AMD Introduces Carrizo-APUs and graphics cards for OEMs




AMD comes with the 7000-series APUs for laptops. It involves chips that were formerly known under the name of Carrizo-L, which have support for the so-called Heterogenous System Architecture. Besides the APUs AMD is also introducing graphic cards for OEMs.

AMD unveiled its new Carrizo-L APUs last year. The APU and the southbridge therein are combined in a single which, making the chip is a soc. Carrizo-L must above all be a lot more efficient than predecessor Kaveri. The processor features Puma + -cores, a development on Puma architecture. The graphic part is provided with the Graphics Core Next GPU, which supports DirectX 12 and Mantle. The chip provides the first APUs full support for the Heterogenous System Architecture, allowing the CPU and GPU memory sharing software and both also share more if one can approach entirely.

These introduced chips are part of the Carrizo-L-series. Which is designed for entry-level and mainstream notebook market. Later this year will also be a ‘normal’ Carrizo-series laptops and hybrid systems with a higher TDP. These APUs feature Excavator cores, an evolution of the Steamroller cores used in-Kaveri APUs.

The top model of Carrizo-L, the A8-7410, features four cores with a maximum clock speed of 2.5GHz. The base clocks of the chips, as well as the configuration of the graphics part of the chip, AMD has not disclosed. The TDP is 12-25W. Meanwhile, the APUs are available in China; when they are first seen in the Netherlands, is not yet known.

Model Cores Maximum memory speed Clock L2 cache TDP
AMD A8-7410 4 1866MHz up to 2.5GHz 2MB 12-25W
AMD A6-7310 4 1600MHz up to 2.4GHz 2MB 12-25W
AMD A4-7210 4 1600MHz up to 2.2GHz 2MB 12-25W
AMD E2-7110 4 1600MHz up to 1.8GHz 2MB 12-15W
AMD E1-7010 2 1333 up to 1.5 GHz 1MB 10W
Besides APUs AMD also comes with graphics cards for OEMs. This fall in the new 300 series, but the architecture has not changed. The maps are based on the same Graphics Core Next architecture as their predecessors. The configurations of amounts stream processors and memory for some cards are different. The R9 380, 370 and 360 and R5 330 have no direct counterparts in the 200 series. The R7 350, R7 340 and R5 340 have though. The first two are compared to their predecessors from the 200 series with the same model number left untouched; only the amount of memory may differ. The maximum clock speed of the GPU-R5 340 is increased relative to the R5 240 at 45MHz.

Model Stream Processors Clock Memory Form Factor
R9 380 1792 to 918MHz up to 4GB GDDR5 Full height, dual slot
R9 370 1024 to 975MHz 2GB or 4GB GDDR5 Full height, dual slot
R9 360 768 up to 1050MHz 2GB GDDR5 Full height, dual slot
R7 350 384 up to 1050MHz 1GB or 2GB GDDR5 / DDR3 Full height, single / dual slot
R7 340 384 to 780MHz 1GB or 2GB GDDR5 or 2GB / 4GB DDR3 Low profile, single slot
R5 340 320 to 825MHz up to 2GB GDDR5 or DDR3 Low profile, single slot
R5 330 320 to 855MHz up to 2GB DDR3 Low profile, single slot
For laptops, AMD introduced the M300 series. These cards are also based on the GCN architecture. As support the cards in the M200 series they DirectX 12 and OpenCL 1.2. The maximum clock speeds of the GPU’s are inflated to over 1GHz, though OEMs can limit the clock speed to control the heat output. Compared to the R7 M260 is the memory bus of the R7 M360 halved. The R9 M375 does retain a 128-bit memory bus, but that card supports relative to the R9 and R9 M260- M270-series no longer gddr5, but the slower DDR3.

Model Stream Processors Clock Memory Memory
R9 M375 640 up to 1015MHz up to 4GB DDR3 2.2GHz 128-bit
R7 M360 384 up to 1015MHz up to 4GB DDR3 2GHz 64-bit
R5 M330 320 up to 1030MHz up to 4GB DDR3 2GHz 64-bit


In: A Technology & Gadgets Asked By: [19121 Red Star Level]

Answer this Question

You must be Logged In to post an Answer.

Not a member yet? Sign Up Now »