All eyes continue to be glued to Intel and its upcoming Core i7 (Nehalem), but AMD has a product release in the wings too, this one for the server market. The struggling chip maker said it's planning to release a new server platform in the second half of 2009 currently code named Fiorano. Built to take advantage of AMD's upcoming 45nm Shanghai processor, Fiorano represents the company's first foray into the server chipset market instead of using chipsets from Nvidia and Broadcom.
The Fiorno platform will fully support the company's chip-to-chip technolgy called HyperTransport 3 while also offering a new virtualization technology called IOMMU, which allows for the virtualization of the system's I/O traffic. Support for the second generation PCI-Express will also be included, but the same can't be said for DDR3 because of cost concerns.
"it will hit once the price of DDR3 comes down," said John Fruehe, who handles worldwide channel market development for AMD's Server and Workstation Division. "The back half of next year is about the time the process changes in DDR3 will happen that will allow the prices to come down."
The first AMD platform to use DDR3 memory will be called Maranello (previously known as Piranha).
We did not expect this. When we first got our hands on Zalman’s CNPS9300 AT, we assumed the company had pulled a “Honey, I Shrunk the CPU Cooler” on its flagship product, the bulky CNPS9700. That’s certainly true if you consider the tale of the tape: The CNPS9300 is 80 percent smaller than its big brother, and its total thermal dissipation area has been nearly halved, from 5,490cm2 to 2,583cm2.
Logic only dictates that this cooler should perform far worse than the Zalman CNPS9700. But the built-for-silence CNPS9300 AT nearly matches its big brother’s performance—as well as that of our top cooler, Thermaltake’s DuOrb (reviewed July 2008).
Instead of a god of the sea, Gigabyte’s midtower Poseidon 310 chassis is a petite prince. But that’s merely a reflection of this case’s size, not its prowess. It clocks in at 7.75”x17”x20”—small enough to fit into that nook in your desk or the space under your bed.
Even given its small size, the Poseidon supports up to five 5.25-inch devices. We’re unsure why this case—or any case, for that matter—still bothers with multiple external 3.5-inch bays. You get two helpings of them on the Poseidon. We would have rather sacrificed these and an additional 5.25-inch bay in favor of more internal hard drive space. Though we’re not complaining about what we get for internal storage: three hard drive bays with included rails.
Full review (with a verdict and everything!) after the jump.
It won't be long before single-core processors will seem as antiquated as single-speed CD-ROM drives, and the case could be made that we're already there. Dual- and quad-core processors rule the landscape, and while Intel's upcoming Core i7 has enthusiasts frothing at the mouth, the chip maker may have something even more mouth watering in the very near future.
If the latest rumor turns out to be true, expect a replacement architecture for Nehalem in 2010 which will double the number of cores per die to eight. Codenamed Sandy Bridge, alleged leaked slides suggest the new architecture will also support hyperthreading, giving the eight-core chip a generous 16 threads to work with. Also look for 16MB of L3 cache to find its way onto the chip.
But for all the hardware goodness, it's the software that may end up playing the biggest role in performance improvements. Intel will reportedly introduce a new instruction set called Advanced Vector Extensions (AVX) that will eventually supersede SSE. AVX will double the size of instructions to 256 bits and will be capable of performing up to four calculations in a single instruction.
With over a year to go before the supposed new architecture makes a debut, will developers be ready by then to take advantage of the additional cores and new instruction set?
Indilinx has completed the development of their Barefoot (IDX22) high-performing solid state drive controller with 90nm process technology which shows an impressive fastest read speed of 230MB/s and supports a capacity of up to 512GB with multi-level cell (MLC) NAND flash. Indilinx claims “phenomenal performance at a competitive price”.
Barefoot supports native SATA 2.0 interface and provides maximum read and write speed of 230MB/s, 170MB/s with SLC NAND flash, and 200MB/s, 160MB/s with MLC NAND, respectively. It uses Indilinx’s unique architecture and technology, including independently operating 4 channels and external DRAM buffer and it enhances stability and reliability by using two types of hardware error-correcting code (ECC).
Those improvements are coming by leaps and bounds in SSDs. It's not clear if this will be competing with Intel’s controller directly. No mention if this is targeted at portable or stationary (or both) PC market.
Talk about a generational leap forward. The SSD revolution has barely begun, but while others are busy focusing on incremental capacity bumps nowhere near the size of the largest HDD, BitMicro says it can now make SSDs with a ginormous 6.5TB capacity.
According to TG Daily, the company made the claim at the Siggraph trade show held at the Los Angeles Convention Center. The rep went on to say that the custom drives can have up to 55,000 input/output instructions per second (IOPS) with sustained (not burst) transfers of up to 230 MB/s. In other words, not only would this wonder drive thoroughly trounce today's SSDs in terms of capacity, but it would be faster too.
The drives would also be physically bigger, with the loose-lipped rep saying the custom SSDs would be about two to three times higher than a regular drive.
Anyone think we'll see 1TB SSDs before long, let alone 6.5TB models?
If you don’t just like Gigabit ports—you love them— Gigabyte’s GA-EP45-DQ6 is the motherboard for you. This mobo has four Gigabit ports that can be teamed together for one seriously fat-ass network connection.
Elsewhere, the board is typical Gigabyte; it includes surface-mounted buttons and the most clearly marked USB and FireWire ports we’ve ever seen. So if you nuke your USB drive because you plugged the USB connector into a FireWire header, it’s your own fault, brother.
If solid state drives (SSDs) ultimately fail to curry favor among enthusiasts, it won't be from lack of effort in OCZ's headquarters. OCZ, who once sold only DRAM products and now offers everything from DIY notebooks to mind-controlled input devices, has been one of SSDs most aggressive pushers, first with its Core Series SSDs, and now with a second revision dubbed V2.
Whereas the original Core Series come in 32GB, 64GB, and 128GB flavors, the new V2 will be available in 30GB, 60GB, 120GB, and 250GB. By offering a 250GB model, a case could be made for going HDD-less in that new build.
Storage space isn't the only feature getting a boost and the V2 series will offer read and write speeds up to 170 MB/s and 98 MB/s respectively. By comparison, the previous Core Series top out at 143 MB/s and 93 MB/s respective read and write speeds. The V2 also touts improved seek times of less than 0.2-0.3ms as part of the new architecture.
One of the more interesting features OCZ brings to the table with the V2 is the ability to update the SSDs' firmware.The drives will come equipped with a mini-USB port, which paves the way for consumers to cash in on speed bumps and other future enhancements simply by installing new firmware.
Of course, the big concern with all SSDs continues to be the long term reliability, and OCZ rates its 2-year warrantied V2 series with a 1.5 million hour mean time before failure (MTBF). No word yet on pricing or availability.
Rambus hoped to make strides with its RDRAM technology, and while it never did take off, it's only a matter of time before a new technology shakes up the random access memory market. Engineers and physicists from Germany think they'll be the ones to do it and have demonstrated what looks to be the quickest prototype yet of an advanced form of RAM, one which challenges the fundamental speed-limit for the process.
Called magnetoresistive random access memory, MRAM uses a faster and more energy efficient version of today's modules, which store a digital 1 or 0 as the level of charge in the capacitor. By comparison, MRAM stores the same information by changing the north-south direction of a tiny magnet's magnetic field, with each variable magnet positioned next to one with a fixed field. To read the stored value, a current runs through the pair to discover the direction of the variable magnet's field.
There are different types of MRAM technologies, and the one most manufacturers are hedging their bets on is called spin-torque MRAM, which involves spinning electrons to flip the magnetic fields. In what could see the technology emerge in as little as just a few years, German researchers have now built a spin-torque system from tiny pillars just 165nm tall that looks to be dramatically faster than any other. The top of the pillar acts as a variable magnet that stores the data, with fixed magnets occupying the bottom. When a current passes through, the electrons reach the pillars' other end and flip the variable magnet region's field to match.
NewScientistTech has much more on the topic, and we want to hear what you think about the emerging RAM technology by posting below.
Toms Hardware reports that Intel’s "Extensible Host Controller Interface (xHCI) draft specification revision 0.9 in support of the USB 3.0 architecture, also known as SuperSpeed USB" is now available. This is a good indicator that we might see the first USB 3.0 demonstrations at next week’s IDF in San Francisco.
xHCI draft specification provides hardware component designers, system builders and device driver developers with a description of the hardware/software interface between system software. It is being made available under RAND-Z (i.e. royalty free) licensing terms to all USB 3.0 Promoter Group and contributor companies that sign an xHCI contributor agreement.
It doesn’t appear that the new spec will be backward compatible past USB 2.0. I find it hard to believe that USB 1.1 devices will be out of luck, so I plan to keep an eye on that aspect. USB 3.0 at 600 MB/s will offer a ten-fold increase in the bandwidth of USB 2.0 at 4.8 Gb/s. That is pretty impressive if it approaches it’s spec yield. USB 2.0 spec rate is 480 Mbit/s but typical USB PC-hosts rarely exceed sustained transfers of 280 Mbit/s.
Will you be wanting USB 3.0 on your future system?