AMD Beema/Mullins Launch Thread

[Abwx]

I havent seen any BayTrail Z3770 Tablet having more than 3-4 hours Battery life at full load. So I guess BayTrail is pathetic by your standards as well.

Nice to know.

It makes no doubt that theses chips power management is still not fully understood if not at all wich lead to assumptions based on scarce datas.

For the sake of the discussion i explain how power management is done in both Mullins and tablet dedicated BTs.

First is defined is a base frequency at wich the device if fully loaded will dissipate its rated TDP, this frequency seems to be in the 1.2-1.4GHz range for Mullins and 1.3-1.46GHz for BT.

From there if temperature is below the limit the chip is allowed to drain 200-250% of its rated TDP until it reach the thermal limit at wich point it will start to throttle down to its base frequency so the tests where we see the chips running at full turbo with all cores are representative as peak performance but in any way it can be called effective performance as it s not sustainable more than the time necessary for running a bench and yet, as you point it, battery life is expressed in respect of a task that would not drain more than 30-50% of the rated TDP, that is 20% at most of the allowed max power drain.

 

[MarkizSchnitzel]

That is all very nice, but as a complete layman, if these Mullins chips are supposed to be used in Windows 8, isn't the lack of support for connected standby an instant nope-nope?

My T100 uses ~2% battery over night. I basically never have to turn it of. I find that very convenient.

 

[Insert_Nickname]

For the laptops and tablets, AMD could just fuse off the unwanted SATA. Based on what I have read from past reviews reducing I/O in this fashion would reduce power as well.

True, but but then you'd pay with die area. These are low cost chips after all. Though I don't think two more SATA controllers would take that much space. AMD must have felt the trade-off worthwhile.

This is actually why I'm pretty interested about what Carrizo bring to the table. The rumour mill has it that it will have an on-die FCH, that gets disabled when you plug it in a desktop board with a traditional off-die FCH. I find this a very interesting architecture, it's really the best of both worlds. There would then be nothing stopping the mainboard manufacturers from releasing boards the use either the integrated FCH or external FCH, depending on what features are needed. What's more there is nothing stopping Intel from taking a similar approach in the future...

 

[rainy]

TechReport and ExtremeTech are both reporting that the A6-6310 has a 2.4 GHz turbo and a 2.0 GHz base clock. The rest of the Beema processors don't implement turbo.

Somehow it's surprising for me - I thought that A4-5000 (A4-6210) need turbo to the bigger degree than A6-5200 (A6-6310).

Btw, anyone could have an idea what is A8 version (25W?) of Beema?
I would hardly expect it could have stronger GPU or dual channel MC.

Those systems are coming in June, and they'll feature a special, A8-branded version of Beema that AMD offers "for select opportunities." (That model isn't part of the standard Beema lineup.)

http://techreport.com/review/26377/a-first-look-at-amd-mullins-mobile-apu/5

 

[Abwx]

Somehow it's surprising for me - I thought that A4-5000 (A4-6210) need turbo to the bigger degree than A6-5200 (A6-6310).

Btw, anyone could have an idea what is A8 version (25W?) of Beema?
I would hardly expect it could have stronger GPU or dual channel MC.

http://techreport.com/review/26377/a-first-look-at-amd-mullins-mobile-apu/5

The A4-6210 is 1.8GHz, it s the E2-6110 that is 1.5GHz like the A4-5000.

http://www.hardware.fr/news/13677/amd-annonce-beema-mullins-maj.html

 

[Ancalagon44]

The A4-6210 is 1.8GHz, it s the E2-6110 that is 1.5GHz like the A4-5000.

http://www.hardware.fr/news/13677/amd-annonce-beema-mullins-maj.html

The A4-5000 is rated to run continuously at 1.5GHz, while the E2-6110 is rated to run at up to 1.5Ghz. At the moment, we dont know how long it can run at that frequency for.

 

[Abwx]

The A4-5000 is rated to run continuously at 1.5GHz, while the E2-6110 is rated to run at up to 1.5Ghz. At the moment, we dont know how long it can run at that frequency for.

Hardware.fr got confirmation of both manufacturers and frequencies of thoses chips from AMD after the sites published their slides as pointed in their article and the 6110 has a fixed 1.5GHz frequency.

http://www.hardware.fr/news/13677/amd-annonce-beema-mullins-maj.html

 

[rainy]

The A4-6210 is 1.8GHz, it s the E2-6110 that is 1.5GHz like the A4-5000.

http://www.hardware.fr/news/13677/amd-annonce-beema-mullins-maj.html

Honestly, it doesn't matter - IMHO, they should implement turbo in lower models of Beema (A4-6210/E2-6110) including also dual cores.

 

[el etro]

TechReport and ExtremeTech are both reporting that the A6-6310 has a 2.4 GHz turbo and a 2.0 GHz base clock. The rest of the Beema processors don't implement turbo.

Puma IPC is the same as Jaguar.

 

[Abwx]

Honestly, it doesn't matter - IMHO, they should implement turbo in lower models of Beema (A4-6210/E2-6110) including also dual cores.

It is obvious that the lower power variants got most of the effort on this front although it s clear that the 6210 should had been granted a 200-300MHz turbo up to 2.0 even if this implied a 1.7 base frequency, same for the 6110 wich would had deserved the same amount, as for the dual cores it s also obvious that they re the most theoricaly turbo friendly by the virtue of their halving.

 

[William Gaatjes]

If amd would release such a chip on a small board for a low price, it could turn into an x86 compatible raspberry pi. This would mean the entire computer interested world would jump on the bandwagon. An addon such a general cheap pic,avr or arm mcu with an usb interface on board as well to use for general purpose input output (GPIO) functions would make it really popular. AMD could do it. I would expect an amd x86soc arduino port to follow soon. It would be really popular and AMD can pull it off at a low price.

 

[strata8]

If amd would release such a chip on a small board for a low price, it could turn into an x86 compatible raspberry pi. This would mean the entire computer interested world would jump on the bandwagon. An addon such a general cheap pic,avr or arm mcu with an usb interface on board as well to use for general purpose input output (GPIO) functions would make it really popular. AMD could do it. I would expect an amd x86soc arduino port to follow soon. It would be really popular and AMD can pull it off at a low price.

Maybe possible with the E1-Micro 6200T. But it really depends how much that costs and it would need a heatsink unlike the Raspberry Pi.

I'd love to see a fanless mini PC using the A4-Micro 6400T or A10-Micro 6700T a la AMD's nano PC concept:

But I just know that it's never going to happen.

edit: Interesting point about STAPM from the TechReport article:

"That gain need not come at the cost of battery life, either. On the contrary, a faster-running processor can complete tasks more quickly and thus spend more time at idle, which "actually saves energy in many common use-cases," the company tells us."

Obviously there's a limit (eg, lower efficiency as clock speeds and voltage increase), but it makes sense.

 

[Homeles]

If Beema/Mullins has taught us anything, it's that AMD's "ticks" are, counterintuitively, better than their "tocks." That is, their architectural refinements and refreshes have proven to be a greater improvement than the new architectures themselves.

Bulldozer was a sidegrade, whereas Piledriver moved forward a fair bit. Trinity, in my opinion, was another sidegrade on the CPU front, while its refinement Richland again was able to move the perf/watt bar forward. Kaveri was a highly-hyped, and again underwhelming design, which appears to have a lot of features and units locked on die, only to possibly unlocked with a future release.

And here we are again with Jaguar and Puma. The former, while inarguably a significant improvement over Bobcat, was a terribly hamstrung design. It's a damn shame that the consoles are stuck with such an unpolished product.

 

[cbn]

Maybe possible with the E1-Micro 6200T. But it really depends how much that costs and it would need a heatsink unlike the Raspberry Pi.

I'd love to see a fanless mini PC using the A4-Micro 6400T or A10-Micro 6700T a la AMD's nano PC concept:

But I just know that it's never going to happen.

I really like the idea of x86 mini-desktop, but the problem they all seem to have is poor value. I'm been thiking this is because they are usually a very low volume item and thus carry niche pricing.

In order for me to be interested it really needs to be cheap, otherwise laptop used as a desktop (or blow out sale mini-itx) seems to be the much better bargain.

 

[TuxDave]

Leakage is proportional to voltage, to reduce CPU leakage by 19% and GPU leakage by 40% they sould had reduced voltage by as much following your logic wich is of course completely wrong given thoses numbers.

Actually a more common method to reduce leakage is to reduce gate width or increase threshold voltage (threshold voltage != supply voltage) at the cost of frequency or iso-frequency if you have "better transistors".

Using classical equations, leakage reduction is linear to gate width and inversely exponential to threshold voltage. Supply voltage shows up as an exponential term but it's divided by threshold voltage. So you can see that width and threshold voltage can play a large factor in reducing leakage without reducing supply voltage.

Edit: Of course, the biggest bang for your buck is architectural improvements such as power gating.

 

[USER8000]

AFAIK,isn't console SOC production moving to GF??

I wonder if it will mean,newer revisions of both consoles,ie,not changing in performance,but perhaps with lower power draw??

 

[Ajay]

And here we are again with Jaguar and Puma. The former, while inarguably a significant improvement over Bobcat, was a terribly hamstrung design. It's a damn shame that the consoles are stuck with such an unpolished product.

Yeah, I'd think that, for a change, Sony and especially MS would be better off moving to Puma. I don't think it would affect developer assets and tool chains but would make a decent boost in performance and perf/watt. Of course, footing the bill for this may not be pleasant, but it seems like the consoles will need more of an update to stay relevant than the last go around. Maybe it could be done at they same time as the 14nm shrink.

 

[monstercameron]

Yeah, I'd think that, for a change, Sony and especially MS would be better off moving to Puma. I don't think it would affect developer assets and tool chains but would make a decent boost in performance and perf/watt. Of course, footing the bill for this may not be pleasant, but it seems like the consoles will need more of an update to stay relevant than the last go around. Maybe it could be done at they same time as the 14nm shrink.

I doubt it, there were no ipc increases in puma cores and console chips are designed to run at constant operation, which is different from how battery powered devices usually operate.

 

[blackened23]

Yeah, I'd think that, for a change, Sony and especially MS would be better off moving to Puma. I don't think it would affect developer assets and tool chains but would make a decent boost in performance and perf/watt. Of course, footing the bill for this may not be pleasant, but it seems like the consoles will need more of an update to stay relevant than the last go around. Maybe it could be done at they same time as the 14nm shrink.

I don't see how this can happen and how MS could possibly spin a randomly updated console with higher performance to their customers. Any performance increase would seem to be minimal as the cores in the XB1/PS4 were already aggressively clocked for console use with turbo and their respective marketing departments could not possibly spin this to consumers.

It would be a situation of "Uh hey guys, we're going to sell the same console which is 10-15% faster (it wouldn't be that much, but theoretically) at the same price". I just can't ever see that happening. A silent update along those lines would make customers rage, ESPECIALLY on the xbox 1 side. They would feel obligated to buy a new console to get better performance, because let's face it, the XB1 aimed far too low (the PS4 did as well, just not quite as bad as the XB1).

There is just no way the marketing departments of sony or MS could positively spin a silent performance update. Now with the xbox 360, they did various die shrinks which did not affect performance. They were simply done to lower costs and NOTHING more. I could see MS and Sony doing that. But a silent hardware upgrade to increase performance? No. That won't happen in the console world IMO.

Think this through. This would hit the internet. And there would be a massive consumer backlash. If there were a silent update that provided x% more performance, particularly on the XB1 side, consumers would NOT be pleased. They would expect to return their consoles for a free update. Which of course , aint happening. I could see die shrinks happening 5-10 years from now whenever the newer nodes reach reasonable prices, but not silent performance updates. That just hasn't happened in the console world, ever, that i'm aware of. Die shrinks have happened, but of course die shrinks only lower costs as to increase profits. The consumer backlash for those who bought older consoles would be considerable.

 

[Idontcare]

I benefit from this post to add that generaly i find all power measurements from various sites as extremely unprofessional, while benches will give about the same scores from one site to another power measurements for thoses low power solutions can vary within almost a 2 ratio but still no one seems to care that such a discretanpcy would be undoubtly questionned if it was Cinebench or Firestrike, most ironic is that if people working in that industry would perform measurements with such random methodology they would be immediatly fired for incompetence.

To their credit though, you are getting what you pay for.

The person working for an employer who would get fired would be getting fired because their employer would not be getting what they (the employer) was paying for (quality data generation and collection).

As the target audience that we are, an audience that refuses to pay for content generation let alone many of us* are using ad-blocker to reduce what little revenue is created by our presence on the web, we get what we deserve and really have no room to complain IMO.

* personal note - I am not using ad-blocker or anything that reduces advertising revenue, but many people do and aren't ashamed to admit it

 

[USER8000]

I don't see how this can happen and how MS could possibly spin a randomly updated console with higher performance to their customers. Any performance increase would seem to be minimal as the cores in the XB1/PS4 were already aggressively clocked for console use with turbo and their respective marketing departments could not possibly spin this to consumers.

It would be a situation of "Uh hey guys, we're going to sell the same console which is 10-15% faster (it wouldn't be that much, but theoretically) at the same price". I just can't ever see that happening. A silent update along those lines would make customers rage, ESPECIALLY on the xbox 1 side. They would feel obligated to buy a new console to get better performance, because let's face it, the XB1 aimed far too low (the PS4 did as well, just not quite as bad as the XB1).

There is just no way the marketing departments of sony or MS could positively spin a silent performance update. Now with the xbox 360, they did various die shrinks which did not affect performance. They were simply done to lower costs and NOTHING more. I could see MS and Sony doing that. But a silent hardware upgrade to increase performance? No. That won't happen in the console world IMO.

Think this through. This would hit the internet. And there would be a massive consumer backlash. If there were a silent update that provided x% more performance, particularly on the XB1 side, consumers would NOT be pleased. They would expect to return their consoles for a free update. Which of course , aint happening. I could see die shrinks happening 5-10 years from now whenever the newer nodes reach reasonable prices, but not silent performance updates. That just hasn't happened in the console world, ever, that i'm aware of. Die shrinks have happened, but of course die shrinks only lower costs as to increase profits. The consumer backlash for those who bought older consoles would be considerable.

Or we could see a second revision PS4 and XBox One when the SOCs are made at GF,with rejigged internals(lower cost cooling if the TDP is reduced) at the end of the year??

 

[blackened23]

No? I'd expect Sony and MS to do die shrinks when economically feasible. Obviously, 20nm is expensive now but it will be economically feasible at some point. And when it is...i'd expect die shrinks. Nothing more. Die shrinks lower costs (when the wafer costs depreciate) and do not change performance.

They will not do a silent performance update or switch SOCs for the hell of it. A silent performance update, in particular, would result in a massive consumer backlash. Also, the cooling costs are not an issue. Of the entire BOM that is the last consideration because it's pennies.

 

[USER8000]

No? I'd expect Sony and MS to do die shrinks when economically feasible. Obviously, 20nm is expensive now but it will be economically feasible at some point. And when it is...i'd expect die shrinks. Nothing more. Die shrinks lower costs (when the wafer costs depreciate) and do not change performance.

They will not do a silent performance update or switch SOCs for the hell of it. A silent performance update, in particular, would result in a massive consumer backlash. Also, the cooling costs are not an issue. Of the entire BOM that is the last consideration because it's pennies.

I am not talking about a performance update. Beema/Mullins are made on a different process node to Temash/Kabini,which is probably contributing to improvements in performance/watt. That means at the same performance level,Beema/Mullins has lower power consumption than Temash/Kabini. The new range tops out at a 15W TDP and the old one at a 25W TDP,which means the newer chips are much easier to cool(officially).

Also,it seems the GF process might be slightly denser too.

Both console SOCs are shifting over to GF at some point this year. If the TDP and power consumption drops,that means a cheaper cooling system and smaller power supply and power delivery system needed. It would still save money overall.

The various versions of the PS3 and XBox360,were cheaper to make not only due to the smaller chips,and unification of various ones,but also down to reductions in power supply size and cooling system size. The latter two saw massive reductions in size and capacity and contributed to a downsizing of the console,and easier assembly.

 

[Rezist]

I don't see how this can happen and how MS could possibly spin a randomly updated console with higher performance to their customers. Any performance increase would seem to be minimal as the cores in the XB1/PS4 were already aggressively clocked for console use with turbo and their respective marketing departments could not possibly spin this to consumers.

It would be a situation of "Uh hey guys, we're going to sell the same console which is 10-15% faster (it wouldn't be that much, but theoretically) at the same price". I just can't ever see that happening. A silent update along those lines would make customers rage, ESPECIALLY on the xbox 1 side. They would feel obligated to buy a new console to get better performance, because let's face it, the XB1 aimed far too low (the PS4 did as well, just not quite as bad as the XB1).

There is just no way the marketing departments of sony or MS could positively spin a silent performance update. Now with the xbox 360, they did various die shrinks which did not affect performance. They were simply done to lower costs and NOTHING more. I could see MS and Sony doing that. But a silent hardware upgrade to increase performance? No. That won't happen in the console world IMO.

Think this through. This would hit the internet. And there would be a massive consumer backlash. If there were a silent update that provided x% more performance, particularly on the XB1 side, consumers would NOT be pleased. They would expect to return their consoles for a free update. Which of course , aint happening. I could see die shrinks happening 5-10 years from now whenever the newer nodes reach reasonable prices, but not silent performance updates. That just hasn't happened in the console world, ever, that i'm aware of. Die shrinks have happened, but of course die shrinks only lower costs as to increase profits. The consumer backlash for those who bought older consoles would be considerable.

The 360/PS3 shrinks were also done to decrease power the RROD and YLOD costs did add up.

 

[blackened23]

The 360/PS3 shrinks were also done to decrease power the RROD and YLOD costs did add up.

Fair enough, but the main thing I was trying to convey was a silent performance update not being realistic. Such a SOC update for increased performance would not be inconsequential with both consoles, but *especially* the XB1; the consumer backlash would be considerable. Most younger folks these days are more in the know on the web with these type of things, and the backlash would be massive to say the least. I'm also not aware of any prior precedent of performance updates in any prior console, i'm pretty certain it has never happened. Although die shrinks will definitely happen when it results in lower BOM for the companies involved.