Showing posts with label Laser. Show all posts
Showing posts with label Laser. Show all posts

Friday, 11 March 2016

AMD launches XConnect, partners with Intel, Razer to drive desktop gaming on laptops


When it comes to gaming, laptop owners have always been stuck between a rock and a hard place. Thin and light systems often portability and convenience, but limited gaming performance. Larger, desktop-replacement systems offer more than enough firepower for gaming, but are typically heavy and fall into the category of “transportable” rather than portable. The handful of external chassis that have been built by various vendors tend to have their own limitations and often require a reboot when changing the graphics mode, or don’t support laptop hibernation. They’re also typically tied to either a specific product family or even a single product SKU.



AMD wants to offer mobile gamers a more flexible option, and it’s partnered with Razer and Intel to make it possible. XConnect is the name of AMD’s driver support and implementation using Thunderbolt external graphics.

AMD XConnect: External graphics done right?

Since we’re talking about the efforts of three different companies, let me break out which manufacturer contributes what.

AMD wrote drivers for its own Radeon GPUs in order to support an external graphics chassis. Radeon cards running Radeon Software 16.2.2 or later are capable of plug-and-play configuration with an external chassis. The new software can monitor which applications are running on an external GPU and offers the option to close current applications and prep the system for safe removal. Unlike previous solutions, you can connect or disconnect an external dock without rebooting the system.

Razer built the first system to support this feature (the Razer Blade Stealth) and is the first company to adopt the BIOS extensions and capabilities required to allow for PCI Express hot-plugging. Razer also built the first desktop chassis (the Razer Core). Right now, the Razer Blade Stealth and Razer Core are the only compatible products on the market, but that will change with future launches.



Intel developed the Thunderbolt 3 standard that makes the entire system function and specifically developed the external graphics specification that ties these components together. While external Thunderbolt chassis have existed for years, these are not necessarily designed for any graphics card and have many of the restrictions mentioned above regarding both hot swapping and reboot requirements.

All three companies have contributed their work back into the general Thunderbolt eGFX system, which means it should be much easier for third parties and other vendors to create their own solutions. The x4 PCI Express 3.0 connection isn’t as fast as a full desktop slot, no, but it’s enough to drive gaming with an average 10% performance hit depending on the title and the GPU.


While these results have been provided by AMD, they’re also running on an ultrabook at 1440p. That’s strong performance all around.

The ecosystem long-term goal

The mobile graphics industry has never developed an ecosystem that allowed for simple or easy graphics upgrades. While the MXM standard defines a specific pin-out for a graphics card, it does not specify the card size or configuration. As a result, two MXM cards from two different systems may be practically incompatible with each other depending on the laptop’s design.
AMD, Intel, and Razer want to solve this problem by introducing a flexible external standard that bypasses all the heat and thermal issues associated with jamming top-end graphics cards into small enclosed spaces. The Thunderbolt eGFX specification isn’t limited to large chassis. At IDF last year, Intel demoed external graphics solutions with a wide number of breakout ports and capabilities that were designed with far more modest TDP requirements. Our sources have indicated those were all using an early version of Thunderbolt external graphics and were powered by AMD hardware.

If laptop manufacturers and third-party component companies buy into this capability, you’ll be able to purchase a chassis from any vendor and use it with any laptop that offers Thunderbolt eGFX. While the chassis would still cost a non-trivial amount of money ($200 to $300 seems a safe bet), the hardware would be upgradeable and compatible with future systems. Intel will maintain backwards compatibility in any future version of Thunderbolt, which means future systems that utilize a hypothetical Thunderbolt 4 should still be backwards compatible with a Thunderbolt 3 dock.

Desktop GPUs also tend to be less expensive than their laptop counterparts and to offer better performance at any given price point. Over time, the chassis could effectively pay for itself by allowing gamers to use less expensive video cards that offer higher overall performance.

Razer has stated that its dock is also compatible with both AMD and Nvidia GPUs, but Nvidia has stayed fairly quiet about this capability. The company added hot-plug support in a driver around CES, but hasn’t said much since. AMD has stated that its XConnect driver will work with the R9 280, R9 285, R9 290, R9 290X, the R9 300 family, and all Fury and Nano products. It will also be compatible with various AMD mobile GPUs and future Polaris hardware.
As someone who has covered the mobile industry for years and watched video cards vanish from lower-end hardware, I’m genuinely excited to see a GPU solution that can meet mobile gamers’ needs without requiring them to buy desktop replacements. This standard appears to solve all of the problems that previously held the industry back. Will customers bite? That’s another question.







Sunday, 6 December 2015

CERN in a Shoebox? Tiny Particle Accelerators Are Coming


Scientists may soon make particle accelerators that can be the size of a shoebox, experts says.
The project, which is still in its early stage, will use lasers instead of microwaves, to accelerate particles to nearly speed of light.
Applying lasers, "you are able to accelerate particles through short distances to obtain  a higher energy," said Joel England, a researcher from SLAC National Accelerator Laboratory at Menlo Park, California, and one of the main researchers involved in the project.
The earliest type of this technology may probably be used in medical physics and experiments to observe atoms in real time, experts say.If the new method works-out, it may gradually be scaled up to size as that of the world's largest atom smashers, and further decentralize the field of particle physics.

Thinking big as well as small
When talking about particle accelerators,virtually all physicists think big. Giant underground rings that cover country borders. Left-over mineshafts and undergoung structures buried deep beneath the Earth. Tunnels with length long enough to measure from one end of Los Angeles to another. The effective physics needs higher energies, and higher energies need longer distances to ramp up speed.
Such ambitious moves are hugely expensive, and that means only a priviledge few are able to perform  particle-physics experiments. In real life,just around 30,000 accelerators are available worldwide, according to Symmetry Magazine. It sounds like a glut, but there's usually a long queue to use even lower-energy accelerators, England said.Even lower-energy particle accelerators still needs huge expanse of space — a luxury only handful universities can afford, he added.

That's because particle colliders recently rely on microwaves to boost particle energy, England said.Since microwaves have a long wavelength — between 0.04 inches and 39 inches (0.1 centimeters to 100 cm) — meaning they take a longer distance to boost a particle's energy.
Recent microwave amplifiers used in particle accelerators resemble a microwave oven in certain ways, England said.

It encloses a cavity within metal container that you pump microwave power into and it sets up a field in within, England said. "Rather cooking dinner, it produces a different type of field with an electric component placed along the axis so that particles passes through will receive a kick."

Laser accelerator
In recent times, laser technology has improved by leaps and bounds.Since lasers have a shorter wavelength (visible light has a wavelength that's between 400 nanometers and 700 nm), that means laser-driven accelerators could reduce in size significantly.
So, England and a huge group of experts have teamed up to make a laser-driven accelerator that can be mounted onto a microchip. [Infographic: How Do Lasers Work?]
"The main aim is to have all of the components needed to accelerate particles to useful energies,using the similar devices on a single silicon wafer," England told Live Science.

Lasers allow for higher field intensities,since they will not damage the metal cavities like microwaves would. In addition, microfabrication could allow researchers to fit hundreds of accelerators in series on a single wafer, England said.
The team proposes to make a working prototype in five years, and the project just received millions of dollars for funding from the Gordon and Betty Moore Foundation to make that come true. In the short run, the team proposes to generate smaller accelerator energies,similar to hospital radiation machines.

Eventually, the technology can be used to scale down the big colliders, such as the proposed International Linear Collider, a next-generation facility which may be built in Japan that would look for new forms of matter. It can also be used to repurpose existing accelerators, such as SLAC, said Robert Byer, an applied physicist at Stanford University involved in the project.
"If we have a smaller accelerator, we will be able to build a version of SLAC that's just 30 meters [98 feet] long, not 3 km [1.8 miles] long," Byer said. "You will save a lot of funds on tunnels and all civil construction."


Lots of obstacles
Although, reaching that point requires a lot of ingenuity.
As such,the team have not stumbled across source of particles that can be made using silicon wafer technology. This means the team will have to make one, possibly looking in the way of diamonds or silicon to emit electrons. Those electrons would then be ramped upto high speed by a focused laser beam, Byer said.
To such electrons within the highly focused beam required for particle acceleration will most likely need making tiny waveguides into the chip. The researchers also have to look for other ways to mount the laser to the other devices on the chip, Byer said.

New applications
One of the numerous possibilities is in the medical-treatment field, Byer said. Existing medical-radiation devices are enormously sized that may occupy an entire room, and radiation often strays to other parts of the body beyond the tumor.

If medical-radiation devices could be miniaturized to fit into a catheter, doctors could irradiate tumors with higher amount of radiation without damaging nearby tissue, Byer said. To do that, doctors could attach a catheter with a tiny accelerator into the body, and then destroy a tumor with electrons at a sufficiently low energy level that all the radiation would be stopped in the tumor tissue, Byer said.
Shoebox accelerators may also assist to reveal the mysterious inner details of the atom. Lasers can now accelerate bunches of electrons at the attosecond timescale, which is about "the same time it takes an electron to orbit the nucleus of an atom," Byer said.

Using such tiny time slices, "we'll be able to take movies of electrons in the orbits of atoms. We'll be able to watch electrons move to make the bonds."