Whereas AMD’s Radeon HD 4830 resembles a Radeon 4870 after a partial lobotomy, the Radeon HD 4850 that sits between these two cards comes with a full complement of 800 stream processors. But don’t make the mistake of thinking you can overclock a 4850 board to achieve the same performance as one based on the 4870: The latter uses GDDR5 memory while the former is limited to GDDR3.
Nvidia’s GeForce 9800 GT is really just a rebadged GeForce 8800 GT, which makes it the only card in our roundup based on a previous-generation GPU architecture: Nvidia’s 65nm G92. Despite its age, however, the G92 helped EVGA’s GeForce 9800 GT best PowerColor’s Radeon HD 4830—at least in terms of gaming performance.
EVGA runs the 9800 GT’s core at 600MHz, but takes full advantage of its 112 shader processors’ capacity for operating at much higher frequencies: 1,500MHz in this implementation. The card has a 256-bit memory interface to a full gigabyte of GDDR3 memory running at 900MHz.
This card is based on Nvidia's most current GPU architecture, the GT200. Priced at $200, it's the least expensive model we tested that's capable of running Crysis at 60-plus frames per second.
If you shop for a GeForce GTX 260 card, make sure you're comparing apples to apples: Core 216 models like the one you see here are manufactured using a 55nm process, and are outfitted with 216 shader processors. Conversely, cards based on the original 65nm GTX 260 GPU remain on the market but possess only 192 processors. Both versions have a 448-bit interface to 896MB of GDDR3 memory.
The 55nm RV770 is one of the best arrows in AMD’s GPU quiver, so it’s a good thing the part has proven to be both versatile and powerful. As deployed in the Radeon HD 4870, the RV770 has a full complement of 800 stream processors—just like the Radeon HD 4850—but in this design, the GPU is paired with GDDR5 memory.
GDDR5 memory boasts a very high data rate (ranging from 3.6Gb/s to 6.0Gb/s, compared to GDDR3’s 1.0Gb/s to 2GB/s). This enables AMD to deliver nearly the same memory bandwidth through a relatively narrow and inexpensive 256-bit bus as it would with a much wider and costlier 512-bit bus.
Nvidia pretty much owns the top end of the GPU market, thanks to the mighty, dual-GPU GeForce GTX 295. But no manufacturer can survive by selling low-volume parts, no matter how pricey they may be. Selling oodles of moderately priced products is where the real money is made. And that’s where the GeForce GTX 275 comes in.
Nvidia would never have concocted the GTX 275 had AMD not launched the Radeon HD 4890. Competition is the consumer’s friend.
Thanks to the recent move to 34-nanometer manufacturing, Intel has been able to create a new series of SSDs, which will (eventually) sport higher capacities, and reduce costs. The new price for the 80GB X25-M drive is $225 (a 60 percent decrease from the $595 price tag a year ago). The 160GB version is down to $440, which is down from its introductory price of $945.
Though, we’ll have to wait a bit for higher capacities. According to Intel’s marketing manager for the NAND Products Group, Troy Winslow, “What we decided to do is split 34-nanometer into a two-step process.” The first step will be to cost-reduce existing 80GB and 160GB drives. “And what we'll do later--and it's not even going to be this year but first half of next year--we will introduce, also on 34 nanometer, a performance enhancement and a doubling of the capacity.”
So what does all this mean? Simply, we won’t see drives over 300GB until next year. Still though, the price cuts are very welcome.
It's been a wild ride for the x86 architecture, which has managed to stay relevant for over 31 years. During that time, AMD has sold a boatload of x86 chips, with today marking the shipment of the chip maker's 500 millionth x86 CPU. This occurred while AMD was celebrating its 40th anniversary.
"AMD's 40th anniversary is a testimonial to our longevity, our employees, our customers, and our unique business approach," AMD wrote on a webpage dedicated to the milestone.
As a "reward to the loyal customers who helped AMD reach this milestone," the chip maker is giving away four HP Pavilion dv2z notebooks via a contest. To enter:
Follow AMD on Twitter (@AMD_Unprocessed), where a new question will be posted every other Monday beginning July 27
Send AMD the answer via a direct message
Those who answer correctly will be entered into a drawing that will take place over an eight week span. More details can be found here.
Back in February of this year, Samsung developed and validated its first 40nm DRAM chip, and now five months later, the chip maker announced it has begun mass producing 2Gb DDR3 using the smaller manufacturing process.
Samsung says the move to 40nm will provide around a 60 percent increase in production capacity over a 50nm process, and it won't all be relegated to the server market, according to news and rumor site DigiTimes. In addition to 16GB, 8GB, and 4GB RDIMMs for servers, Samsung will use the 40nm manufacturing process to build UDIMMs (unregistered in-line memory modules) for workstatios, desktops, and notebooks of up to 4GB.
The energy efficient chips support a data rate of up to 1.6Gbps at just 1.35V, double that of an 800Mbps 1Gb-based dual-die package.
Remember when computer cases were little more than boring beige rectangular boxes with a single fan? My how the landscape has changed since then, as evidenced by NZXT's newest chassis, the M59.
Aimed at gamers, the M59 sports a funky aesthetic even by today's standards, which will probably appeal to the target audience. But looking beyond the exterior facade, NZXT claims exceptional airflow by way of 5 "powerful" fans, however only two are included (side 120mm LED and rear 120mm).
In what's becoming an increasingly popular trend, the M59 boasts an all-black interior, as well the prerequisite side panel window. Other features include pre-drilled watercooling holes, punched holes in the motherboard tray for quick CPU bracket removal, space for two SSD drives, and "space that is especially designed to fit longer 10-inch cards."
The mid-tower chassis will be available in August at an MSRP of $60.
When we first reviewed the original Logitech G9 (November 2007), we didn’t like it. Specifically, we thought it was uncomfortable to hold, using either of the removable shells. In fact, we described it as “not particularly comfortable for day-to-day mousing” before complaining that it was unsuitable for people who use a traditional palming grip.
We were wrong. After we made a few small adjustments to our grip, we fell in love with the G9—at least when using the grippy palm-friendly Precision body. We still don’t like the smooth grip—dubbed Wide Load—and we’re generally not fans of having to adjust our grip to suit a mouse, but the smooth response and power-gamer-friendly features that the G9x delivers make this mouse the best we’ve ever tested.