At Intel , announced during the CES trade show in 2020, the first data on top of video calling X: it is designed for laptops, pc’s and supercomputers), assuming different versions of the — X-LP, X-HP, X-HP, depending on where it is made. The company also announced that it is a dedicated GPU, the DG1for the developers to optimize their apps.
- NUC is 9, Extreme, and Intel preparing a mini-COMPUTER with a chip, a Core i9 and a dedicated GPU
- How to buy a laptop? [Guide & Tips]
The strategy of Intel, it starts in X: a new architecture which promises to adapt itself to every type of computer ranging from a laptop to a giant. Let’s start with the microarchitecture for a laptop: it is called the X-LP is a low-power or low-power consumption and uses approximately 20 watts of power and can reach up to 50 W.
The chip charts, X-LP will keep up with the processors, the Tiger Lake, the 10-nanometer, which will be launched in 2020. Intel promises a performance gain of a “double-digit” in relation to the integrated GPU on its predecessor, the Ice Lake, to the 10th generation.
Intel has put a chip in the performance graph that is similar to the X-LP is a PCIe card and separate the DG1 (Dedicated Graphics of 1, its acronym in English). It has a cooling fan to be tested, no overheating, and it allows developers to optimize their programs for Tiger Lake, and on to the platform, Secure in in general.
The DG1 is going to be distributed only to developers, not to the general public. The aim is to ensure that there are flaws in the programs, which depend heavily on the graphics card, be it a game or a Fine. In fact, the idea of Intel is not to sell video cards are separate, as are the Nvidia and AMD: at the time of the chip is dedicated to their follow-up, the processor in a single package.
In addition, the graphical Implementation will have the support of the so called “dynamic tuning” (tuning, dynamics): the manufacturers will be able to build PCs that make the GPU work faster if the CPU is not under heavy use. It is similar to the technology of SmartShift is found on laptops with an AMD processor and a graphics chip Radeon card.
Intel will have graphics on the Xe-HPC on supercomputers
We also have a microarchitecture that Cl-HP (high-performance for its acronym in English) meant for the desktop, and the Xe-HPC (high performance computing). The latter was able to achieve several teraflops of performance, that is, a trillion operations in a floating point per second, Intel does not disclose the numbers to be more accurate.
The company says that it plans to incorporate the Tests are available upon request at the operating system that the Aurora, which it is developing with the manufacturer, Cray, for the National Laboratory at Argonne for the u.s. Department of Energy of the united states.
The strategy of the Intel for a Gpu dedicated to it in the long run. In the fall of 2017, and she has hired Raja Koduri, who was then the chief architect of the integrated Radeon on AMD. In less than six months later, she got another staff member quite a big thing by Jim Keller, who has worked with Koduri at AMD, and at the time I was in one place to Another.
Then, in 2018, Intel has promised that its first dedicated GPU would reach by the year 2020. She looks like she is going to meet the deadline this time around.
Information: VentureBeat, and The Verge.