The guys over at Micron, who bring you some of the fastest memory chips on the market, introduced a new series of RealSSD solid state drives for enterprise computing and notebook applications. According to the official press release, these new SSDs are providing a dramatic improvement in capacity, power and performance for the applications they serve.

“We are seeing SSD interest in a variety of applications where historically hard disk drives have reigned. For many, the most logical place is in notebook computers, but there is incredible value for SSDs in enterprise server systems,” said Dean Klein, vice president of memory system development at Micron.

Micron’s RealSSD P200 family is specifically designed for enterprise computing and comes in densities ranging from 16 gigabytes (GBs) to 128GBs and is available in a standard 2.5-inch form factor. The overall improvements are facilitated through Micron’s high-performance single-level cell (SLC) NAND technology, which allows the P200 series to provide the industry’s fastest 3Gb/s SATA-based sequential read and write speed of up to a maximum of 250 megabytes per second (MB/s).

The RealSSD C200 lineup is aimed at notebook applications. C200 will be available in 2.5-inch and 1.8-inch form factors, with the 2.5-inchers providing densities of up to 256GBs, while the 1.8-inchers offer densities from 32GB up to 128GB. The notebook line provides a read speed of up to 250 MB/s and a write speed of up to 100 MB/s.

“With our C200 products, we are providing a balanced price to performance solution specifically designed for notebook applications by utilizing MLC NAND technology and highly optimized NAND management algorithms,” continued Klein.

Mass production for both P200 and C200 lineups is expected in the fourth quarter of 2008.

ATI’s HD 4870X2 is a few weeks away from official launch, and although the R700 generation seems to be one of ATI’s best realizations thus far, the Canadians aren’t going to stop here. Some new info on the next generation has been recently leaked and it seems ATI wants to improve the performance per watt aspect, first and foremost.

The n RV870 (codename “Lil Dragon”) next gen GPU is primarily a shrink of RV770. According to NordicHardware, the node following 65/55nm (current gen) would be 45/40nm , as Intel’s and AMD’s CPU demonstrated earlier this year. However, there’s a strong possibility that the next ATI GPU would skip the 45nm phase directly to 40 nm. The main reason to do this is the achievement of a performance per watt ratio twice that of RV770.

Surprisingly, the overall performance gain for the RV870 is set somewhere around 20% over the RV770, but ATI will also drop the power consumption significantly. NordicHarsware estimates that the next gen GPU is going to integrate to 960 unified shaders (192x5), 48 texture units and 24 ROPs. Thus, the TDP of the RV870 should be around 150W, in case ATI maintains the usual frequencies.

Another mystery is represented by the inclusion of the DX11 support. Since DX11 is going to be available in late 2009 and the new GPU will be release next summer, ATI will probably stick to DX10.1 for the time being.

Scythe is known for some pretty good CPU coolers, although these might not be the best one can buy. The Asian company has recently entered the PC case market and their latest Innovator model seems to point out that Scythe could further diversify its line of products.

The Innovator is quite massive, measuring 605x550x208mm (HxDxW). I comes standard with enough room for two ATX power supplies. The innovation here is a special cable that powers the secondary PSU at the same time as you power on the system.

According to Fudzilla, the case is made of steel, so it should provide enough cooling potential, but Scythe thought it would be a good idea to fit a plastic front door to it that hides the six 5.25-inch drive bays. In addition, the Innovator supports two 3.5-inch drive bays and six internal 3.5-inch drive bays. Also in front, users can find four USB 2.0 ports, one FireWire port and a headphone and mic jack.

More features include rubber coated thumb screws for better grip and reduced vibrations, a bunch of extra long thumb screws for add-on cards, especially for graphics cards.

Weighing 14.8kg, it certainly appears to be server-oriented. It is already available in Japan for about €78, but Scythe didn’t mention anything about world-wide availability.

AMD was the first to announce the upcoming Fusion processor that combines the power of CPUs and GPUs. Now Intel strikes back and presents its own Fusion-like solution that is supposed to be released in Q3-Q4 2009.

The Intel hybrid platform is codenamed Havendale and will feature a CPU with integrated GPU capabilities as well as an Ibex peak chipset which will be a single-chip solution, fusing the Northbridge and the Southbridge together. Even though the hybrid processor integrates an iGFX GPU, the motherboards will still support external PCIe graphics.

According to Fudzilla, the CPU part of the Havendale is only a dual-core, Nehalem derived 45nm model. OK, it’s not a six-core, but it comes with integrated memory controller on the CPU side of the chip. The processor is DDR3-only compatible and it interconnects with Ibex peak directly through a memory interface (DMI).

In order to provide decent graphics performance, Intel is going to interconnect the Ibex peak unified chipset with the GPU through Intel's Flexible Display Interface (FDI). Additionally, the new chipset has an input/output part and something known as a Manageability engine, already present in Intel’s 4-series Northbridge.

Although it’s not a high-end processor, the integration of a CPU and a GPU will mean lower power consumption and greater performance at the same time.

We know Roadrunner is the fastest supercomputer on Earth (for the moment at least), but how much does it cost to assemble and maintain such a behemoth? Millions of dollars? Not that much for a government-funded project. I guess the guys who assembled the Roadrunner didn’t take into consideration what NVIDIA CUDA can provide for their endeavors. CUDA forces all unified stream processors found in a GPU to work as separate CPUs. So if we have two NVIDIA 9800GX2 cards, that would amount to 2X256 processors running at around 1,4 GHZ. This setup alone can turn you computer into a miniature supercomputer that can solve complex equations and coordinate impressive simulations. How about 8 of these cards working in parallel?

MIT graduate students Nicolas Pint, David Cox and James DiCarlo have managed to assemble an impressive 16-GPU system composed of eight 9800GX2 video cards donated by NVIDIA. That would translate into 2048 processors that would deliver more than 20 TFLOPS. The CUDA architecture will soon be adopted by ATI, as well, and that means we will get to use 1600 processors on a single Radeon 4870X2 card. Sure, the ATI unified processors are clocked at a lower frequency than those found on NVIDIA’s cards. Still, eight 4870X2 will net you the power of 6400 processors clocked at around 800 MHz.

Ubergizmo informs that the high-throughput method the three students promote can also use other ubiquitous technologies like IBM's Cell Broadband Engine processor (found in Sony's Playstation 3) or Amazon's Elastic Cloud Computing services. What puzzles me is the fact that the team is also involved in the PetaVision project on the Roadrunner, so why didn’t they use the CUDA architecture? That would have cut the costs drastically. I reckon the Roadrunner had been designed before the CUDA architecture actually got released.