[PCPER] beema and mullins

[Abwx]

Whatever, they seems quite confident for theses products.

“AMD is establishing excellent momentum this year in the low-power, mobile computing market and with ‘Mullins’ and ‘Beema’ coming in 2014 we are not standing still. AMD aims to deliver a set of platforms in the first half of next year that will outperform the competition in graphics and total compute performance in fanless tablets, 2-in-1s and ultrathin notebooks,” said Mark Papermaster

http://www.xbitlabs.com/news/cpu/di..._Roadmap_Beema_and_Mullins_SoC_Confirmed.html


Mr Papermaster is not known as being generaly
overly bullish...

 

[ams23]

Nope, I just smelled something fishy on the slides and AMD historical data doesn't help. But if Intel comes tomorrow saying that Skylake will have double the performance per watt of Broadwell you'll find me having a similar reaction, same with Qualcomm, Samsung, etc.

These AMD slides are not trustworthy when it comes to perf. per watt. The only proper way to measure perf. per watt is to actually measure both perf. and power consumed for each specific task. To only measure perf. and apply it to a generic TDP spec is insanely silly given the broad dynamic range of power consumed on these devices.

 

[Phynaz]

what about the arm trustzone ip?

What about it? It's a TPM module, and is required for Microsoft certification for the next Windows release. Same thing as Intel TXT.

 

[sm625]

Windows 8 is so bad on a tablet that I feel sorry for AMD. They come out with these designs that have absolutely no hope of success. Yes, my clover trail tablet would still suck even if the gpu and cpu were 5 times faster. It's just a bad OS. All the software is too power hungry... a faster cpu or gpu would just kill the battery faster. The GBA emulator kills the battery in 2 hours... and this is a battery that lasts for 6+ hours for general usage. Bluestacks wont even run. So really, what's the point in having x86 when half the stuff dont work anyway.

 

[CHADBOGA]

Well we will have to wait and see about this, right?
A question. Would you like this to be true or not?

Too many people wanting things to be true for outrageous AMD claims, is what AMDZone is for.

Normal forums discuss matters on their merits, with past history and what is known about semiconductor limitations, at the forefront of discussions, not some kind of magic pixie dust.

 

[blackened23]

Windows 8 is so bad on a tablet

No, it really isn't. Win 8.1 is pretty great on a tablet, at least. Once the price point got down to a reasonable level (as it did with Bay Trail) people started buying them in droves - I still think the Asus T100 is the #2 top selling portable device on amazon, behind the chromebook.

 

[inf64]

From here:

1. The new 2014 AMD A-Series low power APU platform, codenamed “Mullins,” is expected to deliver up to 139 percent better productivity performance per watt when compared to the previous generation “Temash” platform. Testing conducted by AMD Performance Labs on optimized AMD reference systems. PC manufacturers may vary configuration yielding different results. PCMark 8 - Home score divided by TDP (W) is used to simulate productivity performance per watt; the Mullins platform (4.5W) scored 1809 while the Temash platform (8W) scored 1343. AMD "Larne" reference platform system used for both APUs. Temash-based AMD A6-1450 quad-core APU with AMD Radeon™ HD 8250 Graphics, 2x2GB of DDR3-1333MHz RAM (running at 1066MHz), Windows 8.1, 13.200.11.0 - 03-Sep-2013 driver. Pre-production engineering sample of “Mullins” quad-core APU with next generation AMD Radeon graphics (model number TBD), 2x2GB DDR3-1333MHz RAM, Windows 8.1, and unreleased reference driver. MUN-3
2. The new 2014 AMD A-Series mainstream APU platform, codenamed “Beema,” is expected to deliver up to 104 percent better productivity performance per watt when compared to the previous generation “Kabini” platform. Testing conducted by AMD Performance Labs on optimized AMD reference systems. PC manufacturers may vary configuration yielding different results. PCMark 8 - Home score divided by TDP (W) is used to simulate productivity performance per watt; the Beema platform (15W) scored 2312 while the Kabini platform (25W) scored 1861. AMD "Larne" reference platform system used for both APUs. Kabini-based AMD A6-5200 quad-core APU with AMD Radeon™ HD 8400 Graphics, 2x2GB of DDR3-1600MHz RAM, Windows 8.1, 13.200.11.0 - 03-Sep-2013 driver. Pre-production engineering sample of “Beema” quad-core APU with next generation AMD Radeon graphics (model number TBD), 2x2GB DDR3-1600MHz RAM, Windows 8.1, and unreleased reference driver.BMN-3

 

[Arachnotronic]

These AMD slides are not trustworthy when it comes to perf. per watt. The only proper way to measure perf. per watt is to actually measure both perf. and power consumed for each specific task. To only measure perf. and apply it to a generic TDP spec is insanely silly given the broad dynamic range of power consumed on these devices.

Nothing AMD says about perf/watt is trustworthy. They misled (but never outright claimed) people into thinking Temash 4 core would be a 3.6W TDP part. You have absolutely no idea how many investors were misled into thinking AMD would have a part suitable for competition against Bay Trail by now.

 

[Olikan]

What about TSMC HPM process in play here?

Look at the QQ s600 to s800 transition. Same core, but the speed gains were incredible. Powerusage was way down also.

damm, that might be the answer

 

[inf64]

What I posted in my previous post looks to me pretty much like a claim by AMD... And it's on AMD's website. There is a disclaimer tho, a typical one done by all companies in all branches of business.

 

[Arachnotronic]

What about TSMC HPM process in play here?

Look at the QQ s600 to s800 transition. Same core, but the speed gains were incredible. Powerusage was way down also.

Uhh...S800 throttles really quickly btw.

 

[Ajay]

From here:

Well, it is a press release (i.e. 'PR'), so it's not like AMD is going to boast of their average numbers. Just like Intel and the rest they use the famous "up to" modifier, so this as good as it gets under certain specific settings. But I certainly do hope that these new parts are enough to get AMD some design wins. I just want to have some kind of competition in x86 if at all possible and want to see AMD survive to that end.

 

[Arachnotronic]

Well, it is a press release (i.e. 'PR'), so it's not like AMD is going to boast of their average numbers. Just like Intel and the rest they use the famous "up to" modifier, so this as good as it gets under certain specific settings. But I certainly do hope that these new parts are enough to get AMD some design wins. I just want to have some kind of competition in x86 if at all possible and want to see AMD survive to that end.

The only competition for Intel in the low power space are Qualcomm, NVIDIA, and Samsung. These are all much more powerful and dangerous competitors than Intel has seen in years and it is not at all clear that Intel will emerge from this victorious. Qualcomm's got better radios, NVIDIA's got better GPUs, and Samsung builds 40% of the world's mobile devices, so it controls which chip vendors live/die.

AMD is simply going to be a casualty of Intel's attempts to fight in the low power/low cost space...a market that Intel has traditionally let AMD keep when times were good.

 

[Ajay]

No, it really isn't. Win 8.1 is pretty great on a tablet, at least. Once the price point got down to a reasonable level (as it did with Bay Trail) people started buying them in droves - I still think the Asus T100 is the #2 top selling portable device on amazon, behind the chromebook.

But does BT deliver Core2 Duo like performance on these tablets? I still have and older laptop with a Core2 in it. It sits in a box because the display broke and I find it to be too slow even for web browsing (especially media rich web pages) to be worth fixing (IMHO). Now granted, that was under Vista SP2, so I imagine that Win8.1 would be better, but I don't know by how much.

 

[Arachnotronic]

But does BT deliver Core2 Duo like performance on these tablets? I still have and older laptop with a Core2 in it. It sits in a box because the display broke and I find it to be too slow even for web browsing (especially media rich web pages) to be worth fixing (IMHO). Now granted, that was under Vista SP2, so I imagine that Win8.1 would be better, but I don't know by how much.

Buy one. They're fast.

 

[rainy]

Nope, I just smelled something fishy on the slides and AMD historical data doesn't help.

Somehow I don't believe you - quite obviously you like/prefer all those black scenarios for AMD.

Btw, are you using nvidiot nickname on SA forum (very similar style)?

 

[Ajay]

Buy one. They're fast.

I don't need one right now, but I'll give one a try next time I'm at BB or elsewhere.

 

[Arachnotronic]

I don't need one right now, but I'll give one a try next time I'm at BB or elsewhere.

Definitely worth giving it a try. Have a Dell Venue 8 Pro myself...thing is incredibly quick. It has replaced my Haswell Ultrabook for...a surprising number of tasks.

 

[AtenRa]

Nothing AMD says about perf/watt is trustworthy. They misled (but never outright claimed) people into thinking Temash 4 core would be a 3.6W TDP part. You have absolutely no idea how many investors were misled into thinking AMD would have a part suitable for competition against Bay Trail by now.

I dont remember a single Roadmap slide claiming a Quad core Jaguar at 3.6W TDP. Every Roadmap slide was about 2-4 Jaguar Cores or up to 4 cores and FROM 3.6W TDP.
Investors should pay more attention to what they know do best(gamble on stock) than try to speculate about unreleased hardware.

June 2012

January 2013

 

[Arachnotronic]

I dont remember a single Roadmap slide claiming a Quad core Jaguar at 3.6W TDP. Every Roadmap slide was about 2-4 Jaguar Cores or up to 4 cores and FROM 3.6W TDP.
Investors should pay more attention to what they know do best(gamble on stock) than try to speculate about unreleased hardware.

June 2012

January 2013

Investors don't "gamble on stock" - they do intense research and - particularly in tech - try to predict share shifts and the impacts of new product cycles.

If AMD's Temash were in a bunch of small and 10.1" fanless tablets, you can bet AMD's shares would be much higher than they are now.

 

[AtenRa]

Investors don't "gamble on stock" - they do intense research and - particularly in tech - try to predict share shifts and the impacts of new product cycles.

If AMD's Temash were in a bunch of small and 10.1" fanless tablets, you can bet AMD's shares would be much higher than they are now.

Obviously, those who thought AMD will release a 3.6W TDP Jaguar quad core new shit about transistors, micro architecture and Litho manufacturing.
If you dont have hard data, you "gamble" trying to predict share shifts.

 

[mrmt]

If AMD's Temash were in a bunch of small and 10.1" fanless tablets, you can bet AMD's shares would be much higher than they are now.

I can't but contrast the sober tone of AMD management last year, where they carefully callibrated the expectations for Jaguar and this announcement today, where they seem to have gone wild on the misleading marketing.

This and the 1050 GFLOPS numbers from Kaveri should be a big red light, looks like they are back to their misleading marketing of before.

 

[inf64]

1050Gflops number was what AMD expected a year ago and these things are subject to change(GPU clock obviously is lower than what they expected). There was a Disclaimer you know, don't pretend you don't know about these things.

Can you tell us what is misleading about the Beema/Mullins numbers? You have TDP they list in the Footnote (not SDP), you have the numbers in benchmarks, you have the system specs the chips were in. What you don't have are model numbers and clock speeds which is understandable but does not take anything away from today's presentation. If AMD launches products that have those TDP numbers in real situations and reach those performance figures in those listed benchmarks, AMD has not "mislead" anyone but those who can't understand simple English language.

 

[mrmt]

1050Gflops number was what AMD expected a year ago and these things are subject to change(GPU clock obviously is lower than what they expected). There was a Disclaimer you know, don't pretend you don't know about these things.

Yes, I know the drill... But isn't impressive how frequent and how largely AMD performance forecasts are downgraded? Every time AMD launch a product we can read in the forums things like "It's a pitty that [AMD Product] could not clock as high as advertised", or "If [Foundry] could deliver [AMD Product] would be competitive against Intel". I think we can agree that AMD track record is bad to worse, and this 19% miss from Kaveri just add to the string, and this time is the new management. You can't hide behind a disclaimer with a track record like this.

Can you tell us what is misleading about the Beema/Mullins numbers? You have TDP they list in the Footnote (not SDP), you have the numbers in benchmarks, you have the system specs the chips were in.

Nobody ever got twice the performance per watt with higher clocks at the same node without a completely new architecture. Never. Niemals. Jamais. If someone is promissing this sort of improvements this someone is outright lying or playing with the metrics in order to get the desired numbers.

It's not like AMD is coming from a 5 years old core like Clovertrail to Silvermont, or that they are moving from clock speed uber alles to extremely efficient like Netburst to Core. If they are keeping much of Kabini, what is the secret sauce that allows them to do the same work with half the power? What's your explanation, or your guess? I can't find one, except that AMD is somehow bluffing, either outright lying or toying with numbers to make them look like ok.

 

[Arachnotronic]

Yes, I know the drill... But isn't impressive how frequent and how largely AMD performance forecasts are downgraded? Every time AMD launch a product we can read in the forums things like "It's a pitty that [AMD Product] could not clock as high as advertised", or "If [Foundry] could deliver [AMD Product] would be competitive against Intel". I think we can agree that AMD track record is bad to worse, and this 19% miss from Kaveri just add to the string, and this time is the new management. You can't hide behind a disclaimer with a track record like this.



Nobody ever got twice the performance per watt with higher clocks at the same node without a completely new architecture. Never. Niemals. Jamais. If someone is promissing this sort of improvements this someone is outright lying or playing with the metrics in order to get the desired numbers.

It's not like AMD is coming from a 5 years old core like Clovertrail to Silvermont, or that they are moving from clock speed uber alles to extremely efficient like Netburst to Core. If they are keeping much of Kabini, what is the secret sauce that allows them to do the same work with half the power? What's your explanation, or your guess? I can't find one, except that AMD is somehow bluffing, either outright lying or toying with numbers to make them look like ok.

Sharp, incisive thinking. Also don't forget that AMD is cutting R&D like crazy and lost many key technical people (i.e. most of the K8 team, the low power core team, etc.). AMD cannot work miracles.