• We’re currently investigating an issue related to the forum theme and styling that is impacting page layout and visual formatting. The problem has been identified, and we are actively working on a resolution. There is no impact to user data or functionality, this is strictly a front-end display issue. We’ll post an update once the fix has been deployed. Thanks for your patience while we get this sorted.

AMD Announcement: ARMv8 Opterons In 2014

Page 3 - Seeking answers? Join the AnandTech community: where nearly half-a-million members share solutions and discuss the latest tech.
yes, it boggles my mind how if cash is changing hands, AMD is paying GF and not the other way around.

Still, if GF goes away where would AMD go? I'm sceptical TSMC has enough capacity to serve AMD completely.
It took TSMC 90days to double the number of 28nm wafers they are shipping on a daily basis.

If you want capacity you just have to pay for it. That is part and parcel to the job of effectively managing your fabless business.

The whining comes in when business managers do not internally manage their forecasted product demand well enough to secure capacity in advance (see qualcomm)...so then they complain they are capacity constrained as if they played no role in creating the situation during their own contractual negotions with the foundry in question.

Capacity exists, businesses just don't want to pay the price premium necessary to have the capacity allocated to them versus the businesses it is currently allocated towards. (in business, everything is negotiable, but everyone wants everything for free, naturally)
 
The part of AMD that once upon a time knew how to successfully manage a fabless company (ATI) would have known all this, the part of AMD that appears to have made critical decisions at AMD leading up to 32nm and 28nm appear to have underestimated the criticality of this portion of their job as managers of a then fabless company and it shows in their misexecution and financial situation today. (not Rory's fault, he inherited this mess)

IDC, I don't think that AMD management could do anything about some of its recent foundry woes.

When AMD was almost bankrupt and had to spin off the fabs, it was clear to everyone in the market that the agreement would bound AMD for a lot of time to Globalfoundries as they hadn't any other customer and ATIC would need a lot of time to get new clients. While the IT community saw this as a non-issue, as if ATIC would keep funding the former AMD fab business forever regardless of make money or not, the financial community was always very worried about the deal, because everyone on the market knows that the Arabs are more patient, but they hate to lose money. GLF didn't solve the issue of the unprofitable foundry, just took it out of AMD balance sheet.

Four years later and some billions in payments and new agreements, we can now see that ATIC forced commitments on exclusivity, pricing and volumes, and how ATIC played hard on that. I would not say that they became truly fabless as they were so entangled to GLF that it wasn't really a spin off, but an outsourcing. And with this, even if AMD switched that bunch of incompetents for a full fledged management team I don't think that they could have done anything to improve the GLF situation. It was beyond their control.

AMD past also came back to haunt them. AMD didn't develop processes in-house, they went for others for this, Motorola first, then IBM. AMD simply lacked the expertise in-house to develop a process of their own, and it was this team after the spin off that went ahead to implement an IBM process that IBM didn't have implemented and develop a new 28nm process in HKMG bulk, something they hadn't the smaller experience with. So delays and woes should have been expected.

When they spun off the fabs they became a fabless company in the balance

As for Krishna and Wichita, I tend to agree with you. They could have managed the situation better with GLF, but I wonder if the cuts and leaves that happened after RR had something to do with their cancelling.
 
Brazos sucks in a Big Way. AMD is pushing it in a "full size" notebook market where it sells for the same price, as SB-based Pentiums, which are not only by order of magnitude faster than Brazos on CPU side, but also faster on GPU side. Speaking about netbooks - you can already buy cheap Celeron based netbooks (Acer AO756) which will eat Brazos for breakfast.

Brazos E350 was launched in Q1 2011, Intels 32nm Celerons M was launched in Q1 2012, that is one year later. Not to mention that Brazos 2 E-1800(TSMC 40nm) is far superior in CPU AND iGPU than 32nm ATOM.
 
Brazos E350 was launched in Q1 2011, Intels 32nm Celerons M was launched in Q1 2012, that is one year later. Not to mention that Brazos 2 E-1800(TSMC 40nm) is far superior in CPU AND iGPU than 32nm ATOM.

Atom has different market than E-1800. You wont see atom in notebooks with screen size larger then 10.1" as well as you wont see E-1800 in notebooks with screen smaller then 11.6". Also average atom netbook has much lower price.
 
Brazos E350 was launched in Q1 2011, Intels 32nm Celerons M was launched in Q1 2012, that is one year later. Not to mention that Brazos 2 E-1800(TSMC 40nm) is far superior in CPU AND iGPU than 32nm ATOM.

Do you have Benchmarks cause I want to see them . Intel only recently released the Z2580. For smartphones in a SoC design . I havent seen any clover trail reviews and the power draw is way lower on clover trail. If good graphics is all thats needed in this market. Than remove the cpu all together. You want to compare AMD to an old Intel tech all the time . Haswell below 10 watts will own that highend segment . ARM is for real . They are going to hurt AMD 20x worse than Intel.
 
I have never heard of an x86 exclusivity agreement with GF at any process node. I think there exists a secret agreement that grants x86 exclusivity in perpetuity, in which case expect to see more payments for unheard of 20 nm and 14 nm exclusivity agreements going forward.
 
Do you have Benchmarks cause I want to see them .

Intel new 32nm ATOM 2500 and 2800 vs AMD E350.

http://www.xbitlabs.com/articles/cpu/display/atom-cedartrtail_2.html#sect0

They cant even beat the first gen AMD Brazos 40nm old E350. Not to mention 32nm ATOMs iGPU sucks big time. They dont have even drivers for 64bit windows.

You want to compare AMD to an old Intel tech all the time . Haswell below 10 watts will own that highend segment . ARM is for real . They are going to hurt AMD 20x worse than Intel.

Haswell haven't released yet. As of now, AMDs Brazos 2 is a better product than Intels 32nm ATOM.
 
They cant even beat the first gen AMD Brazos 40nm old E350. Not to mention 32nm ATOMs iGPU sucks big time. They dont have even drivers for 64bit windows.

As if you would need 64 bit windows with more than 4GB on a tablet. Do you know what Brazos problem is, and why the delay for 28nm hurt them badly?

When Brazos launched he could fight with Atom for the same markets AND go above because the competition could not go as low as Brazos in the power ladder. With 32nm Atom went to power levels that Brazos cannot match, and Celeron checked them from upside, as Celeron performance is far better than Brazos.

So what changed is that in the beginning we had a disruptive product with a nice market latitude, and now we have a niche product. Brazos is king on a certain market bracket, but this market bracket is rather small. Need low power? Brazos is a non-starter. Need a lot of performance? Brazos is a non-starter too. Need a delicate compromise between the two? Brazos might be your chip.

28nm was supposed to bring some of this latitude back, but while 28nm in 2012 would be great, 28nm in 2013 will not be so great. Not only Celerons will improve with IVB, further diminishing Kabini potential market upside, we can expect Silvermont a few months later. AMD will lose this one too, but this time they lost because of themselves.

Haswell haven't released yet. As of now, AMDs Brazos 2 is a better product than Intels 32nm ATOM.

Is it? I have a phone with Atom here. Can I have a phone powered by AMD chips?
 
Last edited:
28nm was supposed to bring some of this latitude back, but while 28nm in 2012 would be great, 28nm in 2013 will not be so great. Not only Celerons will improve with IVB, further diminishing Kabini potential market upside, we can expect Silvermont a few months later. AMD will lose this one too, but this time they lost because of themselves.

If 40nm Brazos 2 is faster than 32nm ATOM, then just imagine what a quad core 28nm Kabini will do. No need to speculate about next years products, today even at 40nm Brazos 2 is better than 32nm ATOM(expect power consumption).


Is it? I have a phone with Atom here. Can I have a phone powered by AMD chips?

I dont remember me speaking about phones. Brazos was never meant for the phone market.
 
If 40nm Brazos 2 is faster than 32nm ATOM, then just imagine what a quad core 28nm Kabini will do. No need to speculate about next years products, today even at 40nm Brazos 2 is better than 32nm ATOM(expect power consumption).

I can only wonder why you don't take this unidimensional approach (speed is everything!) when you are talking about Trinity or Bulldozer. Why would you recommend any AMD chips to someone if Intel alternatives are always faster?

Atom was developed for netbooks at 45nm and every single optimization it received in its four years lifespan was to reduce power consumption in order to bring it to phones and other mobile devices. It is obvious that another chip developed without those design constraints was supposed to be faster but use more power, and guess what, this fits Brazos nicely. You can bet that it will fit Kabini too.

Have a look at the big picture. Can Brazos go down with Atom at the power ladder? And what if you are a bit less power constrained, can it beat an IVB Celeron, or even SNB Celeron? Brazos is only good when saw from its small niche as you are doing now, once you go out of it, the chip isn't interesting at all.

I dont remember me speaking about phones. Brazos was never meant for the phone market.

Atom was designed for this market, or at least was optimized to it, and that's why you cannot make direct comparisons between the two chips. Different markets, different aims, different designs trade offs.
 
I can only wonder why you don't take this unidimensional approach (speed is everything!) when you are talking about Trinity or Bulldozer. Why would you recommend any AMD chips to someone if Intel alternatives are always faster?

Because at the same price AMD CPUs are faster. FX4xxx faster than Dual core Intel, FX6xxx faster than Dual Core HT Intel, FX8xxx faster than Quad Core Intel. Also AMD APUs have same CPU Performance and superior iGPU.

Atom was developed for netbooks at 45nm and every single optimization it received in its four years lifespan was to reduce power consumption in order to bring it to phones and other mobile devices. It is obvious that another chip developed without those design constraints was supposed to be faster but use more power, and guess what, this fits Brazos nicely. You can bet that it will fit Kabini too.

32nm ATOM is produced for Desktop and mobile as well, not only for Phones. I own Intel D2700DC and DN2800MT and the iGPU performance is laughable for desktop use. They cant even play 1080p YouTube videos 😵

Have a look at the big picture. Can Brazos go down with Atom at the power ladder? And what if you are a bit less power constrained, can it beat an IVB Celeron, or even SNB Celeron? Brazos is only good when saw from its small niche as you are doing now, once you go out of it, the chip isn't interesting at all.

ATOM has less power consumption only because of the smaller node. At 28nm Kabini will be not only faster but have the same power consumption as 32nm ATOM.
 
Have a look at the big picture. Can Brazos go down with Atom at the power ladder? And what if you are a bit less power constrained, can it beat an IVB Celeron, or even SNB Celeron? Brazos is only good when saw from its small niche as you are doing now, once you go out of it, the chip isn't interesting at all.

If you think if fair to compare IVB Celeron to Brazos, then I think it's fair to compare Trinity to Atom. Brazos is clearly superior than Atom, and the fact that you need to bring Celeron into the picture just stinks of fanboyism.

Why is it so hard for you to admit what everyone already knows? Say it after me, Brazos is better than Atom...
 
If you think if fair to compare IVB Celeron to Brazos, then I think it's fair to compare Trinity to Atom. Brazos is clearly superior than Atom, and the fact that you need to bring Celeron into the picture just stinks of fanboyism.

Why is it so hard for you to admit what everyone already knows? Say it after me, Brazos is better than Atom...

You're kidding, right? Here is netbook Acer Aspire One AO756 for $279 with Celeron 877. Try to find something E-350/450 based at this price range. The closest competitor is C-50/60 which has pathetic cpu performance even when compared to Atom (let alone Celeron).
 
Last edited:
Because at the same price AMD CPUs are faster. FX4xxx faster than Dual core Intel, FX6xxx faster than Dual Core HT Intel, FX8xxx faster than Quad Core Intel. Also AMD APUs have same CPU Performance and superior iGPU.

As the financial guys say, you get what you pay for. AMD chips have superior performance for a few price points but they also have inferior power consumption, which means more heat, bigger cases, more noise, bigger electricity bill....

32nm ATOM is produced for Desktop and mobile as well, not only for Phones. I own Intel D2700DC and DN2800MT and the iGPU performance is laughable for desktop use. They cant even play 1080p YouTube videos 😵

Used in =! designed to

You are complaining that a chip optimized to be on a phone isn't strong enough to your desktop. It would be the same thing of complaining about low fps on playing Crysis on Brazos. You have to use the right chip for the right workload.

BTW, one of the better decisions that AMD management took in the last few years was exactly position Brazos in the huge gap between Atom and Core instead of going straigh to Atom, as their pavlovian behavior pushes them to do.

ATOM has less power consumption only because of the smaller node. At 28nm Kabini will be not only faster but have the same power consumption as 32nm ATOM.

Don't hold your breath. Kabini will have a lot of added complexity and I understand that they are aiming at premium tablets, not low power or ultra-high battery life. They are also aiming at ultra-dense servers too. All this points to an environment much less power-constrained than a phone. So don't expect Brazos to be in the same league as Atom in power consumption.
 
If you think if fair to compare IVB Celeron to Brazos, then I think it's fair to compare Trinity to Atom.

You should read my posts better. I'm saying all along that you cannot compare Atom and Brazos directly, much less Brazos and Celeron, as they are aimed at different markets and different power envelopes.

But if you want to compare a 246mm^2 chip with a 64-100mm^2 SoC, be by guest.
 
2014.


That date seems to make it all completely meaningless - assuming they survive.
They wanna make 28nm ARM\HSA procs against broadwell\atom SoC's?

Is that even gonna.. work?
 
You're kidding, right? Here is netbook Acer Aspire One AO756 for $279 with Celeron 877. Try to find something E-350/450 based at this price range. The closest competitor is C-50/60 which has pathetic cpu performance even when compared to Atom (let alone Celeron).

Why would I do that when the discussion was about Atom vs Brazos and I said that bringing Celeron into the equation stinks of desperation?

However, when I checked the link you provided, I noticed the Item below was an E-300 for the same price, but with 15.6" screen compared to the 11.6" on the Acer.
 
Last edited:
You should read my posts better. I'm saying all along that you cannot compare Atom and Brazos directly, much less Brazos and Celeron, as they are aimed at different markets and different power envelopes.

But if you want to compare a 246mm^2 chip with a 64-100mm^2 SoC, be by guest.

I know you are saying that. I'm saying you're wrong, that you can and must compare Brazos to Atom as they're both netbook chips.

Celeron is a laptop/desktop part and the only reason Intel is sticking it in netbooks is because Atoms come up short against Brazos.

You can claim Atom is better than Brazos if you want, but you wont find any of the reviews coming to the same conclusion as you.
 
I know you are saying that. I'm saying you're wrong, that you can and must compare Brazos to Atom as they're both netbook chips.

Celeron is a laptop/desktop part and the only reason Intel is sticking it in netbooks is because Atoms come up short against Brazos.

You can claim Atom is better than Brazos if you want, but you wont find any of the reviews coming to the same conclusion as you.


+1

Additionally both chips are getting long in the teeth. The points of key interests are the new atom and jaguar, when released. The impact of this generation on the market is known, but what is next?
 
Back
Top