Will Bobcat be the home run AMD is looking for?

[ModestGamer]

Yep, Cortex A15 is one interesting chip (from my layman's point of view).

But how will MS respond to these processor developments?

Will MS try to increase their x86 OS sales in more countries (using Bobcat) or do they think their ARM based OS stands a better chance for that type of expansion (even if Google is considerably ahead of them in development)?

MS will invest in whatever gives them the greatest cross platform portability while reducing costs.

 

[grimpr]

MS and ARM? A happy marriage.

http://www.macdailynews.com/index.php/weblog/comments/microsoft_deepens_deal_with_arm_holdings/

http://www.eetimes.com/electronics-news/4209600/Why-Microsoft-Windows-8-will-run-ARM

 

[cbn]

http://www.anandtech.com/show/4023/the-brazos-performance-preview-amd-e350-benchmarked/3

Anandtech E-350 preview updated with Core i3-330UM benchmarks.

The Core i3-330UM manages a 46% performance advantage over the E-350. Even in the ultraportable Arrandale ULV space at lower clocks, AMD still leaves a lot of CPU performance on the table. The advantage here will be cost. A single E-350 is less than 40% of the die area of a Core i3-330UM. You may not get the same CPU performance, but performance per mm^2 is much higher.

 

[Xpage]

I wonder why AMD doesn't use SOI for these processors, a 10% increase in cost, is worth it if they can get an equal amount of CPU performance, so they can catch up to CULV processors intel has in CPU performance, since they can pass the cost onto the consumer. I'd pay an extra $5-10 for 300 more mhz CPU speed

 

[IntelUser2000]

If we're going to look a year out, I think it's much more interesting to compare against ARM's upcoming A15, which is very architecturally impressive (3-wide, out of order loads, 128 bit FPU? Compare that to STARS!)

We'll see how that turns out. x86-isn't the most efficient ISA but the time when ISA mattered a lot is gone. I wouldn't be surprised if AMD stops ARM from advancing on the high-end, and Intel on the low end/low power(ironic I know).

I am also a bit surprised at Idontcare's comment about process tech. Pretty sure TSMC does have some advantages for its process that fits the products better than Intel at the same generation. Kinda like how IBM says their 28nm SoC SRAM bit cells have 4-10x lower leakage than 32nm SoC Intel's but Intel has absolute advantage at the higher end.

There is probably padding hiding some of the truth, but are they all lies? Intel's process with high drive currents fit well with their strategies which use high clocked CPUs, that don't need so much focus on leakage/power, while with TSMC products like GPUs and low power SoCs need lower drive currents and better density/power.

 

[Riek]

http://www.anandtech.com/show/4023/the-brazos-performance-preview-amd-e350-benchmarked/3

Anandtech E-350 preview updated with Core i3-330UM benchmarks.

well the gpu intensive taks are the reverse. Not sure if rendering, encoding are things i want to do, but good to know that if those are things you are looking for in a cheap/low power notebook you will have to go for the ULV brands (which was already known). But if you do something that will render on gpu or need gpu performance you are better of with zacate (or atom+ion for that matter). Would be interesting to see the power consumption of zacate and ULV.

Will SB ULV near the end of 2011 change this? current i3UM is dual core, SB would be 1core+HT with more focus on gfx.

 

[IntelUser2000]

Will SB ULV near the end of 2011 change this? current i3UM is dual core, SB would be 1core+HT with more focus on gfx.

IMO it won't. It will not only arrive a year later but ULV parts feature significantly lower GPU clocks.

With a 850MHz core clock and a 3GHz CPU it performs like the HD5450. If you bring that to 250MHz and pair it with a 1.2GHz CPU it'll be far below the HD5450. They are better off with Penryn based Celeron ULV part that's supposed to be arriving in early 2011. The G4x based chipset on that Celeron ULV will be just as capable playing HD videos and will be equally insufficient playing games anyway while being cheaper thanks to the older platform and process tech.

 

[Khato]

IMO it won't. It will not only arrive a year later but ULV parts feature significantly lower GPU clocks.

With a 850MHz core clock and a 3GHz CPU it performs like the HD5450. If you bring that to 250MHz and pair it with a 1.2GHz CPU it'll be far below the HD5450. They are better off with Penryn based Celeron ULV part that's supposed to be arriving in early 2011. The G4x based chipset on that Celeron ULV will be just as capable playing HD videos and will be equally insufficient playing games anyway while being cheaper thanks to the older platform and process tech.

The real unknown in that equation and what actually matters are the turbo clocks. The normal mobile parts are supposedly going to go from 650MHz base to 1300MHz turbo. The ULV may well be 250MHz or even lower base, but it could easily still be 1000MHz turbo. No way to know until it's announced.

As for how sandybridge graphics will actually perform... I'd wait to see numbers from an actual review instead of a preview with a large number of unknowns.

 

[IntelUser2000]

The real unknown in that equation and what actually matters are the turbo clocks. The normal mobile parts are supposedly going to go from 650MHz base to 1300MHz turbo. The ULV may well be 250MHz or even lower base, but it could easily still be 1000MHz turbo. No way to know until it's announced.

As for how sandybridge graphics will actually perform... I'd wait to see numbers from an actual review instead of a preview with a large number of unknowns.

You won't get 100% efficiency from Turbo, especially when the difference from base clocks are large like for graphics.

650/1300 isn't free. The real performance will be somewhere in between, probably trending closer to 1300 than 650. The low voltage chips ALL do this. They lower the base clocks but keep the Turbo mode clocks same.

Of course having Turbo is better than nothing. But I'm not expecting something amazing from ULV chips. CPU clock speeds have been stagnating for 6-7 years for the ULV parts. 1.1GHz 0.13u to 1.4GHz 45nm. The dual cores and Core i series chips do increase performance a lot but they sacrifice battery life lot too, especially the latter.

Let's look at graphics

915GMS @ 133/160MHz
945GU @ 250MHz
GME965 mini and sub note: 266MHz/320MHz
GS45 @ 320MHz(533MHz for HD video)
Core i7 ULV @ 166-500MHz

How do they perform? The low power versions of the graphics perform roughly equal to standard power version of the previous gen. I expect the trend will be similar with Sandy Bridge. If the "Zacate competitor" Sandy Bridge 1 core is a Celeron, it'll even cut on caches.

 

[Nemesis 1]

http://www.anandtech.com/show/4023/the-brazos-performance-preview-amd-e350-benchmarked/3

Anandtech E-350 preview updated with Core i3-330UM benchmarks.

Maybe whoever did this so called review of a chip thats is not out on any products being tested against present products. AT has a SB . Why not down clock that and test . By the time both are out. Intel will have SB and Oaktrail. This is only a disillusionment attempted by AMD to show a product that is not on the market on products its going against that it will not be up against. It will not have the energy efficiency Oaktrail. probably better than oaktrail in games remains to be seen . But SB culv not a chance in hell.

Its like when the SB reviews come out , No need to include AMD processors in that review . Just test against Intels last generation , There is no point in doing a amd review with SB none what so ever. It reminds me of the disillusionment of doing a GPU review with an i7 O/Ced at 3.8 to show off power of said gpu . Thats BS, Read the CPU section here or read the AT cpu reviews with AT giving the AMD cpu as the price performance leader . If thats true along with all the users that say they bought AMD performance for the buck cpus . These GPUs reviews are pure hype nothing more. Unless you have the same high dollar Intel setup. Sure it shows full gpu potentical.. But according to AT the AMD cpu is the price performance king . That being said. AT should back up their strong talk by using the CPUs they recommend to CPU buyers in their reviews . I see this as being 2 faced. Recommend this but show results with something else when testing gpu.
In todays cpu section much talk is about multi threaded . But in this zacata review that same multi threaded performance is downplayed. Single threaded performance is the focus . I can't help but hold head in hand and shake head. Its all about selling us the readers out. For a few bucks. If anyone here thinks they will get the same performance using a NV580 on amd cpu at stock speeds your crazy. Sure in some games that are gpu bound at high enough resolutions it might be close. Hell many games the differance is 50% between intel/AMD in games. Yet all say buy a knew GPU for better performance . True statement . But NOT TRUE if you let me pick the games tested . Sell outs thats what reviewers are now days some worse than others.

 

[jaydee]

This is only a disillusionment attempted by AMD to show a product that is not on the market on products its going against that it will not be up against. It will not have the energy efficiency Oaktrail. probably better than oaktrail in games remains to be seen . But SB culv not a chance in hell.

Why does this strawman arguement keep getting brought up? Is there anyone arguing that AMD is pitting Bobcat against SB ULV? Or are you bringing this up because you think that Intel is going to price SB ULV anywhere near Bobcat?

Something new is going on here. There's going to be high-end/low-power parts, low-end/low-power parts and stuff in-between. Bobcat is clearly low-end to mid-range, with price and efficiency advantages over SB. That's CPU-only. As far as GPU goes, it's a high-end part per it's power draw. You make it sound like there are no trade-off's whatsoever.

As far as Oaktrail, do you have anything other than pure speculation backing your words up?

 

[StrangerGuy]

Why does this strawman arguement keep getting brought up? Is there anyone arguing that AMD is pitting Bobcat against SB ULV? Or are you bringing this up because you think that Intel is going to price SB ULV anywhere near Bobcat?

Something new is going on here. There's going to be high-end/low-power parts, low-end/low-power parts and stuff in-between. Bobcat is clearly low-end to mid-range, with price and efficiency advantages over SB. That's CPU-only. As far as GPU goes, it's a high-end part per it's power draw. You make it sound like there are no trade-off's whatsoever.

As far as Oaktrail, do you have anything other than pure speculation backing your words up?

Shhh...Don't feed the trolls. We all know Intel will crush the under $500 market with SB ULVs with the asking price of Bobcats.

 

[Nemesis 1]

Why does this strawman arguement keep getting brought up? Is there anyone arguing that AMD is pitting Bobcat against SB ULV? Or are you bringing this up because you think that Intel is going to price SB ULV anywhere near Bobcat?

Something new is going on here. There's going to be high-end/low-power parts, low-end/low-power parts and stuff in-between. Bobcat is clearly low-end to mid-range, with price and efficiency advantages over SB. That's CPU-only. As far as GPU goes, it's a high-end part per it's power draw. You make it sound like there are no trade-off's whatsoever.

As far as Oaktrail, do you have anything other than pure speculation backing your words up?

On Oaktrail . Google . for backup . You act as if I made the name up . Zacata 8 watt will be up against oaktrail 5 watts.

 

[Nemesis 1]

Shhh...Don't feed the trolls. We all know Intel will crush the under $500 market with SB ULVs with the asking price of Bobcats.

Again I made it perfectly clear . That on the higher end SB culv wins easily . On the bottom end Oaktrail (5 watts)wins (Battery ) Zacata 8 watt loses here . Zacata 18 watt not a player here. Zacata has I small spot in the middle only . Zacata 8 watt won't make it into tablets.

 

[veri745]

On Oaktrail . Google . for backup . You act as if I made the name up . Zacata 8 watt will be up against oaktrail 5 watts.

Hopefully Oaktrail will have a chipset that doesn't use 9x the power of the CPU, unlike the previous Atom products.

I always thought that if they didn't pair the Atoms with that beast of a 945 chipset, I might actually buy one...

 

[Nemesis 1]

Oaktrail is going into tablets also . That should tell ya alot . Also It has NO pci-e its also 32nm. As far as I know. Its also SOC . so no pcie says alot. Intel is allowed by special rule to make these without pci=e until 2013. By than intel will have a 22nm version and will fab 100s of millions so as not to run out of supply after 2013 . I suspect a 6 year supply . Than intel can stop NV from playing on any intel product . And intel will . UNLESS NV pays bigtime . like billions . LOL.

 

[cbn]

Hopefully Oaktrail will have a chipset that doesn't use 9x the power of the CPU, unlike the previous Atom products.

I always thought that if they didn't pair the Atoms with that beast of a 945 chipset, I might actually buy one...

More information on Oaktrail here.

Moorestown can’t run Windows. It has no PCI bus, and without one you can’t run Windows. Oaktrail solves this problem.

Take Lincroft (Atom Z6xx series SoC) and pair it with a new PCH, codenamed Whitney Point and you get Oaktrail. Whitney Point is effectively Langwell plus SATA, HD Audio, HDMI and a bunch of legacy I/O (HPET, GPIO, RTC, DMA). Oaktrail is roughly the same footprint as Moorestown and although it’ll consume more power it’ll use less than Pine Trail.

 

[cbn]

There's going to be high-end/low-power parts, low-end/low-power parts and stuff in-between. Bobcat is clearly low-end to mid-range, with price and efficiency advantages over SB. That's CPU-only. As far as GPU goes, it's a high-end part per it's power draw.

That is my thinking also. The bobcat cores look like the super efficient part of the Zacate APU equation (until OpenCL matures). In fact, I just have to wonder how much of the 18 watt Zacate TDP is coming from the higher clocked cpu vs the higher clocked (500 Mhz) GPU?

I mean just look at the size of the GPU silicon area compared to the size of the bobcat cores.

Increasing a large amount of silicon's clock speeds by 78% (500 Mhz vs 280 Mhz) vs. increasing a small of amount of silicon's clock speeds by 60% (1.6 Ghz vs. 1.0 Ghz).

Anyone want to speculate on the TDP of an APU comprised of the Zacate clocked bobcat cores + Ontario clocked GPU?

 

[Eug]

I had a Samsung NF310 in my Amazon cart, ready to purchase pending the results of the Brazos/Zacate performance previews. I took it out of my cart. I will await Zacate-based solutions now.

These are the numbers that matter to me:
Support for 4gb DDR3.

Load power is good. Idle power is great. This is based on an ES as well, so I expect the power usage on production chips will drop. My netbook does not spend much time under load. It spends most of its time idling.

Looks like Zacate >>> Atom for those of us who produce rather than consume with our netbooks.

The bottom line is that single core Atom-based netbooks no longer provide 'good enough' computing for me. I played with a dual core Atom netbook a few days ago and it is 'good enough.' However, dual core Zacate-based netbooks will be better all-around, from performance to endurance to heat production. I see no real reason at this point to recommend Atom-based ultraportables to friends until Intel refreshes its Atom platform. I'm also eager to start building Zacate-based nettops for friends. It will be nice to deliver 'good enough' desktop computing for $300 or less including the display.

Perhaps, but as I said earlier, cost is going to be significant factor. Extrapolating from the AMD slides, it seems they will price it higher than Atom.

Unfortunately for AMD, Intel CULV is already starting to trickle down into the lower end market. I was waiting for Zacate for an 11.6" netbook but pulled the trigger last week for a dual-core 1.3 GHz Pentium SU4100 11.6" laptop with Intel 4500MHD instead. (See sig.) Why? Because it was less than $400, exactly the price point I was waiting for for Zacate, and I don't have to wait until 3-5 months from now. Also, I betcha when Zacate comes out, most 11.6" netbooks won't be <$400 in Canada. If they debut at $449 or $499, then it may be some months before they drop below $400.

I'll know soon enough if it was a good or bad decision to forego the AMD APU's GPU for Intel 4500MHD, but for those of us who produce on these things using standard productivity apps, gaming performance doesn't matter. The only thing I'm really concerned about is HD video playback since I like watching these on the plane when I'm on a business trip, but so far it's OK. What it cannot do is play native 1080p HD H.264 Quicktime files in QT Player completely smoothly, but I'm not surprised, because QT Player is not GPU accelerated. However, I'm sure Zacate wouldn't be able to either, given that its CPU is even slower than SU4100. Neither should have a problem with WMP and CCCP though.

As for that Zacate nettop... there are some good X2 3250e deals showing up. I'm tempted not to wait for that Zacate nettop either.

 

[ModestGamer]

Perhaps, but as I said earlier, cost is going to be significant factor. Extrapolating from the AMD slides, it seems they will price it higher than Atom.

But the boards will be cheaper as the require fewer parts. One slightly more exspensive chip vrsus two chips.

Ie cpu $5 gpu $5 for intel

zacate might be $7.50

So while one is higher the combined unit is cheape.r

At least that is my impression.

 

[IntelUser2000]

Oaktrail is going into tablets also . That should tell ya alot . Also It has NO pci-e its also 32nm. As far as I know. Its also SOC . so no pcie says alot.

Oak Trail: 45nm Lincroft chip + Whitney Point companion chip

http://news.cnet.com/8301-13924_3-20019410-64.html

Combined TDP of the CPU and the companion chip is supposed to be 4W.

Hopefully Oaktrail will have a chipset that doesn't use 9x the power of the CPU, unlike the previous Atom products.

I always thought that if they didn't pair the Atoms with that beast of a 945 chipset, I might actually buy one...

Nettop Atom: 4W
Netttop 945G: 22W

Netbook Atom: 2.5W
Netbook 945G: 4W

 

[CTho9305]

We'll see how that turns out. x86-isn't the most efficient ISA but the time when ISA mattered a lot is gone.

I think ISA still matters. Of course someone selling x86 chips will claim it doesn't... but if you actually look at the complexity involved in making x86 fast, I think it really adds up both in terms of design effort and silicon area/performance. Anyone claiming the penalty is contained in the decoder alone is misleading you. Sure, it's not technically x86 in the RISC/VLIW/other-CISC backend, but that backend is optimized for making x86 go fast and has to handle all the ridiculous things that x86 allows. Maybe if a company were willing to say "anything but AMD64 long-mode with all segment bases set to 0 and aligned memory accesses will be emulated with a 10X performance penalty" you could make a low-overhead x86...but I have yet to see any implementations do that.

 

[Eug]

I'll know soon enough if it was a good or bad decision to forego the AMD APU's GPU for Intel 4500MHD, but for those of us who produce on these things using standard productivity apps, gaming performance doesn't matter. The only thing I'm really concerned about is HD video playback since I like watching these on the plane when I'm on a business trip, but so far it's OK. What it cannot do is play native 1080p HD H.264 Quicktime files in QT Player completely smoothly, but I'm not surprised, because QT Player is not GPU accelerated. However, I'm sure Zacate wouldn't be able to either, given that its CPU is even slower than SU4100. Neither should have a problem with WMP and CCCP though.

Nope. Spoke too soon. I can't get this 4500MHD to play nice with CCCP and WMP.

No DVXA at all with WMP with my settings. I can't get it to work. With 1080p H.264 I get 100&#37; CPU utilization. I can get DVXA to work under Media Player Classic Home Cinema, but for some strange reason I sometimes still get stutters even when the CPU is only at 30%.

So I downloaded Splash, and it works perfectly. CPU usage is 25-35% with completely smooth 1080p H.264 playback. In fact, some really high bitrate stuff that stutters on Atom 330 / ION with WMP/CCCP plays perfectly in Splash on both platforms (with a bit lower CPU utilization on Atom 330/ION). Also, Splash's interface is about 50X better than MPC-HC, and about 5X better than WMP's.

On the other hand, for general OS usage SU4100 feels noticeably faster than Atom 330. Not surprising. Cinebench's CPU score is 0.50 on Atom 330, and it's 0.76 on SU4100 (over 50% faster). I'm guessing Zacate would be somewhere in between. On my MacBook Pro 2.26 GHz (OS X) it's 1.30.

So, with Zacate 1.5 GHz dual core it should feel noticeably faster for general OS usage than dual core Atom, albeit not as fast as SU4100, but the fast GPU should help it too compared to some of the Intel offerings, even SU4100 because support for Intel ClearVideo is hit and miss. That's assuming AMD provides good DVXA drivers for Zacate and CCCP/MPC-HC, WMP, Splash work well with it.

Intel would do well to encourage better support for its ClearVideo, not just with 4500MHD, but esp. with its new "flagship" Intel HD graphics. If it can get better support for ClearVideo, that would go a long way to help keep Zacate down.

 

[Nemesis 1]

More information on Oaktrail here.

Its all so confusing .

Intel has until 2013 to add an interface to the tablet-centric version of its Atom processor, as device makers are targeting the chip for Netbooks, the Federal Trade Commission said in a "modified settlement order" today.


"The agreement reached in August is now final but there is a minor modification to it," Intel spokesman Chuck Mulloy said today. "That modification is related to Oak Trail, which is for a tablet computer design. Some OEMs (device makers) have decided to design it into a Netbook-type device and under the definitions of the agreement, that made it a computer," Mulloy said.

The net effect is that the PCI Express interface will have to be added by 2013. "So, by the time we get to 2013 you could expect some kind of successor product that would be in compliance with the FTC consent decree," Mulloy added.

The FTC settlement with Intel in August ordered Intel to, among other things, maintain that key computer interface on its chips for at least six years "in a way that will not limit the performance of graphics-processing chips." The FTC said at that time that it hoped to provide a path for Nvidia and others to offer "complementary, and potentially competitive, products



Read more: http://news.cnet.com/8301-13924_3-20021504-64.html#ixzz15cNxau2j

 

[Nemesis 1]

More information on Oaktrail here.

Its all so confusing .

Intel has until 2013 to add an interface to the tablet-centric version of its Atom processor, as device makers are targeting the chip for Netbooks, the Federal Trade Commission said in a "modified settlement order" today.


"The agreement reached in August is now final but there is a minor modification to it," Intel spokesman Chuck Mulloy said today. "That modification is related to Oak Trail, which is for a tablet computer design. Some OEMs (device makers) have decided to design it into a Netbook-type device and under the definitions of the agreement, that made it a computer," Mulloy said.

The net effect is that the PCI Express interface will have to be added by 2013. "So, by the time we get to 2013 you could expect some kind of successor product that would be in compliance with the FTC consent decree," Mulloy added.

The FTC settlement with Intel in August ordered Intel to, among other things, maintain that key computer interface on its chips for at least six years "in a way that will not limit the performance of graphics-processing chips." The FTC said at that time that it hoped to provide a path for Nvidia and others to offer "complementary, and potentially competitive, products



Read more: http://news.cnet.com/8301-13924_3-20021504-64.html#ixzz15cNxau2j