Just looked at a teardown report of the Epic 4G (hat tip to los1223). It appears that the Epic 4G does not even use the same GPS chip as its GSM cousins. This is certainly news to me, because reading the tea leaves I have assumed that all Galaxy S phones shared the same GPS hardware, even if the software was in different stages of development and debugging.
According to the teardown, the Epic 4g uses the multifunction Qualcomm QSC6085 for GPS functions. (Presumably this is because of the CDMA radio, so the Fascinate design is probably similar.) Previous analysis the GSM variants, based on the international I9000 model as well as the Captivate and the Vibrant, showed they use a Broadcom BCM4751 chip.
This would explain why the symptoms of the GPS problems on the Epic, and the firmware modules, are not the same as those on the GSM variants. But amazingly, it seems that Samsung managed to screw up two different GPS designs in two different ways.
Is one chip better than the other?
I know my co-worker's TMobile Galaxy S (Captivate, Fascinate, Masticate...whatever) just recently received the GPS update and he states it is now spot on and very quick to lock.
Didn't ask about any hard coded location, though, like ours. I'll have to do that on Monday.
Thats ame gps as the motorola droids use I think
The new QSC6085 features include:
Support for EV-DO Rev. A with backward compatibility to EV-DO Rel. 0 and CDMA2000
Equalizer technology for increased EV-DO data throughput
Powerful GPS location capabilities, including gpsOne(R) multi-mode hybrid assisted GPS technology, enabling a broad variety of location services such as navigation, social, safety and security applications
Up to 3 megapixel camera support with anti-shake capability
15 fps video encode and decode capability
72-voice polyphonic ringtones with support for multiple audio codecs
Multiple technologies to improve network capacity, including QUALCOMM Linear Interference Cancellation (QLIC(TM)) with pilot and traffic cancellation, and Fourth Generation Vocoder(TM) (4GV(TM)), published as the EVRC-B 3GPP2 standard codec, which gives network operators the flexibility to prioritize voice quality or network capacity
Support for multiple bands, including 800 Mhz, 1900 Mhz and the AWS band (1700/2100 Mhz), with mobile receive diversity
Cost-effective 65 nm process technology and a 12 mm x 12 mm package size
othan1 said:
Is one chip better than the other?
Click to expand...
Click to collapse
Let me first lead off by same I'm not an expert on chipsets. That being said, I did a lot of research on the QSC6085 when I was looking to see if it supported TV out or if we were never going to have that feature (I can honestly tell you that I don't know enough about this stuff for a conclusive answer, but right now I'm guessing no for the 3.5mm port).
Anyhow, the QSC6085 is an integrated processor that does a lot of this with one chip rather than having multiple chips that do many individual things. The purpose was to reduce costs and potentially, create power savings as well.
The other galaxy S phones (at the least the GSM ones) have a dedicated GPS chip to where the Epic's/Fascinate's baseband modem (QSC6085) handles all kinds of things, including GPS.
Other devices that use the QSC6085 (like the Droid, Droid 2 and my Sierra Wireless modem) do not seem to have the same GPS problems as the Epic/Fascinate which tells me it's a Samsung implementation issue and not a problem with the chip.
It's either the software (highly likely, no reason for us to not get a fix within seconds even if it is horribly inaccurate) or an insensitive GPS antenna. The GPS antenna is easily spottable on the Vibrant and Fascinate. It's covered by black tape next to the SD card on the Fascinate, it should be the same on the Epic 4G.
heygrl said:
It's either the software (highly likely, no reason for us to not get a fix within seconds even if it is horribly inaccurate) or an insensitive GPS antenna.
Click to expand...
Click to collapse
the snr numbers and the cold start fix tell us it is both, mainly software though.
Related
curious if anyone has flashed any of the asian ROMs and found the FM radio chip in our VIbrant?
IIRC its not a software issue, the vibrant doesn't have the hardware. Though it wouldn't be the first time an FM radio was hidden in a phone.
Sent from my SGH-T959 using XDA App
was hoping it had been found to still be there
I thought the same chip by Broadcom was used in all Galaxy S variants that integrates BT, GPS and FM radio
http://www.intomobile.com/2010/02/1...-with-media-playback-and-recording-functions/
I really have no idea about the hardware, but maybe that's why there is problems with the GPS, they didn't use the chip that has the BT/GPS/FM in it and used something else that took out the FM part and screwed up the GPS. Just a wild off the wall speculation.
there was an initial post by TravUK identifying the chipset as the Broadcom BCM 20751 and that is a 3 way chip (BT/GPS/FM)
but later he corrected that as a mistake - others found in the UK version of the i9000 and the US Tmobile delivered Vibrant (T959) and i believe in the AT&T Captivate, that the chip was a Broadcom 4751, which is a singular function chip, GPS only
so the FM chip must reside by itself or with the BT chip
I guess without doing the full autopsy we won't know what chipset they used unless someone who knows inner land of Samsung can leak it - bummer, I wish there was dmesg or similar output that you can parse to get the idea of the hardware used.
As said the BCM 4751 is the gps chip, not the aforementioned 20751. However, the phone still needs a wifi/bluetooth chip, and from one of the captivate threads I believed it was identified as the BCM4325, which also has fm capability. I think it is the only broadcom chip with bluetooth 3.0 and wifi b/g/n. So fm has to be on the chip. Whether or not everything is still connected on the pcb is still a question, but there doesn't seem to be any reason for samsung to have changed anything if they're still using the same chips.
Even if the chip supported FM, there is a lot more to it in terms of powering those circuits, integrating them into the audio circuitry, and providing leads for an antenna. Even if you could "activate" the chip you may not be able to pick up anything or hear anything.
All current Note users, anyone here can tell me if any of the currently released Note's have NFC? Also read a spec sheet somewhere stating it to be optional, any chance we can get this as an add-on accessory?
I'm also keen to find this out. The UK devices have NFC listed as 'optional' which doesn't really reveal anything useful. Can anyone who has one identify whether the capability is there but not enabled or just not present? It would be nice to think that the devs here could enable it if it was available.
NFC: Optional
It will be supported by a connection through the micoUSB *sigh*
However, the Korean models will get internal NFC and LTE (4G) support. I was going to hold out for that model but aterwards I learned they will not have the Exynos but instead a Snapdragon (Qualcomm S3) chip instead.
I prefer an ARM chip (TI OMAP, Samsung in-house, ST-Ericsson, Nvidia Tegra) rather than Qualcomm = usually more devices, better performance, less battery drain, fewer driver issues, possible to recylce drivers from one device to another.
http://sammyhub.com/2012/03/11/samsung-announces-galaxy-tab-7-7-lte-for-south-korea/
Samsung’s first Galaxy Tab with AMOLED display has been announced for its home market. The Galaxy Tab 7.7 LTE sports a 7.7-inch Super AMOLED Plus Display (1280×800 WXGA) and will be available in South Korea via SK Telecom.
The Galaxy Tab 7.7 LTE is touted to be the first LTE tablet to support voice calls, enabling users with improved voice call quality in addition to faster surfing speeds. Moreover, the 7.9mm thin tablet weighs just 345g and comes with Android 3.2 onboard, a dual-core 1.5GHz CPU, Wi-Fi 802.11b/g/n, Bluetooth 3.0, 3MP and 2MP rear and front-facing cameras respectively, T-DMB, 16 or 32GB internal memory, microSD card slot, 5100mAh battery and various Hubs like Learning Hub, Social Hub, Readers Hub etc.
The tablet will be available in silver color for a price of 800,000 won (US$716).
I'm waiting to be able to buy this locally in Canada...
Really wouldn't be any different than the 7.7 that's out right now except for the added 800mhz band. South Koreans use 800/1800 bands for LTE while Canada uses 700/2100 like the US, neither uses 800mhz at all.
Big difference is that Verizon version has voice removed
The only notable thing there is that it's a 1.5GHz CPU. I wonder if that's a mistake or they are running a different SoC--I've never seen an Exynos stock clocked at 1500MHz, not that it's impossible but it seems unlikely, and that is a common clock rate for Snapdragon S3s (which appear in nearly every Samsung LTE device aside from the I815).
Does it have the infrared port that is missing in the international version?
charliechocolate said:
Big difference is that Verizon version has voice removed
Click to expand...
Click to collapse
So you don't think it's a big difference that they use different LTE bands so that the Korean version won't work on any North American LTE data networks? Good luck with that.
Guys who hack for a hobby, or who work on Android development can probably pass on this thread, as they will, or should already know the following.
For everyone else:
Not all Galaxy Note smart-phones are equal. The same applies to every other smartphone, computer, computer chip, computer component, or aeroplane. http://www.youtube.com/watch?v=-kHa3WNerjU This aeroplane was hard-bricked because the programmers didn't understand how we fly aircraft, and the pilot not understanding the programming logic, probably because he was never told. BUT, there might be another reason. Quality control of the computer chip. In the days I'm talking about, every chip was manufactured as a 486 or 386 chip, under-testing, if it failed, it would then be tested as an SX chip (the math co-processor disabled) if it passed the lesser test, it would pass as a SX chip, instead of the premium DX chip. So if anyone bought a computer with a SX chip, they bought a failed chip. Just like the guys who built that aeroplane, it had 6 or 7 SX (failed) chips driving the whole thing. Of course, the chip and plane makers insist there was nothing wrong with the product, it must be the pilots fault. In a way they are correct, the components met specification, but when the pilot pulls up, and commands full power.....and it still hard-bricks itself....hmmm...someones lying.
Quality control over chips is pretty much the same today as it was then. To expect 100% perfect chips to be installed into 100% smartphones would result in most chips winding up in landfill and your phone costing £$kk. So we have to compromise, and find acceptable limits. When your phone was made and software loaded for the first time and tested, it worked, but doesn't mean it will work the next time upgrade software is loaded. That is one reason for hard bricking a phone or aeroplane. So when you flash up the latest software, a warranty would be nice.
The next reasons for hard-bricking your phone is many fold, and many will be familiar with this. I'm now talking about peripheral devices, magnetometers, giros, accelerometers and on... The same still applies to these devices as the chips when it comes to quality control. So that's another point. But are ALL devices in your phone exactly the same, from the same manufacturer, in EVERYONES phone? Personally I doubt it. Buy any make and model of computer you want, and I'll guarantee that I can find differences in components for that brand and model. So what? Well this is what, those different devices may well have differing driver requirements. The same applies to your smart phone.
Case Study:
I speced up my own faster than me PC, had the component list checked by an expert. We all felt happy and I bought it, it was also assembled by experts. It powered up, passed all the tests, memory, bench marking, you name it, we did it. All passed. Happy? NO! In use I got random BSOD. fault traced to an instruction sent from the chip to the mobo which the driver did not recognise. Found offending setting in the BIOS and stopped the instruction being sent. Now have a very stable and fast PC. Both Intel and the mobo makers said their product was perfect. They are correct, but with certain combinations of hardware, this glitch comes up. Guess what, its the same for your smart phone.
There are many reasons why Samsung and all the other makers 'regionalise' the product. Language comes top of the list to most people, but it goes further. Because the components wont all be the same, from the same factory, or the same maker, phones will be made in batches. All the phones with giro 'A' go to the USA, all phones with giro 'B' go to India and so on. Assume Giro 'A' needs a different driver to giro 'B'. When we upgrade to a new android version, the USA will get a different upgrade to India, and one will be released before the other. Translate this to all of the other peripheral devices, and you can then begin to understand why there are so many different ROMs out there, and why Samsung roll out in slow time. They have to make sure, as best they can, that the upgrade has the correct coding for your phone, because your phone is different.
There are some misconceptions about what is an official ROM. To me, an official ROM is one that Samsung selected for me, via KIES. A ROM destined for users in say, Australia, is not an official ROM for me with an EU phone, otherwise, why would Samsung make life hard for themselves by having extra ROM releases? Just pause and think, that's all.
One more point. It would be helpful for those that can, post a hardware list when they spot faults, bugs and hard bricks. No one seems to do that, but on PC forums, that is the norm. Maybe we could get much more accurate with our hack and so on if we could do this. here is what's under my bonnet:
Sensor - Accelerometer: K3DH Acceleration Sensor, Vendor: STMicroelectronics, Version: 1, Power: 0.2, Resolution: 0.0048, Max Range: 19.6.
Sensor - Gyroscope: K3G Gyroscope Sensor, Vendor: STMicroelectronics, Version: 1, Power: 6.1, Resolution: 0.0003, Max Range: 8.7.
Sensor - Pressure: BMP182 Pressure Sensor, Vendor: Bosch, Version: 1, Power: 0.1, Resolution: 0.0100, Max Range: 1100.0.
Sensor - Magnet: AK8975 Magnetic field Sensor, Vendor: Asahi Kasei Microdevices, Version: 1, Power: 6.0, Resolution: 0.0600, Max Range: 2000.0.
Sensor - Orientation: AK8975 Orientation Sensor, Vendor: Asahi Kasei Microdevices, Version: 1, Power: 7.8, Resolution: 0.0156, Max Range: 360.0.
Sensor - Light: GP2A Light Sensor, Vendor: Sharp, Version: 1, Power: 0.8, Resolution: 1.0, Max Range: 3000.0.
Sensor - Proximity: GP2A Proximity Sensor, Vendor: Sharp, Version: 1, Power: 0.8, Resolution: 5.0, Max Range: 5.0.
Sensor - Gravity: Gravity Sensor, Vendor: Google Inc., Version: 1, Power: 0.2, Resolution: 0.0048, Max Range: 19.6.
Sensor - Linear Acceleration: Linear Acceleration Sensor, Vendor: Google Inc., Version: 1, Power: 0.2, Resolution: 0.0048, Max Range: 19.6.
Sensor - Rotation: Rotation Vector Sensor, Vendor: Google Inc., Version: 1, Power: 6.2, Resolution: 0.0000, Max Range: 1.0.
Awesome post man! I too always wondered why there were so many different "official ROMs". Very nice post.
Sent from my GT-N7000 using XDA
So, based on your reasoning, we should assume that the German LPY firmware should be more safe for German phones.
Could we test such assumption? Did superbricked devices come most likely from non-German devices?
Sent from my GT-N7000 using xda premium
debernardis said:
So, based on your reasoning, we should assume that the German LPY firmware should be more safe for German phones.
Could we test such assumption? Did superbricked devices come most likely from non-German devices?
Sent from my GT-N7000 using xda premium
Click to expand...
Click to collapse
That is how the reasoning works. Thats why its important to post hardware detail and the Samsung pre-installed version ROM. We have to know the start point, know the success, and the failures, to get some reasoning into why some phones brick, and others don't.
long read, but after the first paragraph.. i think u are FOS! if sx chips are failed dx chips then that would mean the manufacturer would have go check every chip to make to find out if it's sx or dx.. FOS! it's more likely there was a separate manufacturing process for the sx and dx chip.. i couldn't be bothered reading the rest
Interesting argument, but I think a simpler reason for there being multiple ROMs is purely due to ease of distribution OTA by the carriers on specific national networks rather than due to hardware differences.
It's certainly true that not all smartphones of the same name all have identical components inside (the US version of the S3 is a case in point), but this current issue with ICS seems to be due to a revision of firmware causing damage to flash chips. I would guess that all flash chips on all variants of the Note are susceptible until proven otherwise - e.g. no issues with those who have "German" Notes flashing LPY.
Coincidentally, I have a "German" Note, but I have no interest in being a guinea pig. I'll wait for the software fix .
Replying to the OP... While you are somewhat correct about CPUs 20 years ago it's not quite the same today. Generally CPUs are speed tested and you are more likely to grade a CPU on reliable clock frequency rather than retarding it's functionality in any way. A lot more effort is put into manufacturing processes for reliability than 20 years ago and the small % of devices that fail nowadays are more likely to be thrown away or recycled.
Your case study is flawed and somewhat insulting to engineers. To suggest you stringing a few components together and calling it a well designed PC because an 'expert' had given it a thumbs up and then installing a generic OS on it is the same as a mobile phone that has been bespoke designed by a team of engineers, with firmware especially developed to work with that hardware is way off. And, yes, you are getting a bespoke build of Android on the Note.
I'm not suggesting that our phones will be bug free, but all the hardware and firmware will have been exhaustively tested on multiple examples of the device under different loads.
Apple v Microsoft is a classic example here. MS designed the OS and left component manufacturers to do the drivers. Result, nobody really knows how well the components will work together and the solution is only as good as the combination of drivers, etc. Apple decided to strictly control what hardware is in their machines and could optimise the OS to work reliably with that hardware.
bamboo12 said:
long read, but after the first paragraph.. i think u are FOS! if sx chips are failed dx chips then that would mean the manufacturer would have go check every chip to make to find out if it's sx or dx.. FOS! it's more likely there was a separate manufacturing process for the sx and dx chip.. i couldn't be bothered reading the rest
Click to expand...
Click to collapse
Well, FOS, that's just your opinion and that's your right. BUT that's EXACTLY how it used to happen. I'm not going to start writing up history lessons here, I'll leave the research to you. Maybe you'll learn something along the way, including some manners. I'm always up for reasoned discussion, but personal insult just hits a brick wall with me.
emuX said:
............
Your case study is flawed and somewhat insulting to engineers. ......
Click to expand...
Click to collapse
Not sure how I insult engineers, that is not my intent. So if there are any engineers out there who feel insulted, I apologise.
bamboo12 said:
long read, but after the first paragraph.. i think u are FOS! if sx chips are failed dx chips then that would mean the manufacturer would have go check every chip to make to find out if it's sx or dx.. FOS! it's more likely there was a separate manufacturing process for the sx and dx chip.. i couldn't be bothered reading the rest
Click to expand...
Click to collapse
As far as the 386/486 DX/SX goes, it is true. I guess even today's lower cache processors are defect chips where they turn off specific part of the memory and brand as medium/entry level chips.
So what's the thread title have to do with this? It should read: (THEORY) Flashing ICS Region Other than Original Manufacturer POSSIBLY Dangerous
Or
(THEORY) Samsung Bins Parts ....like everyother manufacturer out there
I'm not knocking what you're saying. Look at Intel, most cpus cut from the same wafer but just binned according to their resistance and that's how you get a $200 processor vs a black box $1000. They will even turn off cores
Well,
the other thread is having discussion mainly about the Snapdragon 800 version, but since we Europeans can't have it, it is better to have the discussion about the Exynos models here.
Do you think Heterogeneous Multi-Processing Capability will be implemented in the Note 10.1 2014?
http://youtu.be/fLrSTJECVaU
http://youtu.be/8LNPxExkLMo
http://youtu.be/1t-6jqhELVk
The wifi only model will be exynos based. I live in the US and only want the exynos version.
Why do you think that we (Europeans) won't get the Snapdragon 800 version?
In Europe
WiFi = Exynos
LTE = Snapdragon 800
Live4Racing said:
In Europe
WiFi = Exynos
LTE = Snapdragon 800
Click to expand...
Click to collapse
Isn't that true everywhere?
Speaking of the Exynos version. How does the community support for those devices stack up against their snapdragon
siblings (thinking of s4 and so on)?
Aletheia said:
Isn't that true everywhere?
Speaking of the Exynos version. How does the community support for those devices stack up against their snapdragon
siblings (thinking of s4 and so on)?
Click to expand...
Click to collapse
from what i have read and heard, the snapdragon siblings seem to have much more dev support than the exynos ones - that being said, with the new note 10.1, we might be able to expect some pretty damn good dev support so long as the rumors/claims about HMP being fixed are true. i think the biggest downfall of the previous octa chips is that they either use all 4 A15 chips or all 4 A7 chips, not a combination of them or anything else - the cache translation function is broken (i think thats what it is called... i may be wrong on the name).
but i think if they have fixed it, and if the new note actually offers HMP, then there should be some good dev support i would suspect - not everyone wants to pay for an extra contract for their device when they can just tether or hotspot to their phone, which they always have on them anyways.
On that *note* (pun intended) does anyone know if, on this new device, the 3g and lte models will have phone and sms functionality via headset of some sort? i can't imagine why would would purposely disable this feature, as the device has the proper radios etc for it. and if it is somehow disabled, does anyone know if that is something that can be brought over to the device with a new rom/kernel or other mod?
just my 3 cents, yea that's right, 3 cents....
asaqwert said:
On that *note* (pun intended) does anyone know if, on this new device, the 3g and lte models will have phone and sms functionality via headset of some sort? i can't imagine why would would purposely disable this feature, as the device has the proper radios etc for it. and if it is somehow disabled, does anyone know if that is something that can be brought over to the device with a new rom/kernel or other mod?
just my 3 cents, yea that's right, 3 cents....
Click to expand...
Click to collapse
Make a "note" to look into a program called Tablet Talk.
You can make any tablet into a sms/mms and phone... so worth the price.
Itchiee said:
Make a "note" to look into a program called Tablet Talk.
You can make any tablet into a sms/mms and phone... so worth the price.
Click to expand...
Click to collapse
Have definitely used and like tablet SMS on my new nexus but I just find it a hassle as the tablet always needs to be tethered anyways so I need my phone with me regardless. The idea behind having full talk and text on the tablet is so I don't absolutely need both devices with me all the time... I guess with a tablet that has 3g or late then you don't really need the phone with you anymore, and I can see an upside of this setup.... You can use just one phone number and have it accessible on two devices...just kind of seems ****ty to basically be forced into having two separate phone bills each month though if the option could have very easily been there to have the talk and text directly active and working on the tablet. I wonder if there is some sort of mod or hack that could be used to gain this functionality considering that on a 3g or late device the rradios are all there to support this.....
asaqwert said:
On that *note* (pun intended) does anyone know if, on this new device, the 3g and lte models will have phone and sms functionality via headset of some sort? i can't imagine why would would purposely disable this feature, as the device has the proper radios etc for it. and if it is somehow disabled, does anyone know if that is something that can be brought over to the device with a new rom/kernel or other mod?
just my 3 cents, yea that's right, 3 cents....
Click to expand...
Click to collapse
In all review that i saw, all tablet can phone and send sms/mms.
There are the apps for this.
So, i will buy the 4G version because i want the Snap800, and because it will have a lot of support then Exynos base...
Guich said:
In all review that i saw, all tablet can phone and send sms/mms.
There are the apps for this.
So, i will buy the 4G version because i want the Snap800, and because it will have a lot of support then Exynos base...
Click to expand...
Click to collapse
what is exepcted price difference between Wifi only and 4g
what is the support that makes u willing to spend that difference ! is it from Samsung it self, or free developers !
will it be in just Applications ! or the Android system updates running !
this is going to be my first device,and i am already highly considering the 3G version as i dont have smartphone, and hope i can use the note (beside its main functions) as a smart phone + to connect to net when no wifi around or on the road , with headset .. but i will wait till after release to see reviews about that .
if the 3g not making call as a phone, i will get wifi, and for internet connection i will use usb flash modem (by OTG adaptor)
Dr_Muhsin said:
what is exepcted price difference between Wifi only and 4g
i dont have smartphone, and hope i can use the note (beside its main functions) as a smart phone + to connect to net when no wifi around or on the road , with headset .. but i will wait till after release to see reviews about that .
Click to expand...
Click to collapse
The price difference between the wifi only and 4g models can be $100-$200 for 16GB models, and sometimes you can't even find the higher capacity models in 4g or wifi. It can be a hassle trying to find the model you want. Data plans for cell based tablets can be ridiculously high for 4g models so have a carrier in mind if you want to go the 4g route. Keep in mind Im using 4g here as a generic term for cell based tablet (4g/3g/lte). Im sticking with getting a wifi based note 10.1 this time around because I hope to use it for content creation and media playback.
Kernel source
So, is this good news as far as dev support for the Exynos versions is concerned?
sammobile.com/2013/09/20/samsung-releases-kernel-source-for-galaxy-note-10-1-2014-edition
In john lewis web its only left the white wifi version. I hope that its because they are receiving next week the new one
'Note' that the tablet with not ship with HMP/GTS out-of-box.
They were recently finalized. They will ship with cluster migration/core migration logic initially. Later there will be some kernel/patch upgrade to have HMP enabled.
CLARiiON said:
'Note' that the tablet with not ship with HMP/GTS out-of-box.
They were recently finalized. They will ship with cluster migration/core migration logic initially. Later there will be some kernel/patch upgrade to have HMP enabled.
Click to expand...
Click to collapse
i was reading that in the octa-core you could only have 4 cores being used at a time; and that the lower ghz quad cores were to save power.
but if android's cpu governor will fluctuate the cpu speed up and down based on the workload, what's the diff?
doesn't that make the other 4 cores useless?
so in effect a 2.3Ghz quad-core vs a 2.4Ghz quad+1.7Ghz quad would have the same performance??
now if you can use more than 4 cores at a time that's a different story.
am i way off base here?
-Tony
ncohafmuta said:
am i way off base here?
Click to expand...
Click to collapse
A bit. The architecture of each 4-core chip is very different; designed for different tasks. The weaker core is A7 while the more powerful core is A15. Your comparison is kind of like cylinder deactivation in a car's engine. In a V8, all eight cylinders are identical and they are shut off in pairs to save fuel and match performance to load. Think of Octa as a discreet 4 cylinder engine (A7) alongside a V8 engine (A15). Each, to your point, can be controlled so any or all of their cores can be active at a given time. The 4 cylinder's barely able to reach the low point of the V8's performance curve. Similarly, the V8 can't reach the 4 cylinders lower efficiency range. 85% of apps don't leave the A7 core so that's why it's more efficient than just shutting off some of the A15's cores; especially true at idle. It's actually a pretty neat approach to performance vs. efficiency.
BarryH_GEG said:
A bit. The architecture of each 4-core chip is very different; designed for different tasks. The weaker core is A7 while the more powerful core is A15. Your comparison is kind of like cylinder deactivation in a car's engine. In a V8, all eight cylinders are identical and they are shut off in pairs to save fuel and match performance to load. Think of Octa as a discreet 4 cylinder engine (A7) alongside a V8 engine (A15). Each, to your point, can be controlled so any or all of their cores can be active at a given time. The 4 cylinder's barely able to reach the low point of the V8's performance curve. Similarly, the V8 can't reach the 4 cylinders lower efficiency range. 85% of apps don't leave the A7 core so that's why it's more efficient than just shutting off some of the A15's cores; especially true at idle. It's actually a pretty neat approach to performance vs. efficiency.
Click to expand...
Click to collapse
FWIW the 2.4 vs 2.3 was a typo. Should be the same number there.
I understand where you're coming from, thanks for that. I just never heard it explained that a A7 at Clock X was different power usage wise than a A15 at Clock X.
If it is, so be it, that answers that. Would be interesting to see real world benchmarks behind it.
-Tony