stored somewhere we don't know and too vulnerable to government, hackers, marketers etc before any app can improve and serves u better they had to take most ur of data down to personal voice conversation for the algorithm improvement, this is not safe. Most of this organization are getting away with it because they are American companies, if it was non American or western companies they would be dead by now based on the amount of invasion of privacy and personal data they take.

A woman scrutined Amazon and realized all her life is on Amazon server, from shopping to phone conversation and the frightening thing is that non phone conversation is also stored on the server.

The purpose of this is to analyse individual and personalize item for them based on machine learning but who know who will use such data against her.

emae009:


For maps it's likely impossible except in cases of pre-determined route where ML can learn and store your route in the phone. Real time updates will need Google server as it must rely on information outside your phone.

Taking Apple's Find my Phone as a possibility for maps which can work offline by using UWB to constantly scan frequencies. It could work for walking, hiking etc. but driving? the speed at which a car moves make that sort of tech redundant

seunfly:


Honestly I'm waiting to see how it will be implemented, we are server dependent system.
I want to see how google earth will work, what about YouTube, I learnt Gmail is also part of the plan. One thing I v learnt is what we can do with technology and law of physics can't be predictable, u can achieve anything with it if u master it enogh.

Another way I'm looking at it is may be, the apps will collect the data, process it on your phone and send the final result to the server instead of massive data taken to the server for processing which then send to end users.

Bytedance is also looking in that direction, since they were accused of data theft to d extent of copying what u are typing on ur phone in 2019, they are slowly implementing the normal processing on phone and final stuff is being used for their algorithm on the server.

Those data theft is what made bytedance creates one of the most powerful algorithm that can personalize ur life to perfection.

seunfly:


They wanted the phone to be powerfull when u are doing single but complex tast such as machine learning, image processing, artificial intelligence processing, speech recognition, google assistance etc but google doesn't want to over burden the phone with too much power hungry chips since X1 is very, very power hungry so they step down to Arm cortex A76 to reduce the power load on the battery, this decision saves tensor from being a battery killer chips and it also makes it less powerful because SD 888 and exynos 2100 used Cortex A78 while Huawei Kirin 1000 makes use of Cortex A77, this place tensor behind those chip in raw power.

What google simply does was to focus on their intention for the phone, they are not here for gaming or multitasking, they are here for Artifical intelligence, using ur phone to control ur house, cars, electronics, pertern recognition etc, it is extremely powerful in this particular aspect, it will be very fast in such tast while it will only be capable in other aspect.

Google is going in another direction entirely from modern phones, they wants all our day to day activities to be socs based instead of normal cloud base. This simply means that all the google earth, google map, speech recognition, android assistance capability will be handled by the chips instead of usual google server.
We are seeing the end of cloud based processing and more of phone based service processing, as soon as they get things right, u won't need to be connected to the internet or Google server to be able to use services like Google map and co.

seunfly:


Honestly I'm waiting to see how it will be implemented, we are server dependent system.
I want to see how google earth will work, what about YouTube, I learnt Gmail is also part of the plan. One thing I v learnt is what we can do with technology and law of physics can't be predictable, u can achieve anything with it if u master it enogh.

Another way I'm looking at it is may be, the apps will collect the data, process it on your phone and send the final result to the server instead of massive data taken to the server for processing which then send to end users.

Bytedance is also looking in that direction, since they were accused of data theft to d extent of copying what u are typing on ur phone in 2019, they are slowly implementing the normal processing on phone and final stuff is being used for their algorithm on the server.

Those data theft is what made bytedance creates one of the most powerful algorithm that can personalize ur life to perfection.

emae009:


I'm just pointing out that there isn't much we can do on our phone no matter how sophisticated they are without the Internet on a wider scale. Once the Internet is involved, data must be collected and store in a server that shares these information.
All they can do going forward is to follow Apple's path where any data that's leaving your phone to server uses a randomized unique ID to get the info you need at the time. Without that there's nothing much ML or AI can do without the server.

To me though, we are making a big deal out of this privacy issue. Yes the concern is great but not everything is black or white. I mean how many people right now would willingly use the first Android version in 2021 or the first version of Facebook, Instagram, YouTube, and the rest? I willing to bet not a lot but these softwares couldn't have improve over time without data collection.

atheistandproud:


I like this comment but I disagree with the privacy issue. A bigger deal should be made out of it.

Where were you on this day in 2003? It will take you a while to recollect. Imagine that company has all of that. Everywhere you have been, your seach history for the last 10 years, all your photos you put online, a full psychological profile of who you are etc. and is willing to sell to the highest bidder.

Guy it's scary as f.

seunfly:


They wanted the phone to be powerfull when u are doing single but complex tast such as machine learning, image processing, artificial intelligence processing, speech recognition, google assistance etc but google doesn't want to over burden the phone with too much power hungry chips since X1 is very, very power hungry so they step down to Arm cortex A76 to reduce the power load on the battery, this decision saves tensor from being a battery killer chips and it also makes it less powerful because SD 888 and exynos 2100 used Cortex A78 while Huawei Kirin 1000 makes use of Cortex A77, this place tensor behind those chip in raw power.

What google simply does was to focus on their intention for the phone, they are not here for gaming or multitasking, they are here for Artifical intelligence, using ur phone to control ur house, cars, electronics, pertern recognition etc, it is extremely powerful in this particular aspect, it will be very fast in such tast while it will only be capable in other aspect.

Google is going in another direction entirely from modern phones, they wants all our day to day activities to be socs based instead of normal cloud base. This simply means that all the google earth, google map, speech recognition, android assistance capability will be handled by the chips instead of usual google server.
We are seeing the end of cloud based processing and more of phone based service processing, as soon as they get things right, u won't need to be connected to the internet or Google server to be able to use services like Google map and co.

seunfly:


I agree with u but Google knows best, that is their plan and that is the direction they wants to move.

We don't know what they know.

I remember some years when they started the cloud thing, many analysts think they are not realistic, they want to go cloudless people are saying it is not possible.

Samething they told Apple it is not possible to operate the internet of today on phone but Apple figured it out. Never say never because this guys are constantly researching and pushing boundaries.
Google made the decision because of privacy issue.

Today's machine learning and AI algorithm relied too much on server stored somewhere we don't know and too vulnerable to government, hackers, marketers etc before any app can improve and serves u better they had to take most ur of data down to personal voice conversation for the algorithm improvement, this is not safe. Most of this organization are getting away with it because they are American companies, if it was non American or western companies they would be dead by now based on the amount of invasion of privacy and personal data they take.

A woman scrutined Amazon and realized all her life is on Amazon server, from shopping to phone conversation and the frightening thing is that non phone conversation is also stored on the server.

The purpose of this is to analyse individual and personalize item for them based on machine learning but who know who will use such data against her.

emae009:


You put the photos online right?
Search history? I know how much time I've from not clicking on a site I've already searched previously that didn't give me the answer.
Like I said, it's a gray area not black or white. People get killed from cyber stalking not that they bought your data from Google or others.
You can make a case of using our data to condition how we behave and what we believe which is true to some extent.

I know people get shifty when privacy is discuss but have you ever try to search for something on DuckDuckGo? because they don't have a lot about you, your search will have many irrelevant result especially if it's not something generic.
yes some personalized data shouldn't be collected but you can't argue that those data enhances our user and online experience

Studymore123:


A cybersecurity professional once told me. "The less private data we have online the less search rewarding experience we have "
I don't have issues with privacy, I only have issues when personal data is exploited to violate the trust of users.

atheistandproud:


I like this comment but I disagree with the privacy issue. A bigger deal should be made out of it.

Where were you on this day in 2003? It will take you a while to recollect. Imagine that company has all of that. Everywhere you have been, your seach history for the last 10 years, all your photos you put online, a full psychological profile of who you are etc. and is willing to sell to the highest bidder.

Guy it's scary as f.

emae009:


You put the photos online right?
Search history? I know how much time I've from not clicking on a site I've already searched previously that didn't give me the answer.
Like I said, it's a gray area not black or white. People get killed from cyber stalking not that they bought your data from Google or others.
You can make a case of using our data to condition how we behave and what we believe which is true to some extent.

I know people get shifty when privacy is discuss but have you ever try to search for something on DuckDuckGo? because they don't have a lot about you, your search will have many irrelevant result especially if it's not something generic.
yes some personalized data shouldn't be collected but you can't argue that those data enhances our user and online experience

Meksbig:


I personally don't think it's about collection of data that is the problem but transparency in collection and use of data. What I mean is the user should be in charge of how and how much of their data is collected and used.

I think more people would be open if they were sure of what their data is used for. Waking up to find that your data was sold to a company that wants to rig an election kinda leaves a sour taste in the mouth.

atheistandproud:


That's what the Ultra low power engine is for, I think. Coupled with the fact that it is a 5nm SoC.

I obviously was thinking that they could clock the two X1 cores near peak power. And only use them when necessary. They're not going to be used all the time.

That's what the A76 and A55 are for.

But like you said, Google knows best and after considering a lot of things...

Many of the AI tasks may pop up regularly causing the X1 cores to come online and use battery throughout the day. Clocking the X1 lower to maximize battery consumption is possibly the most ideal option.

So it's starting to make sense.

We can never know better than the engineers that made it.

Studymore123:


Nice.
But the reality is we are becoming more cloud dominant than less cloud based. True, capabilities can be built chips and todays chips are more powerful than they are truly utilized. I don't need to be online to use maps, I have it all downloaded on my cellphone but I need to be online to know the present traffic along my route and other real time updates. The same is applicable for most other apps including the voice typing and translate features. Real time enhances the capabilities.

Google is designing these chips to leverage their ML/AI capabilities which they have developed over the years. These chips take advantage of the Google ecosystem to deliver amazing performance (they still have limitations eg conversational version of voice translate requires internet connection to use the cloud based TPUs) and when combined with the cloud capabilities of Google servers they have no rival.

I don't really understand what you mean by going cloudless but Google is increasing their cloud physical infrastructures by the day and spending billions on Cloud research, data centers, smart technologies and iots.