[Christian Troy]

Tzul, on 06 February 2013 - 06:42 PM, said:

Update on the current situation:

It looks like Actions / Ainol removed the "CPU part" line from /proc/cpuinfo in their latest firmwares!
A bit late for cover up tactics, makes them only look worse.

LOL! Like the fake 1.2 GHz of the allwinner a10!

[Nickos-V]

Jiefurui, on 07 February 2013 - 08:35 AM, said:

Right now I'm having a problem with the computer not recognizing the tablet once it has been connected so I could update the firmware. Anyone knows how it should be done?

Start a new thread for that. This one is talking about whether ATM7029 has Cortex A9 or A5 CPU and how this affects tablet performance and Android benchmarks like Antutu.

Christian Troy, on 07 February 2013 - 09:14 AM, said:

LOL! Like the fake 1.2 GHz of the allwinner a10!

I think Actions and AllWinner are cousins!!! =) Both like to tell big lies about their specs. You forgot the A31 chip. I heard 1.4 GHz when first coming out and it ended up being 1 GHz instead.

Actually, a few tablets clocked the A10 @ 1.2 GHz but it was marketed as 1.5 GHz chip by AllWinner. Though most tablets ran A10 @ 1 GHz like you say.

A10 - marketed 1.5 GHz - actual 1 GHz (few coming w/1.2 GHz overclocked speed)
A31 - marketed 1.4 GHz - actual 1 GHz
ATM7029 - marketed 1.5 GHz Cortex A9 - actual 1.2 GHz Cortex A5. Huge lie!!! They have just trashed their reputation by doing this.
RK2918 - marketed 1.2 GHz - actual 1 GHz
8726-MX - marketed 1.5 GHz - actual 1.2 (or 1.3?) GHz

Rockchip even lied with RK2918 because their site said 1.2 GHz but my tablet & others ran RK2918 @ 1 GHz. At least it was a small lie by Rockchip and they told the truth with RK3066.

Edit: made a couple of changes and added in 8726-mx

This post has been edited by Nickos-V: 07 February 2013 - 11:42 AM

[Christian Troy]

The worst thing in that Allwinner lie was that any frequency >= 1008000 was reported in the sysfs cpufreq_cur_info as 1200000 (so you had no chance to see if it was running a 1, 1.2 or even 1.5 GHz, but 99.9999999999999% where really running at 1GHz)

[munka]

Nickos-V, on 07 February 2013 - 11:38 AM, said:

Start a new thread for that. This one is talking about whether ATM7029 has Cortex A9 or A5 CPU and how this affects tablet performance and Android benchmarks like Antutu.


I think Actions and AllWinner are cousins!!! =) Both like to tell big lies about their specs. You forgot the A31 chip. I heard 1.4 GHz when first coming out and it ended up being 1 GHz instead.

Actually, a few tablets clocked the A10 @ 1.2 GHz but it was marketed as 1.5 GHz chip by AllWinner. Though most tablets ran A10 @ 1 GHz like you say.

A10 - marketed 1.5 GHz - actual 1 GHz (few coming w/1.2 GHz overclocked speed)
A31 - marketed 1.4 GHz - actual 1 GHz
ATM7029 - marketed 1.5 GHz Cortex A9 - actual 1.2 GHz Cortex A5. Huge lie!!! They have just trashed their reputation by doing this.
RK2918 - marketed 1.2 GHz - actual 1 GHz
8726-MX - marketed 1.5 GHz - actual 1.2 (or 1.3?) GHz

Rockchip even lied with RK2918 because their site said 1.2 GHz but my tablet & others ran RK2918 @ 1 GHz. At least it was a small lie by Rockchip and they told the truth with RK3066.

Edit: made a couple of changes and added in 8726-mx

The Amlogic chip is a 1.5ghz part that Ainol underclock because their implementation of it gets too hot. I don't really blame Amlogic for that, its still their highest supported speed if you give it an opportunity to cool itself and enough juice.

[Tzul]

Christian Troy, on 08 February 2013 - 12:29 AM, said:

The worst thing in that Allwinner lie was that any frequency >= 1008000 was reported in the sysfs cpufreq_cur_info as 1200000 (so you had no chance to see if it was running a 1, 1.2 or even 1.5 GHz, but 99.9999999999999% where really running at 1GHz)

Wow, that's really bad. Haven't heard about this before, but I don't own an Allwinner based tablet anyway.

munka, on 10 February 2013 - 07:02 AM, said:

The Amlogic chip is a 1.5ghz part that Ainol underclock because their implementation of it gets too hot. I don't really blame Amlogic for that, its still their highest supported speed if you give it an opportunity to cool itself and enough juice.

This "too hot" card is often pulled as an excuse. Is there any proof for that? I'm guessing that the absence of 1.5GHz on the AML8726-MX has more to do with stability and power consumption than with heat. According to its quick reference manual, the AML8726-MX's CPU may run with "up to 1.5GHz" and may get quite hot.
 qrm.png (80.81K)
Number of downloads: 23
They haven't even listed an absolute maximum temperature there. The "recommended" maximum is 125°C. However, that's the internal junction temperature. The outside / ambience of the chip must always be quite a bit less than that during operation, due to imperfect heat conductivity.
Does Ainol even actually make the firmware? I've had the impression that it's done by Amlogic, especially the lowlevel stuff. In that case, Amlogic would underclock their implementation themselves.

Anyway, Feiyu's latest "happy spring festival" firmware for Ainol Amlogic tablets has 1.5GHz unlocked, and it appears to run fine. On some tablets at least. There could very well be copies of the AML8726-MX that just aren't stable at 1.5GHz and reasonable voltages in their given tablet environment.
I'm sure the AML8726-MX could do 2GHz and more if you crank up the voltage and battle the heat with liquid nitrogen, but does that mean they should market the chip as 2.0GHz?

This post has been edited by Tzul: 10 February 2013 - 11:00 AM

[Nickos-V]

munka, on 10 February 2013 - 07:02 AM, said:

The Amlogic chip is a 1.5ghz part that Ainol underclock because their implementation of it gets too hot. I don't really blame Amlogic for that, its still their highest supported speed if you give it an opportunity to cool itself and enough juice.

Actually, the chip is not designed well and that is why it was clocked lower. Same thing goes for AllWinner A31.

Dual core 40nm in theory can go to 1.5 GHz. Dual core 28nm in theory can reach 2 GHz. Quad core versions are 200 MHz less. ie. Tegra3 @ 1.3 GHz. A perfect chip design would reach those max available speeds and only get slightly hot while 3D gaming. ie. A31 should have been able to run at 1.3 or 1.2 GHz had the chip design been really great. 8726-MX @ 1.5 GHz. But that is not the case for either of those chips and why they have to be clocked to lower speeds.

RK3066 (40nm) is able to handle 1.6 GHz while still working very well. 8726-MX tried for 1.5 GHz but was getting very hot while 3D gaming. A complaint I heard a couple of times from people. That is a clear sign the chip was not able to properly handle the higher clock speed. 8726-MX is very likely a 1.2 GHz chip running at higher clock speed.

You can take any CPU out there and overclock it, but you run into higher heat and stability issues. 8726-MX @ 1.5 GHz was set way too high for the chip to properly handle. It was clocked that high to compete against RK3066 @ 1.6 GHz.

For those that want to 3D game. Are they supposed to play 5 mins. at a time to allow 8726-MX to stay cool??? A chip should be able to run stable and slightly hot for at least 1 or more hours while 3D gaming. My desktop, laptop & RK3066 tablet can all do this but 1.5 GHz 8726-MX tablet certainly cannot. I hear it gets extremely hot while 3D gaming @ 1.5 GHz.

This post has been edited by Nickos-V: 10 February 2013 - 12:00 PM

[Tzul]

Nickos-V, on 10 February 2013 - 11:56 AM, said:

Actually, the chip is not designed well and that is why it was clocked lower.

That's quite an allegation. Do you have any proof? Amlogic designed the chip for up to 1.5GHz according to their documentation, and it can actually deliver that. So far munka is right.
I was just questioning whether "it gets too hot" is the real reason you don't see 1.5GHz on Amlogic that often.
It's similar to a chicken and egg problem:
1. Is 1.5GHz not used because the chip gets too hot?
2. Is 1.5GHz not used because some chips would be unstable without a voltage bump, which in turn would increase heat too much?

Nickos-V, on 10 February 2013 - 11:56 AM, said:

Dual core 40nm in theory can go to 1.5 GHz. Dual core 28nm in theory can reach 2 GHz. Quad core versions are 200 MHz less. ie. Tegra3 @ 1.3 GHz.

Maybe that's a rule of thumb for ARM Cortex CPUs, but it's not true in general. Given a voltage boost and appropriate cooling system, you can easily reach higher frequencies than that. However, cooling is indeed a problem in mobile computing, as is power consumption.
Also, Cortex A9 dual core 28nm can reach 3.1GHz in practice.

Nickos-V, on 10 February 2013 - 11:56 AM, said:

RK3066 (40nm) is able to handle 1.6 GHz while still working very well. 8726-MX tried for 1.5 GHz but was getting very hot while 3D gaming. A complaint I heard a couple of times from people. That is a clear sign the chip was not able to properly handle the higher clock speed. 8726-MX is very likely a 1.2 GHz chip running at higher clock speed.

The thing with forums such as this is that you are much more likely to hear complaints and problems than to hear that everything is working fine.

Nickos-V, on 10 February 2013 - 11:56 AM, said:

For those that want to 3D game. Are they supposed to play 5 mins. at a time to allow 8726-MX to stay cool??? A chip should be able to run stable and slightly hot for at least 1 or more hours while 3D gaming. My desktop, laptop & RK3066 tablet can all do this but 1.5 GHz 8726-MX tablet certainly cannot. I hear it gets extremely hot while 3D gaming @ 1.5 GHz.

You heard wrong.
I know what "extremely hot" is. I once had a notebook with AMD Turion dual core CPU that reached 95°C under load, so hot that it started throttling the frequency to reduce heat.

As I mentioned before, I'm just testing Feiyu's "happy spring festival" JB 4.1.2 release on my Aurora 2.
CPU at 1512MHz, GPU at 333MHz or 400MHz. Antutu 3.1.1 score of 9942, Vellamo Metal 580. It does get a bit warm under stress, but so far nothing I'd call "extremely hot". And it ran absolutely stable so far, no crashes yet.
Also, keep in mind that most games are either CPU or GPU bound. It is very unlikely to come across a game that pushes both CPU and GPU to their limits at the same time. I've been running 10 minutes or so of the Epic Citadel benchmark, for example, which clearly is limited by GPU perfomance. Whether the CPU runs at 1320MHz or 1512MHz doesn't make a noticeable difference there.

The power consumption is quite alarming, though. I measured battery discharge currents of nearly 1.7A and power consumption of over 6W - with the screen at maximum brightness (which I barely use in practice).

[Nickos-V]

Tzul, on 10 February 2013 - 01:20 PM, said:

That's quite an allegation. Do you have any proof? Amlogic designed the chip for up to 1.5GHz according to their documentation, and it can actually deliver that. So far munka is right.
I was just questioning whether "it gets too hot" is the real reason you don't see 1.5GHz on Amlogic that often.
It's similar to a chicken and egg problem:
1. Is 1.5GHz not used because the chip gets too hot?
2. Is 1.5GHz not used because some chips would be unstable without a voltage bump, which in turn would increase heat too much?

Only way for me to prove it is by buying an 8726-MX tablet, having proper test equipment to monitor chip temps and then running lots of tests to stress CPU and CPU+GPU. Otherwise, I can only base my assumptions on what I hear and see.

If the chip was really meant to run @ 1.5 GHz then Ainol would not have under-clocked it. The chip is being pushed past its limit. Ainol realized this and reduced the clock frequency on it. ie. getting too hot @ 1.5 GHz. A tablet maker reducing the frequency of the chip is not a good sign! Points to some type of issue with the chip. Simple as that.

Just because you can run a chip (whether CPU, GPU or combined) at a certain speed does not mean it was meant to handle that high clock speed. Some chips can be clocked too high which causes issues like heat and instability. Too much heat shows chip is pushed way too high. Also, people should not have to boost voltage to run a chip stable. Doing this means the chip is not meant to run at the higher clock speed in 1st place or not the highest quality to begin with.

Last I checked this is an ARM site so clock rates listed should be assumed for 40nm & 28nm "ARM chips" unless specified otherwise. =) I was not talking x86 but ARM. I have seen almost no posts for x86 on this site.


PS, try running Asphalt 7 @ 1.5 GHz for 1 hour and see how hot that 8726-MX gets. A real intense 3D game will push that ARM chip to the max. ie. pushing both the CPU & GPU to extremes.

PPS, the high heat maybe caused by the GPU clock speed instead. Both the CPU & GPU are integrated on one chip. One way to know for sure if CPU is the actual problem is by looping a CPU benchmark/test for 60 mins and monitoring chip temp & checking for errors. That will remove the GPU factor. Then can run bench that stresses both CPU & GPU and any temp higher than recorded just for CPU test would be because of the GPU. High detail 3D gaming tests out both at the same time and really pushes and stresses out an ARM chip.

I cannot say for sure if the problem is because of the CPU, GPU or CPU+GPU clock speeds contributing to the very high heat. Whatever the case may be; it is a problem that Ainol decided to fix by lowering CPU clock speed. Ainol realized the high heat in the 8726-MX chip was an issue and clocked it lower to reduce the heat problem. I have seen multiple people complain about their hands getting really hot when 3D gaming with 8726-MX. One person saying he owned both RK3066 & 8726-MX tablets and with Rockchip getting slightly hot while Amlogic becoming very hot while 3D gaming.

[Jote]

i heard that on first firmwares the 8726 was clocked 1.5GHz but was undeclocked(undervolted) on later versions.

And all this is offtopic.

[vvuv]

In fact AllWinnerTech A10 ran stable on 1.2ghz. As of rockchip, Officially Rockchip RK2918 always advertised as 1Ghz (max 1.2ghz)

Proof:
http://www.rock-chip...%81%232_max.jpg

The Action ATM7029 GPU is Gc1000 Plus from Vivante, Not really same as GC1000. Quite hard to say whether ATM7029 utilize the a5/a9.

Too early yet to say about AllWinnerTech A31. Quite sure they will clock it at around 1.2-1.4ghz in further firmwares. Though it appear like Rockchip is the best among them.

[Tzul]

Nickos-V, on 10 February 2013 - 03:06 PM, said:

If the chip was really meant to run @ 1.5 GHz then Ainol would not have under-clocked it. The chip is being pushed past its limit. Ainol realized this and reduced the clock frequency on it. ie. getting too hot @ 1.5 GHz. A tablet maker reducing the frequency of the chip is not a good sign! Points to some type of issue with the chip. Simple as that.

Wrong. The chip is meant to run at 1.5GHz. But it's entirely possible that Ainol's cooling system is undersized for that (it's basically just a piece of graphite paper connecting the SoC and RAM chips to the metal back plate of the LCD assembly). In that case, the chip isn't pushed past its limit, but Ainols cooling system is. Simple as that.

Nickos-V, on 10 February 2013 - 03:06 PM, said:

Just because you can run a chip (whether CPU, GPU or combined) at a certain speed does not mean it was meant to handle that high clock speed. Some chips can be clocked too high which causes issues like heat and instability. Too much heat shows chip is pushed way too high. Also, people should not have to boost voltage to run a chip stable. Doing this means the chip is not meant to run at the higher clock speed in 1st place or not the highest quality to begin with.

Just because you see a certain CPU underclocked in some tablets does not meant that it wasn't meant for higher clock speeds.
Also, how much heat is too much? Depends entirely on the cooling system. You are aware that with clock frequency scaling (CPU running at different speeds depending on load, to save power), each frequency state usually has its own voltage, right? If you run a chip at X GHz and the voltage is too low for that frequency, it'll get unstable and crash. If the voltage is too high, it'll run stable but get hotter than necessary. If the voltage is MUCH too high, the chip will get so hot that it might be damaged - if it doesn't have a throttling mechanism implemented (no idea if ARM CPUs do).
The point is, for maximum efficiency, you need to find the perfect voltage for each frequency state. Not too low and not too high. This can vary from chip to chip, even if they're of the same type.

So, if somebody complains about too much heat with an AML8726-MX at 1.5 GHz, I'd first ask about the firmware used (unfortunately, the CPU voltage settings aren't easily obtainable, AFAIK).
Whatever Feiyu/Amlogic did in Feiyu's latest release works really well for my Aurora 2.

Nickos-V, on 10 February 2013 - 03:06 PM, said:

Last I checked this is an ARM site so clock rates listed should be assumed for 40nm & 28nm "ARM chips" unless specified otherwise. =) I was not talking x86 but ARM. I have seen almost no posts for x86 on this site.

Um, I was talking about ARM (except when mentioning the excessive heat of an AMD Turion X2).
This is an ARM chip that clearly proves your rule of thumb wrong. Because it's not just the manufacturing size (40nm, 28nm, ...) but also other factors of the process that influence power consumption and thus the maximum usable clock speed. TSMC offers four difference choices for 28nm, for example.

Nickos-V, on 10 February 2013 - 03:06 PM, said:

PS, try running Asphalt 7 @ 1.5 GHz for 1 hour and see how hot that 8726-MX gets. A real intense 3D game will push that ARM chip to the max. ie. pushing both the CPU & GPU to extremes.

I'll try Asphalt 7 when I get the chance. However, as I said before, it's very unlikely to come across a game that pushes both CPU and GPU "to extremes" simultaneously. Games are usually limited by either CPU or GPU, and with 3D heavy games it's often the latter, of course. If you don't believe that, look it up, there are plenty of articles about this fact. Mostly for PC games, but the same principles apply. The difference between running the Epic Citadel benchmark at 792MHz and 1512MHz (almost twice the speed) on my tablet is literally just 1-2 fps. You can't get a clearer sign that this is GPU limited and doesn't even begin to push the CPU to extremes...

Jote, on 10 February 2013 - 05:02 PM, said:

i heard that on first firmwares the 8726 was clocked 1.5GHz but was undeclocked(undervolted) on later versions.

And all this is offtopic.

Depends on the tablet. With the Elf 2, Aurora 2 and Crystal, it certainly was not at 1.5GHz from the beginning. On the Fire and Hero perhaps.

Yeah, you're right, this is off topic, sorry about that.

vvuv, on 11 February 2013 - 07:31 PM, said:

The Action ATM7029 GPU is Gc1000 Plus from Vivante, Not really same as GC1000. Quite hard to say whether ATM7029 utilize the a5/a9.

Have you somehow missed the first two pages of this topic?
It's crystal clear that the ATM7029 has a Cortex-A5 as CPU!

This post has been edited by Tzul: 12 February 2013 - 10:06 AM

[Christian Troy]

Anyway what matters to me is that the soc manufacturer is cooperative and gives access to full source code to be able to build third party ROMs.
In this case amlogic has been flawless, where ainol didn't answer they did, if people asks me for advises I'll say to buy an amlogic product.

Rockchip does good products but they don't offer any kind of support to develop an aosp based ROM.

[vvuv]

I can not say anything bad regarding the AMLogic AML8726-MX. Some tablets are clocked at around 1.3ghz others at 1.5ghz.

In fact I have found once a custom firmware for a tablet that the official firmware was clocked to 1.3ghz. Installed it and tested various games such as asphalt7/dead trigger ran it for quite long and can not say that it was getting hot. Just very slightly warm. Though I can not say that the same games ran slower on 1.3ghz, They ran as fast as they should be under both, No lags etc. The only thing however, Stable Android 4.1 was released for RK3066 much quicker than AML-8726MX. And of course RK3066 perform better due to the fact that the gpu is quad core mali400 comparing to the dual core mali400 in amlogic. (Not to mention the additional 100mhz per core)

This post has been edited by vvuv: 12 February 2013 - 02:59 PM

[Nickos-V]

Tzul, on 12 February 2013 - 10:06 AM, said:

Wrong. The chip is meant to run at 1.5GHz.

Not just Ainol doing this. Same for Zenithink too. They run C93 @ 1.2 GHz and probably their other 8726-MX tablets too.
http://www.aliexpres.../344858872.html
http://zenithink.com/Eproducts_C93.php

Wait. Same thing for Onda 8726-MX tablets too. 1.2 GHz.
http://chinatablets....?p=2143&lang=en

It is not just Ainol (1 tablet maker). 3 tablet makers I have found so far searching quickly. Tablet makers would prefer to run 8726-MX @ 1.5 GHz if it could handle high heat and/or be stable but it just cannot. You think tablet makers like having to under-clock 8726-MX to 1.2 (or 1.3) GHz??? They do not like doing this but they realized the chip was only meant for 1.2/1.3 GHz max speed to run with lower heat and/or stable.

Quote

Um, I was talking about ARM (except when mentioning the excessive heat of an AMD Turion X2).
This is an ARM chip that clearly proves your rule of thumb wrong.

OK. You found a post saying 28nm can reach 3.1 GHz. GREAT!!! Where can I buy a tablet with 3 GHz ARM chip??? Please tell which tablet has a 3 GHz dual or quad core ARM chip. I and many others would love to buy one!!! Oh wait, you mean I cannot get one. So, 3.1 GHz is more of a "theory" whereas 2 GHz is actual (or reality). Show me a 28nm ARM chip inside some tablet that runs at 3 GHz and I will say you are right!!! I wonder why Rockchip is only giving us a 1.6 GHz quad then with RK3188? I guess they should have been able to do 2 GHz without any issues since 3 GHz appears to be the max speed? Or maybe I am right instead?

You may have missed the part that read "ARM A9 at TSMC 28HPM delivers performance speed range from 1.5GHz to 2.0GHz, suitable for mobile computing." 3 GHz is for high performance. So, maybe the chip can get pushed to 3 GHz but I can only imagine the heat it will produce at that speed.


Edit:
TMSC created a 3.1 GHz Cortex A9 dual to test out the max speed achievable. No information what type of cooling was used for it. Can only assume that it will require heatsink & fan. For tablets & phones, 2 GHz is the max speed you will see with 28nm.
http://www.geek.com/...3-1ghz-2012054/

This post has been edited by Nickos-V: 12 February 2013 - 06:04 PM

[vvuv]

Maybe tablet makers thought that 1.3ghz would be optimal for most customers and the additional 200mhz per core would not give much advantage?

Maybe when clocked to 1.5ghz it takes significantly more battery without giving any significant advantage in term of performance?

This post has been edited by vvuv: 12 February 2013 - 07:56 PM

[Nickos-V]

vvuv, on 12 February 2013 - 07:55 PM, said:

Maybe tablet makers thought that 1.3ghz would be optimal for most customers and the additional 200mhz per core would not give much advantage?

It was actually because of high heat. Too many people complaining tablet was getting hot where the ARM chip is while 3D gaming. Otherwise these tablets makers would have kept it at 1.5 GHz. I think only a few 3D games caused this to happen.

Games like Asphalt 7 & WildBlood are very intense and really push the GPU to the limit. 8726-MX appears to have trouble running those two games in normal mode?
http://www.slatedroi...post__p__459050
http://www.slatedroi...post__p__603986

Anyways, this thread has gone way off-topic so best to continue this in another thread instead. Though I've said enough. Also, I get part of the blame for not moving my 1st off-topic post into a new thread and linking to it instead.

Back to ATM7029 chip discussion. Later,

This post has been edited by Nickos-V: 13 February 2013 - 12:02 AM

[F&LSuperstore]

New Heros and Crystal AML8726-MX are running at 1.5Ghz and very stable. They can get hot after gaming for an hour, but they are stable. Originally it was heat, then it was stability. Now they seem to be running good at 1.5Ghz. Possibly a board revision. So Amlogic came through, it just took them awhile to get everything perfect.

[rui no onna]

Nickos-V, on 13 February 2013 - 12:02 AM, said:

It was actually because of high heat. Too many people complaining tablet was getting hot where the ARM chip is while 3D gaming. Otherwise these tablets makers would have kept it at 1.5 GHz. I think only a few 3D games caused this to happen.

So what? Just because the cooling system on tablets can't sustain a 1.5GHz clock doesn't mean the chip itself is unable to reach it. People have a much lower threshold for heat than silicon does. What is the ATM7029 rated for? 125ºC or something? You can get a 3rd degree burn at much lower temps than that. The problem is Chinese tablet manufacturers are advertising the ATM7029 at the max possible frequency when they have it running underclocked. When HP underclocked the Snapdragon S3 CPU on their TouchPad from 1.7GHz to 1.2GHz, they didn't market the device as 1.5GHz. They marketed it as having a 1.2GHz CPU because that's the clockspeed it was running at.

[Rob B]

Is this thread about 8726-MX or ATM7029? Im confused now

[Christian Troy]

It's about comparisons and people should be aware they're buying a shiny bright piece of sheet.