AMD Brings Kabini to the Desktop
AMD Packs a Punch at Low Power
Perhaps we are performing a study of opposites? Yesterday Ryan posted his R9 295X2 review, which covers the 500 watt, dual GPU monster that will be retailing for $1499. A card that is meant for only the extreme enthusiast who has plenty of room in their case, plenty of knowledge about their power supply, and plenty of electricity and air conditioning to keep this monster at bay. The product that I am reviewing could not be any more different. Inexpensive, cool running, power efficient, and can be fit pretty much anywhere. These products can almost be viewed as polar opposites.
The interesting thing of course is that it shows how flexible AMD’s GCN architecture is. GCN can efficiently and effectively power the highest performing product in AMD’s graphics portfolio, as well as their lowest power offerings in the APU market. The performance scales very linearly when it comes to adding in more GCN compute cores.
The product that I am of course referring to are the latest Athlon and Sempron APUs that are based on the Kabini architecture which fuses Jaguar x86 cores with GCN compute cores. These APUs were announced last month, but we did not have the chance at the time to test them. Since then these products have popped up in a couple of places around the world, but this is the first time that reviewers have officially received product from AMD and their partners.
Big Step up from Brazos
Several years ago AMD released the Brazos platform which combined the Bobcat CPU architecture with the Evergreen graphics component in a product called Zacate? Enjoying all the code words? Wikipedia keeps a nice record of them all. This was AMD’s first low power APU, and it was a pretty solid success for the company. Two Bobcat cores proved to be adequate competition to Intel’s Atom based products, and the graphics portion was a significant step above what the rest of the industry had to offer in that particular power envelope.
The Bobcat architecture is a big step away from what AMD was doing with Bulldozer at the time. IPC was important, but so was power. They went with a lower clockspeed and a lot of pretty clever design decisions to get IPC up there without sacrificing too much die space or power to get there. Keeping the clockspeed low allowed them not only to save power, but to really manipulate the core architecture to really push the work per clock as compared to Bulldozer which relies on higher clockspeeds to achieve good performance.
Jaguar builds upon this to a great degree. AMD has doubled the core count from two to four with Kabini, and they have also improved IPC and power consumption. No real sacrifices were made to achieve this, or at least so the engineers claim. No souls were lost in the design of this APU. AMD did extensive work on the front end (including a beefier branch predictor), re-arranged the integer and SSE/MMX/AVX pipelines to balance out the workload, and improved the caches. Here is a quick reference to what work was done.
Jaguar also embraces all of the latest instruction sets. This includes BMI, AVX, FC16, SSE 4.2, and AES. Unlike Bulldozer and its variants, Jaguar does not utilize a “modular” architecture. Each core in Jaguar is self sufficient with its own floating point unit, as compared to the shared unit with each dual core “module” in Bulldozer/Piledriver/Steamroller. The only thing shared is the decently sized L2 cache, which is 2 MB in size. Caches typically use up a lot of power, so the smaller the L2 cache the better… until it gets so small that it has a big negative impact on overall performance. 2 MB seems to be a pretty good balance between performance and power.
Kabini sports the latest iteration of AMD’s GCN architecture. It has support up to DX 11.2, OpenGL 4.3, and OpenGL ES 3.0. Kabini is not fully HSA compliant, but it does support a lot of HSA-like features. It fully supports OpenCL 1.2, DirectCompute, and C++AMP. All of the HSA optimizations in Kaveri did not trickle down to Kabini, but for a product in this particular niche, this is not necessarily a disadvantage. Kabini features 2 GCN compute cores with 128 stream units. In this case there are 4 x 16 wide vector units per compute core with a scalar co-processor. Multimedia features include UVD 4.2 (Universal Video Decoder) and VCE 2.0 (Video Coding Engine).
Kabini is made on TSMC’s 28nm HKMG process as of now. This could change in the future, but for the moment AMD is sticking with TSMC for this particular product. TSMC has been producing 28nm products in a variety of flavors for over three years now. If the reader has not gathered by now, Kabini is not new. It was introduced in Q2 of last year and Ryan covered the mobile implementation here. This chip has been used in low end notebooks as well as higher powered tablets.
Something that sometimes gets glossed over is that Kabini is a true SOC. It does not need a separate southbridge for I/O functions. Onboard it features 2 SATA 6G, 2 USB 3.0, and 8 USB 2.0 ports. It has multiple PCI-E 2.0 lanes coming out that can connect to peripherals such as Ethernet and wireless controllers.
Since power is the primary concern, AMD stuck with a single DDR-3 channel running at a max speed of DDR-3 1600. The chip can push up to three displays at one time. It natively supports DisplayPort, HDMI, DVI, and VGA outputs.
Jaguar is the basis for the latest generation of consoles from Sony and Microsoft. These custom designed SOCs feature 8 core Jaguar implementations attached to much more robust GCN based units than we see here. Still, it is interesting to see that Jaguar is going to be the basis for a lot of games that will be released on both consoles and PCs for many years to come.
Good review. Nice to see AMD
Good review. Nice to see AMD back in the upgradeable low-end arena again unlike it’s competitor. I believe in emerging markets this might be a hit as well as HTPC/router/NAS for the rest of us.
Question though how well does this play 4k high bit-rate videos? Does it play the puppy video well?
Have you tried streaming the
Have you tried streaming the 4K videos posted on YouTube? I would also like to know how well it would handle Bluray playback too.
Nice article 🙂
AMD did
Nice article 🙂
According to Kanter over at Real World Tech, that branch predictor proved to be quite good so AMD implemented it also in Piledriver. I think a evolution of Jaguar keeps getting more and more interesting 🙂
Quarter of the power, half of
Quarter of the power, half of the performance.
^This is what tweaks my nethers!!!!!
EPIC JOB AMD!
“It is unfortunate that I was
“It is unfortunate that I was unable to get a working Intel Bay Trail D based product in for testing, but all indications point to these AM1 parts outperforming the Intel J1900 and J2900 products across the board. ”
Unlikely.
Anandtech has a few dual core bay trail D numbers, bay trail out performs in some (single threaded), not in others (multithreaded i.e. dual vs quad). The quad core bay trail D, I suspect would come in equivalent to slightly better on the CPU side at lower power while the GPU will be behind, naturally.
Here’s a good comparison at
Here’s a good comparison at the Tom’s Hardware article-
http://www.tomshardware.com/reviews/athlon-5350-am1-platform-review,3801.html
Hey Josh, good review, but
Hey Josh, good review, but can we get a more sensible approach to the power consumption tests? Testing a 25W Kabini with a 500W PSU seems a poor way to go about it. Might I suggest that you guys have a picoPSU for the low wattage Atoms, Kabinis and the like? These chips are likely to end up with external PSUs and in miniature form factors anyway, thus I figure utilizing lower wattage PSUs that are more efficient at these power consumptions, or even make more sense with respect to form factors, seems far more sensible.
Thanks.
I finally did get access to
I finally did get access to one of those really small power supplies. I will see if I can't get that up and running this week to test power consumption!
Awesome! Thanks for the
Awesome! Thanks for the response.
if only they allowed
if only they allowed overclock it would be a fun toy
Good review!
Good review!
Good benches but please get
Good benches but please get someone to proofread your stuff before you publish it. Your writing is awful. The sentences are downright confusing, the structure is terrible and your use of punctuation is erratic at best. If it weren’t for your profile suggesting otherwise I would have assumed that English is not your first language.
This article provides good
This article provides good benches, but please employ someone to proof-read your articles before they are published. Your writing is awful; the sentences are confusing, the structure is terrible, and your use of punctuation is erratic at best. If it were not for your profile suggesting otherwise, I would have assumed that English is not your primary language.
Thanks for reading!
That’s a burn. A+.
That’s a burn. A+.
That was awesome!
That was awesome!
can the sempron or athlon
can the sempron or athlon (2014 version am1 socket) be crossfired or just have an graphics card like a simple r7 250 or hd7770 or something in it and does it handle stable? plz reply by mailing me @ ownertje@hotmail.com
greets
luuk AKA AMDfan 😀
i have a fx8350/r9 270x rig just asking if it could be a sort of mini streaming and light gaming pc like skyrim @ low or something
You could use a single
You could use a single low-to-low/mid-range graphics card, but CrossFire is basically out of the question. The Kabini chips only support 4 PCI-E 2.0 lanes, but there are AM1 Platform motherboards with x16 slots (that run at a max of x4 electrically).
Actually, they support PCIE
Actually, they support PCIE 3.0 x4, not 2.0 x4, which is equivalent to PCIE 2.0 x8.
Nope, Kabini is at PCI-E 2.0.
Nope, Kabini is at PCI-E 2.0. AMD did this for power and money reasons. PCI-E 3.0 does take up more power as compared to 2.0. Certification for 3.0 does take more time and money in the development process. So for a low cost/low power part like this, they stuck with PCI-E 2.0.
Man, I would love to see an
Man, I would love to see an 8-Core Version with a GPU with 1152 Steam Processors for future steam machines. That may make a nice game system.
Josh
Kabini A6-5200 is
Josh
Kabini A6-5200 is manufactured at TSMC. But Athlon 5350 seems to be manufactured at Globalfoundries, Dresden. The “Diffused in Germany” marking on the heat spreader in the image on this page gives clues that its manufactured at Globalfoundries. Jaguar is a highly portable CPU core design and so is the GCN based GPU. With Athlon AM1, Kaveri, Berlin, Seattle and some console APU production (starting likely in H2 2014) all being manufactured at Globalfoundries, AMD should have no problems meeting the USD 1.2 billion 2014 wafer commitments.
AMD can’t give official word
AMD can't give official word right now on that due to them being in their "quiet period" but it does look like you are correct. Good for GF, I guess. Odd that it is diffused in Germany, rather than utilizing the US fabs in NY. I thought FAB 8 was the primary site for 28 nm HKMG… guess not.
Josh griped on the podcast
Josh griped on the podcast that the testing motherboard only had 2 SATA ports, so you can’t have fast storage (SSD) AND big storage (spindle) AND optical storage. Optical drives work pretty well on USB 2…
This is true. Good plan.
This is true. Good plan.
With such low power
With such low power consumption, if a board manufacturer made a raid capable board with 6+ Sata III ports, would this make a good NAS / media server? It sounds ideal.
For everyone who need
For everyone who need additional SATA port, you can always add it using PCIe card. Considering PCIe 2.0 x4 has 2GB/s bandwidth and x1 has 500 GB/s bandwidth (comparing to SATA1 150 MB/s, SATA2 300 MB/s, SATA3 600 MB/s), there are sufficient bandwidth to make SATA III run smoothly using PCIe expansion card x4 or you can use cheap 10 buck USD PCIe 2.0 1x for optical drive while the integrated SATA3 ports are used for storage drive. With 1 slot of PCIe 2.0 x4 and 2 slot x1 slot, i think we can add plenty of SATA port.
^ of course you need
^ of course you need motherboard with that feature. Kabini platform itself capable of providing 3 x1 and 1 x4 PCIe 2.0 lane.
I want to see this progress
I want to see this progress its way to having R9 GPU cores and maintain the 2ghz quad core. 60mm copper/aluminum water cooler with a variable speed pump run off CPU fan header. get someone to make a single channel dimm board with no pci-e slot. have x1 slot for wifi chip. 1 8 gig stick of 1866 with a built in cache slot on board like used in older laptops. Would make a pretty nice really small form factor gaming rig in the 4-500 dollar range. Kinda almost have to rethink the whole mini itx thing tho. laptop disc drives onboard power adapter.say like 4.5×4.5×4.5 gaming rig that was almost silent.