Bookmark and Share


Advanced Micro Devices on Tuesday finally began to sell its new-generation A-series “Kaveri” accelerated processing units. The new chips boast with all-new Steamroller x86 cores, heterogeneous system architecture (HSA), new-generation graphics engine, improved media processing units as well as a set of AMD’s other latest technologies.

New Chips Detailed

General specifications of AMD A-series “Kaveri” have been known for a long time: the new APUs feature two dual-core x86 Steamroller modules (two integer units, one FPU with two 128-bit FMACs, one MMX unit per module) for a total of four cores; Radeon R7 graphics processing unit with up to 512 stream processors (eight GCN compute units), dual-channel DDR3 memory controller with hUMA technology, various special-purpose accelerators, including AMD TrueAudio DSPs [digital sound processors] second-generation video coding engine (VCE) and fourth-generation unified video decoder (UVD) units that enhance quality of playback and add new video encoding formats.

At present AMD sells A10-7850K and A10-7700K accelerated processing units with unlocked multiplier and 95W thermal design power. Later in the quarter the company will ship A8-7600 models with reduced power consumption. The most advanced AMD APU (model A10-7850K) costs $173 in 1000-unit quantities, whereas its younger brother (A10-7700K) is priced at $152. The A10-7850K and A10-7700K APUs will be bundled with EA’s Battlefield 4 video game.

“AMD maintains our technology leadership with the 2014 AMD A-Series APUs, a revolutionary next generation APU that marks a new era of computing. With world-class graphics and compute technology on a single chip, the AMD A-Series APU is an effective and efficient solution for our customers and enable industry-leading computing experiences,” said Bernd Lienhard, corporate vice president and general manager of client business unit at AMD.  

Heterogeneous System Architecture

The new APUs support heterogeneous queuing that defines how processors interact equally and enables uniform visibility into the entire memory space by both GPU and CPU. Kaveri’s hUMA [heterogeneous unified memory architecture] shared memory controller fetches data to both x86 and GCN processing units as well as acts like a dispatcher/hub.

Thanks to HUMA and heterogeneous queuing, applications can easily run processes on different types of cores, such as x86 CPU or GCN GPU compute units. As a result, AMD claims that its new Kaveri features twelve compute units in total. While the claim is technically correct, it should be kept in mind that not all applications can offload computing to stream processors of graphics adapters, hence, a number of programs will take advantage of only four x86 processors.

The new series of APUs support all the technologies that AMD introduced along with its latest family of graphics processors – which feature improved second-gen GCN architecture – including TrueAudio as well as DirectX 11.2 and Mantle application programming interfaces.

APU Optimized Manufacturing Process

For previous generations of its high-performance APUs (i.e., Llano, Trinity and Richland), AMD chose 32nm silicon-on-insulator process technology that was developed by the company itself and GlobalFoundries with high-performance microprocessor units in mind. Those process technologies allowed AMD to ensure high clock-speeds for its x86 cores, such as Bulldozer and Piledriver. Unfortunately, that manufacturing technology was not exactly tailored for high transistor density of graphics processing units. As a consequence, AMD had to make certain compromises, which did not allow it to integrate all the hardware it wanted.

With Kaveri, AMD chose to use GlobalFoundries’ 28SHP (28nm super high performance) process technology that was designed with various types of chips in mind. Improved transistor density and thinner elements of 28nm process tech allowed the company to integrate whopping 2.41 billion of transistors into a 245mm2 Kaveri die, which is 85% higher transistor count compared to Trinity/Richland design (246mm2, 1.303 billion transistors, 32nm SOI). Now the company has four Steamroller cores, hUMA [heterogeneous unified memory architecture] memory controller, Radeon graphics processing unit with 512 stream processors (AMD calls them eight GPU compute units) and a lot of various special-purpose hardware inside its flagship APU.

The high-density “universal” process technology has its drawbacks too: it lacks thick metal stacks that make possible to run microprocessors at high frequencies. As a result, the company sacrificed extreme clock-rates for x86 general-purpose processing cores in favour or forward-looking APU design with a strong focus on heterogeneous computing.

While overclockers and performance enthusiasts may not appreciate AMD’s choice, the company claims in its documents that thanks to more advanced Steamroller micro-architecture (with up to 20% higher instruction per clock [IPC] performance), the Kaveri will not be slower than currently-available Richland APUs and will even beat them in many cases.

A Look Ahead

The AMD A-Series APU processor-in-a-box (PIBs) for the AMD A10-7850K and AMD A10-7700K, which started shipping in Q4 2013, are available starting today. The AMD A8-7600 will be shipping in Q1 2014.

AMD believes that thanks to scalability of its Kaveri design (which is supposed to be available in 15W – 95W thermal design power envelopes), it will be able to successfully address various personal computers, including desktops, notebooks and servers, with its new A-series accelerated processing units in 2014.

Tags: AMD, Kaveri, Radeon, GCN, Steamroller, Fusion, 28nm, Globalfoundries


Comments currently: 62
Discussion started: 01/15/14 01:13:48 AM
Latest comment: 01/17/14 06:51:22 PM
Expand all threads | Collapse all threads


I Just Saw Good Benchmark from Anandtech. AMD Kaveri offers Good Performance in GPU side. Worth Upgrading or Replace for AMD Llano
9 5 [Posted by: xentar  | Date: 01/15/14 01:13:48 AM]

This is the most important processor since AMD64. And this is why -

It kicks Intel i7 to the kerb in OpenCL at half the TDP. Now imagine it at 14nm with 24 compute cores and further evolution of HSA features and software. Powerful stuff.

Now a quick comparison between:

Phenom II X6 1100T 6xCPU @ 3.6GHz 125W TDP
Kaveri A10-A7850K 4xCPU + 8xGPU HSA @ 3.7GHz 95W TDP

Using Phoronix Test Suite 4.8.2

x264 (frames per second) 106.07 vs 94.45 (non OpenCL accelerated)
OpenSSL (RSA 4096-bit Performance - signs per second) 106.87 vs 350.60
C-Ray (seconds to complete) 57.33 vs 37.00
John The Ripper 1.8.0 Blowfish 990 vs 3631
John The Ripper 1.8.0 MD5 36124 vs 61971
John The Ripper 1.8.0 Traditonal DES 5197667 vs 9059667
Himeno Benchmark (linear solver test - MFLOPS) 1388.15 vs 958.10

These early benchmarks show Kaveri's massive advantage in parallelism whilst also showing the 4 core CPU as being not much weaker than the 6 core Phenom in linear calculations.
12 7 [Posted by: linuxlowdown  | Date: 01/15/14 01:17:46 AM]
- collapse thread

GCN really work well to AMD Kaveri. AMD APU winner for GPU side. My time to replace the old AMD Llano
9 4 [Posted by: xentar  | Date: 01/15/14 01:39:35 AM]
Ola, down there! Long time no see.

Couldn't resist but to pour some anger here.

It is obvious how some sites that claim to be an authority in tech analysis, e.g. AnandTroll fail to see the obvious truth. Let's leave the lack of objectiveness in their "reviews" for another topic.

You pointed it very well with your link, yet they are still like "meh, get an i3 it is better, blah blah". They can't see the truth behind Kaveri (can't they really, or ?) and that my friends is what we've seen times and times before. I'm talking about x64, dual-core, 1GHz core clock, etc. Now we have HSA, or hUMA to be precise.

AlMightyD plans are not to compete clock-per-clock with overpriced and overhyped crap. The idea is to bring the game to a whole new level. We all can see the huge improvements in performance once hUMA and Mantle are well utilized. IPC and GHz now mean nothing. NOTHING!

Once the major SW vendors adopt it the computing as we know it will change forever. Imagine when Java get hUMA acceleration! And Mantle just makes the CPU obsolete.

And the real deal here is that everything is open to everyone willing to join. No proprietary crap, the likes of which we are seeing every day from camp Green.

Let everyone in the game with clear rules. Once everyone is on the same wagon, the only way to stay on it is to be the best. That is what a true competition is. Not some bribed optimizations and vaporware marketing.

Once AlMightyD pull the whole thing out, the computing landscape won't have anything to do with what it is today. It's inevitable.

Prepare your anuses, blue/green fans, they are soon to become red, whether you like it or not. Soon, you'll all use technologies you've never ever dreamed of, whether you like it or not!

8 5 [Posted by: zlobster  | Date: 01/15/14 05:58:33 AM]
I don't know how we're going to answer this new outstanding AMD technology. Our Iris Pros just don't cut it. I guess we better go down a node again to compete. Jeez this is starting to cost us in the current PC slump.
6 5 [Posted by: Brian Krzanich  | Date: 01/15/14 06:24:48 AM]
show the post
0 4 [Posted by: zlobster  | Date: 01/15/14 10:25:05 AM]
But if we pack more transistors into our shit GPU designs, then surely we're gonna win! You can't beat the brute force of transistor count. This is my plan. But the only snag is that it's costing us lots of dollars to keep nodes ahead of the industry - and we can't maintain it for much longer. I'm really not sure how we're going to tackle it at this point. It'll take us at least 3 years to engineer our own HSA type architecture. AMD is going to have the same type of lead they had with AMD64, except this time Rory Read is competently managing the company. So we'll probably abuse our monopoly position with our tried and true tactics: -

1) Pay OEMs not to use AMD designs (carrot)
2) Blackmail OEMs by delaying or withdrawing new processors if they use AMD designs (stick)
3) Give discounts in the channel for singular promotion of Intel processors
4) Employ professional trolls to spread FUD about AMD
5) Pay moles inside AMD
6) Head hunt key AMD engineers to disrupt operations, even if we don't need their talent
7) Take Republican senators to lunch more often
3 0 [Posted by: Brian Krzanich  | Date: 01/16/14 05:18:26 AM]
For once I have to admit this is quite a humorous comment.

I'm just not sure whether there are tons of sarcasm in your post or you just managed to miraculously describe the InHell tactics so far.

On another topic, I believe that shrinking the node you are also forced to use lower clock freqs. Let alone the increased heat dissipation generated by all the extra transistors (if you have ever seen a transistor that is ;P). It is not just a win-win. Oh, and one more thing, it is not how much transistors you have, it is how you use them.

Now go have some quality time with some of your PBAs. You earned it once again with your guerrilla marketing and FUD.
2 1 [Posted by: zlobster  | Date: 01/16/14 07:16:12 AM]
show the post
2 5 [Posted by: amdzorz  | Date: 01/15/14 11:43:47 AM]
porn movies in your case zorz
2 2 [Posted by: MHudon  | Date: 01/15/14 02:03:43 PM]
I can tell by his persuasion that he is moved by homoerotica. This troll will be silenced for any blatant pro Intel/Nvidia comments without sensible reasoning in the future.
1 2 [Posted by: fanboyslayer  | Date: 01/16/14 01:41:48 AM]
Those 8 GPU "cores" can actually context switch between up to 8 separate [threads]* {queues} per ACE, for a total, if I am correct {of 64 queues, of varing lengths, of Tasks, so that is 64 gueues with a varying amount of tasks to be fed into GPU's 8 compute "cores", 8 queues per GPU "core" }

[ of 64 threads of simultaneous CPU like processing]*

* Removed words in brackets[] replaced with words in curly braces{} edited 1/16/2014 4:05 PM EST.

, in addition to the 4 ARM cores and their SIMD instructions, and I'll have to agree, linuxlowdown, with you that this is of the AMD x86 64 bit ISA innovation scale, or better, quality of new innovation, on AMD's part! The Impliations for Ray Tracing, with respect to this, point to a great ability, once the software ecosystem catches up with the Kaveri hardware!
3 1 [Posted by: BigChiefRunAmok  | Date: 01/15/14 02:31:49 PM]

show the post
2 9 [Posted by: Tristan  | Date: 01/15/14 08:25:26 AM]
- collapse thread

***** Silenced by Fanboyslayer *****
5 3 [Posted by: fanboyslayer  | Date: 01/15/14 08:45:16 AM]
You silence them good, bwoy! Keep it up!
2 1 [Posted by: zlobster  | Date: 01/15/14 10:26:22 AM]

By any legitmate review AMD has scored a great victory with Kaveri and it's just the beginning of a new era in X86 computing. The HSA architecture is so good that virtually all major PC players are onboard. Laptop users will see even more benefits to Kaveri so it's all good. Kaveri has proven that CPU core performance alone no longer is the deciding factor is system performance. Those days are long gone. AMD once again delivers a superior PC solution and at a very affordable price.
7 4 [Posted by: beenthere  | Date: 01/15/14 08:27:13 AM]

A very important step for AMD and PC arch. as a whole. I was just a bit surprised that the drivers for HSA wouldn't be ready for a few more months. They should have been ready at launch.

It'll be a great day, but compilers won't have HSA support for probably a year now.
3 1 [Posted by: KeyBoardG  | Date: 01/15/14 10:55:44 AM]

show the post
3 8 [Posted by: amdzorz  | Date: 01/15/14 11:53:37 AM]
- collapse thread

***** Silenced by Fanboyslayer ****
4 2 [Posted by: fanboyslayer  | Date: 01/16/14 01:22:23 AM]

show the post
3 8 [Posted by: amdzorz  | Date: 01/15/14 11:54:54 AM]
- collapse thread

**** Silenced by Fanboyslayer *****
3 2 [Posted by: fanboyslayer  | Date: 01/16/14 01:24:44 AM]
Can you Find on Intel i3 price that will offer Playable FPS Like 24FPS, your i3 can deliver only at 10FPS, it cant compete in Graphics Performance without the of Descrete GPU.

Now your System will Cost Higher. compare to AMD Built.

But if im more serious on High Resolution and Graphics. well my Choice will be a Higher CPU and GPU Like Intel i5 Build ATI R Series or Nvidia GTX Series
1 1 [Posted by: xentar  | Date: 01/16/14 05:34:33 AM]

show the post
3 7 [Posted by: amdzorz  | Date: 01/15/14 11:55:07 AM]
- collapse thread

***** Silenced by Fanboyslayer *****
3 2 [Posted by: fanboyslayer  | Date: 01/16/14 01:26:15 AM]

show the post
3 7 [Posted by: amdzorz  | Date: 01/15/14 11:56:55 AM]
- collapse thread

***** Silenced by Fanboyslayer *****
3 1 [Posted by: fanboyslayer  | Date: 01/16/14 01:26:23 AM]

show the post
4 8 [Posted by: amdzorz  | Date: 01/15/14 12:09:07 PM]
- collapse thread

It mine's Dogecoins better than any Intel or NVidia chip. Wow.
2 1 [Posted by: KeyBoardG  | Date: 01/15/14 05:12:23 PM]
***** Silenced by Fanboyslayer *****
3 2 [Posted by: fanboyslayer  | Date: 01/16/14 01:26:40 AM]

show the post
3 8 [Posted by: amdzorz  | Date: 01/15/14 12:12:19 PM]
- collapse thread

***** Silenced by Fanboyslayer *****
2 2 [Posted by: fanboyslayer  | Date: 01/16/14 01:28:17 AM]

i can't wait for VIA to but out Amd..
3 4 [Posted by: amdzorz  | Date: 01/15/14 12:12:56 PM]
- collapse thread

Can you post "I'm done writing to myself when you'll be done with it? Time to get breastfeed again honey.

Adults would like to share intelligent point of views here.
3 2 [Posted by: MHudon  | Date: 01/15/14 02:11:58 PM]
Boy, Fanboyslayer will be busy for quite some time with this one. :D

What happened with the moderation here? This site was once a neat place for many intelligent people to communicate constructively. Now mostly butthurt trolls are lurking between the posts...
2 2 [Posted by: zlobster  | Date: 01/15/14 07:13:03 PM]

I have to leave xbitlabs. Honestly 70% of all comments across any topic are fanboy trolling. It doesn't change opinions, just ruins the atmosphere of this site. If that was the goal, mission accomplished.

There are other new sites with less nonsense.
4 1 [Posted by: KeyBoardG  | Date: 01/15/14 11:10:16 PM]
- collapse thread

It's just one or two people who have narcissistic personality problems. As you would remember, things are not as bad now as they were 12 months ago. But the mod must take control again. In his absence, Fanboyslayer will silence them. You can also do your bit by voting them down to collapse their comments - to be never seen and never heard.
2 2 [Posted by: fanboyslayer  | Date: 01/16/14 01:33:33 AM]
As far as I know this section is not moderated.
Our forums here are..
2 1 [Posted by: Sauron_Daz  | Date: 01/16/14 07:22:53 AM]

I still don't understand the testing regimens of the testers as a chip with a build in graphic processor should be tested against chips with inbuilt graphics. Using a separate graphic card in the test is stupid and pointless exercise. Even more pointless is using a graphic card from another manufacture. If the tester want to use a graphic card in their test use a compatible card. viz. an R7. as the chips in question are able to run these cards in crossfire mode. Any other way is like comparing a 4cylinde engine of about 2 lts. with a v12 Varon. Every one knows this is silly so why do the so called experts do this? Is it placate their advertisers or sponsors?
1 1 [Posted by: tedstoy  | Date: 01/16/14 06:57:47 AM]
- collapse thread

Spot on again Ted.
0 1 [Posted by: linuxlowdown  | Date: 01/17/14 07:51:13 AM]


Add your Comment

Related news

Latest News

Wednesday, November 5, 2014

10:48 pm | LG’s Unique Ultra-Wide Curved 34” Display Finally Hits the Market. LG 34UC97 Available in the U.S. and the U.K.

Wednesday, October 8, 2014

12:52 pm | Lisa Su Appointed as New CEO of Advanced Micro Devices. Rory Read Steps Down, Lisa Su Becomes New CEO of AMD

Thursday, August 28, 2014

4:22 am | AMD Has No Plans to Reconsider Recommended Prices of Radeon R9 Graphics Cards. AMD Will Not Lower Recommended Prices of Radeon R9 Graphics Solutions

Wednesday, August 27, 2014

1:09 pm | Samsung Begins to Produce 2.13GHz 64GB DDR4 Memory Modules. Samsung Uses TSV DRAMs for 64GB DDR4 RDIMMs

Tuesday, August 26, 2014

10:41 am | AMD Quietly Reveals Third Iteration of GCN Architecture with Tonga GPU. AMD Unleashes Radeon R9 285 Graphics Cards, Tonga GPU, GCN 1.2 Architecture