Monday, August 22nd 2011

AMD Intros New C-Series and Two New E-Series APUs

AMD launched a trio of new E-series and C-series accelerated processing units (APUs), these include the E-450, E-300, and C-60. The three come in BGA packages, and are based on the Brazos platform, making use of the Bobcat high-efficiency x86 micro-architecture by AMD. All three are dual-core, and are built on the 40 nm process. The trio is topped by E-450 "Zacate". The processor component of the E-450 is clocked at 1.65 GHz (E-350 has 1.60 GHz). The Radeon HD 6320 GPU core is clocked at 508 MHz, and features TurboCore technology, that can drive it up to 600 MHz.

The E-300 lacks TurboCore on both its CPU and GPU components, its CPU component is clocked at 1.30 GHz, and Radeon HD 6310 GPU core at 488 MHz. The C-60 is AMD's ultra low power APU with its TDP at just 9W (E-450 and E-300 are 18W). The C-60 features TurboCore technology on both its CPU and GPU components, the CPU is clocked at 1.00 GHz, that can increase to 1.33 GHz, while the Radeon HD 6290 GPU core is clocked at 276 MHz, which can increase to 400 MHz. On all three APUs, the GPU component packs 80 stream processors, and is DirectX 11 compliant. All three feature a single-channel DDR3 memory controller supporting DDR3-1333 MHz memory. The integrated GPUs of all three support HDMI 1.4a and DisplayPort, apart from the usual DVI and VGA.
Add your own comment

14 Comments on AMD Intros New C-Series and Two New E-Series APUs

#1
jpierce55
That 9w deal should be great in tablets or netbooks!
Posted on Reply
#3
TheLaughingMan
by: Semi-Lobster
Still waiting for the E-450 :/
Its in the post, so what are you waiting for? Besides, it doesn't have a realistic improvement over the E-350. 0.05 MHz CPU speed bump and turbo mode added to the GPU will not improve its performance in any realistic manner. It should have been called the E-360 or E-355.

Let me know when there is an E-600 dual core at 1.7 GHz (2 GHz turbo) and a HD 6450D with 160 SIMPs clocked at 450 MHz (turbo 525 MHz) and Dual Channel Memory with a TDP of 18W and I will be excited.
Posted on Reply
#4
MxPhenom 216
Corsair Fanboy
im looking at getting a A-series 3500m APU laptop for school
Posted on Reply
#5
TheLaughingMan
by: nvidiaintelftw
im looking at getting a A-series 3500m APU laptop for school
I look forward to the irony of your name and owning an AMD laptop for school.
Posted on Reply
#6
Thefumigator
The MSI CR430 featuring the E450 will be my next laptop for sure.
I had a CR400 in my hands and I loved it, but my worries only lies on its 14" screen, which is smaller than any 14" screen, because its wide (16:9)
Posted on Reply
#7
RejZoR
Do want the C-60 in a netbook :D
Posted on Reply
#8
Semi-Lobster
by: TheLaughingMan
Its in the post, so what are you waiting for? Besides, it doesn't have a realistic improvement over the E-350. 0.05 MHz CPU speed bump and turbo mode added to the GPU will not improve its performance in any realistic manner. It should have been called the E-360 or E-355.

Let me know when there is an E-600 dual core at 1.7 GHz (2 GHz turbo) and a HD 6450D with 160 SIMPs clocked at 450 MHz (turbo 525 MHz) and Dual Channel Memory with a TDP of 18W and I will be excited.
My bad missed it :p

But the biggest improvement is memory going from DDR3-1066 to DDR3-1333 which helps noticebly with the GPU in the APU
Posted on Reply
#9
Completely Bonkers
When will there be architecture or cache enhancements to gain 40%+ improvements in performance? Releasing new versions with 5% gains is boring. Atom was great 3 years ago when launched, but has FAILED to follow MOORES LAW. Atom today should be 3x-4x faster than the original if it followed Moore's Law. Looks like AMD is in the slow lane too. (Although they do have a much better baseline starting point - much better than Atom)
Posted on Reply
#10
TheLaughingMan
by: Semi-Lobster
My bad missed it :p

But the biggest improvement is memory going from DDR3-1066 to DDR3-1333 which helps noticebly with the GPU in the APU
That will get you about 13% overall performance improvement. Nice, but they could get another 40% to 50% buy just going to dual channeled memory.

by: Completely Bonkers
When will there be architecture or cache enhancements to gain 40%+ improvements in performance? Releasing new versions with 5% gains is boring. Atom was great 3 years ago when launched, but has FAILED to follow MOORES LAW. Atom today should be 3x-4x faster than the original if it followed Moore's Law. Looks like AMD is in the slow lane too. (Although they do have a much better baseline starting point - much better than Atom)
Well AMD is not in the slow lane. Getting the power consumption down at 40nm is next to impossible at this point. They have pushed the K10.5 architecture to its limit with these chips. The only thing that can and well get them any real improvement is the Trinity chips (APU based on Bulldozer at 32nm).

With the CPU being 20% smaller at 32nm, plus area gained from the shared resources in the core design, should net enough room to get both dual channel memory and more GPU SIMPs on the die. It will also generate less heat bring TDP down, and use less power for better battery life. Or they can forgo TDP and power improvements and bring the clock speed up to the 2.4 GHz range. Either way, we get a new series of chips from APU that should prove to be very, very nice come Trinity release.
Posted on Reply
#11
Thefumigator
by: Completely Bonkers
Although they do have a much better baseline starting point - much better than Atom
Specially in the GPU area. Well it takes like half of the chip area, and instead of getting a CPU with a GPU, to me its more like you get a GPU featuring a CPU.
Posted on Reply
#12
mastrdrver
by: Completely Bonkers
When will there be architecture or cache enhancements to gain 40%+ improvements in performance? Releasing new versions with 5% gains is boring. Atom was great 3 years ago when launched, but has FAILED to follow MOORES LAW. Atom today should be 3x-4x faster than the original if it followed Moore's Law. Looks like AMD is in the slow lane too. (Although they do have a much better baseline starting point - much better than Atom)
Atom fails in PCs because it can not do OoO operations. It's main focus is in tablets and cell phones which OoO shouldn't matter and would cost power in a market that is power sensitive.

As far as the AMD stuff, your looking for Wichita which is 4 Bobcat cores and FCH on a chip or know as SOC. All at 28nm at 20w verse 18w for Zacate (E350/450) plus what ever the FCH needs (5w?). link

Should be here by beginning of next year or 1 year after Zacate/Ontario launch which is really good when you consider that they went two shrinks in one year from 40nm to 28nm (skipping 32nm).
Posted on Reply
#13
Nesters
Not really two shrinks... CPUs go like this 45nm->32nm->22nm, GPUs 40->28->20(?)

And AMD SoC are following gpu shrinks. Nothing skipped there.
Posted on Reply
#14
mastrdrver
To be technical (for TSMC) the 28nm process is a full node shrink from 40nm. So going to 28nm is only skipping the half node shrink of 32nm.

The only reason the AMD SoC is following GPU shrinks is because they are using the same fab, TSMC.

Also the CPUs and GPUs can shrink the same way, its just dependent on where they are made. for example, if AMD's GPUs were being made at GF, then they could have a 32nm GPU.
Posted on Reply
Add your own comment