On Thursday, Advanced Micro Devices (NASDAQ:AMD) officially released a number of SKUs from both of its 2013 APU product lines, Richland and Kabini/Temash. While only Kabini/Temash represent a new architectural design, both are designed to attack specific parts of the current mobile computing market where its main rival Intel (NASDAQ:INTC) is fundamentally weak.
I have talked at length here about how strong I believe the Jaguar core-based Kabini/Temash core design is. Up until recently, however, we had only AMD official marketing benchmarks and the odd bit of engineering silicon to give us a clue as to where they would fit in the market. The full lineup is now available and marketing slides can be found all over the web. I won't bore you with them here, but I do recommend looking at the full lineup of chips, if for nothing else than to try and make sense of AMD's naming convention.
Performance for the 15W quad-core A4-5000 is very comparable to that of a 17W Sandy Bridge Core i3 2377M and these benchmarks here confirm numbers that we've speculated on for a while now since the initial numbers were published by AMD back in February. And as these are a pure SoC with integrated north and south bridges, this is more like a 12W TDP part comparing very well with a 17WCore i3 in CPU intensive and mixed loads that will yield superior graphics performance for low-end gaming and not eat your battery in a matter of minutes. These are Sea Islands GPU's - GCN 1.1-not HD6000 series - VLIW4, the best AMD currently makes.
Why this is significant is price. Simply put, these SoCs are cheap. The embedded system versions of them will run between $39 and $72 and these retail chips should be cheaper. There is nothing in the upcoming Haswell lineup that will be able to compete with these chips at these prices and Bay Trail is six months away. To give you an idea of where Intel is on pricing, it released an Ivy Bridge Celeron at 1.0Ghz for $102 last month because its Ivy Bridge-Y series starts at $200 for a Core i3-3229Y.
Case in point, H-P released its updated line of laptops for the summer and the one that stands out is the $399 11.6 touchscreen Pavilion likely running a Temash A4 or A6 that has already grabbed headlines because of its price.
I haven't said much about Richland because there wasn't much to talk about until we got closer to real silicon, not just for it but also for Haswell. Richland is a mid-cycle refresh to last year's Trinity and for all intents and purposes it is the chip that Trinity was supposed to be - and frankly, needed to be. What is important to note is that while this is a refresh that boosts clock speeds by around 10%, the overall tweaks to power management have created a lot more thermal headroom and now allows for dynamic switching of power between the CPU and GPU on an as-needed basis. The result is a dramatic drop in power usage in GPU intensive tasks while creating the opportunity to improve performance, especially at the low end of the power curve, due to being able to stay at turbo clock-speeds longer.
If AMD's engineering slides are to be believed, then 15% to 50% battery life is expected over Trinity in common tasks. The Piledriver cores are exactly the same, so CPU performance scales with clock speed. Similarly, the GPU is the same Cayman-based one that was in Trinity. Again, clock speeds are up slightly across the board and the new ULV versions come in 17 to 35W varieties.
Now, remember, AMD is not fighting Intel for the performance crown. Even its marketing slides show this.
But, that said, where Richland gets interesting is that all of these little tweaks, clock speeds, power management, and higher supported memory speeds have created a lot of synergy in terms of performance. And there are enough independent numbers out there already that show AMD's claims of 20-40% graphics performance over Trinity, with only a 5% increase in clock speed on average, is quite impressive for what is -- essentially a stop-gap part while we wait for Kaveri.
Adding in some of the ancillary functionality coming with Richland -- facial recognition/login, gesture control of Windows 8 and low-latency wireless monitor support -- there is the beginning of AMD trying to create something new for the jaded PC buyer.
The Value Proposition
Both Kabini and Richland are very good products for this moment in computing history. Neither are game-changers for the industry. But they are good enough that given the current state of the competition that they should drive market share gains and put AMD on more solid footing for 2014. Kabini is head and shoulders better than AMD's most successful chip, Brazos and Intel has not responded with a price or power competitive part.
With Richland AMD will be price and performance competitive in the mid-end of the mobile market and provide some added value if the bundled gesture and facial recognition software are better than gimmicks. One of the things that the PC industry has failed to do in a long time is capture the imagination of buyers. Buying a computer as about as interesting as buying corn chips. The Nintendo (OTCPK:NTDOF) Wii had that wow-factor that created a runaway hit without having the best technology. Apple's (NASDAQ:AAPL) iPad and Microsoft's (NASDAQ:MSFT) Kinect for the Xbox did as well. PC's have rested on their functionality laurels for too long and the industry needs to do something or be subsumed by sub 5" mobile devices within two years.
Product differentiation is what drives success in the market. A $400 touchscreen netbook that can play League of Legends well and act as an HTPC is a great product depending on the battery life and one that has me far less interested in a tablet than I would have been two months ago. For what we can expect from Haswell, I don't see that on the horizon, not when the CPU itself will practically cost that much.
Disclosure: I have no positions in any stocks mentioned, and no plans to initiate any positions within the next 72 hours. I wrote this article myself, and it expresses my own opinions. I am not receiving compensation for it (other than from Seeking Alpha). I have no business relationship with any company whose stock is mentioned in this article.