TuxHiggs writes: Last month, Forbes wrote that Microsoft was preparing a cross-platform smartwatch with the ability to continuously track your heart rate and sync the data to your devices. A trusted source with knowledge of the development has verified some of that information and provided Tom's Hardware with additional details about the device. The source confirmed previous rumors that the device is cross-platform compatible, and added there would be open APIs as well. The source also confirmed that the display is on the inside of the wrist as opposed to the outside. Design-wise, Microsoft has gone with a slim band design that is said to resemble a thinner, flatter version of the Nike Fuelband. While details about the hardware are scant, the source did reveal that there are 11 sensors under the hood and a mix of chips, including some from TI and Atmel. Finally, the release for this device is apparently set for October.
TuxHiggs writes: Fresh out of MWC 2014 in Barcelona comes a big comms announcement from Intel’s Mobile and Communications Group. The Santa Clara-based chip giant once again has high hopes for the mobile sector, with a number of launches planned for LTE products in 2014.
First up is the XMM 7160 module, which supports 2G, 3G, and 4G LTE across 15 bands. While the 7160 is of course available for gadget manufacturers, it will also be released inside the company’s own M.2 PCI express module for PCs. Intel announced that the 7160 would be arriving in products from Asus, Dell, Lenovo, and Samsung among others.
Intel’s second announcement is the XMM 7260, a global LTE-Advanced solution that the company plans to have shipping in devices by Q2 this year. With over 30 bands on a single global SKU, including bands used by Chinese telecoms, the XMM 7260 could prove to be a very tempting comms solution for device makers.
Intel not only expects to have its products available in every major LTE market by year’s end, but the company also plans to be the second largest provider of LTE technology on the planet.
TuxHiggs writes: Neat article from Tom's Hardware in which they setup 24 Intel SSD DC S3700 drives in a massive RAID 0 (striped) configuration. So what kind of performance does $50,000 worth of top-of-the-line enterprise-grade SSDs working in unison net you? The answer: up to 3,136,000 IOPS, or in excess of 12,000 MB/s. Extraordinary!
TuxHiggs writes: Two days ago, NVIDIA officially unveiled its newest flagship single-GPU consumer graphics card, the GeForce GTX Titan. Being based on the GK110 GPU which powers NVIDIA's latest server boards, the GTX Titan could provide the company with a much-needed boost in general processing performance, where AMD's products have a strong foothold. Today the benchmark results are finally in. Tom's Hardware puts the GTX Titan through a gauntlet of benchmarks versus NVIDIA's dual-GPU GTX 690 and it's single-GPU cousin the GTX 680 (both of which are based on the consumer-oriented GK104 GPU), along with AMD's current flagship, the Radeon HD 7970 GHz Edition. Performance was not only measured in average framerate, but also in consecutive frame latency. This metric reveals brief framerate stutters that adversely effect gameplay which are typically just absorbed into average FPS results. SLI scaling, heat, noise, and power were also examined.
When looking exclusively at gaming performance, the Titan is consistently slower than GTX 690, yet faster than the AMD card. Those of you hoping to see NVIDIA finally gain a foothold against AMD in general processing will be sorely disappointed. While notably improved over the GK104, the GK110 just can't touch AMD's Tahiti in GPGPU tasks. The GTX Titan's heat, noise, and power levels fall right in line with Nvidia's other cards, meaning that it runs cooler and quieter than AMD's card, while consuming less power.
At the end of the day, the Titan is a good fit for users with multiple hi-res displays or Mini-ITX builders. However, at the same price point of $1,000, the dual-GPU GTX 690 is a better fit for the extreme gamer with a case to accommodate it. Meanwhile, two Radeon HD 7970 GHz Editions can be bought for $200 less, making that the smartest buy in high-end graphics today.
TuxHiggs writes: This morning Nvidia introduced not one, but two new Kepler-based graphics cards. At $110, the GeForce GTX 650 is Nvidia's cheapest dual-slot Kepler, holding the same GK107 GPU as the $100 single-slot GeForce GT 640, but with a slightly higher base clock. The real performance boost comes from pairing the 650 with GDDR5 memory, doubling the bandwidth per clock cycle over the 640's DDR3. Testing shows the GTX 650 can go toe-to-toe with AMD's comparably-priced Radeon HD 7750.
At $230, the GeForce GTX 660 is the first Kepler-based card to enter the lower mid-range market. It also debuts the GK106 GPU, which Nvidia claims is fully-utilized — having no disabled resources which could be unlocked later. Performance-wise, the GTX 660 hits a sweet spot in the market, evenly between the AMD Radeon HD 7850 at $210 and the 7870 at $260. Reviews of the GTX 660 are also available from Anandtech, HotHardware, TechReport, and TechSpot.
Tom's Hardware points out the beginning of more naming scheme shenanigans, as Asus is manufacturing an OEM GTX 660 with a stripped down version of the GK104 graphics processor found in more expensive cards, even though the retail GTX 660's center exclusively on the GK106. They also conducted additional memory bandwidth testing on the GTX 660 as well as the GTX 660 Ti (which received varied benchmark results from the press). The testing proves these cards have a disadvantage when enabling 4x MSAA and above. While AMD cards show a steady decline in performance as MSAA is increased, the Nvidia cards take a nosedive starting at 4x MSAA due to the 192-bit memory interface.