Two years away after AMD FX model released 9590, came among those from Intel to launch a processor whose operating frequency of 5 GHz threshold.
Success, otherwise worthy of praise, not a word processor for ordinary consumers, but a model Xeon server addresses systems and workstations. Thus, Xeon E5-2,602 V4 is the first Intel which officially exceeds 5 GHz barrier.
Using the manufacturing process on a 14 nm “Broadwell-EP”, the CPU core with plates addressed socket LGA2011v3 has specified a consumption 165W TDP obtained with difficulty and only 4 of the 8 processor cores available has been disabled. Quad-core Xeon E5-2,602 V4 is aided with 10 MB L3 memory cache, quad-channel controller compatible with DDR4 memory modules and HyperThreading technology.
AMD is one of the largest manufacturers of processors on the planet, and through a complex video clip accusing Intel of BAPco and companiaile trisarea in SYSmark benchmark to favor Intel. More specifically, the folks at AMD have demonstrated that although benchmarking program SYSmark difference in performance between laptops with Intel and AMD hardware is 50%, in its own software and PCMark, the difference is less than 10% between the components produced by the two companies. Test the folks at AMD have used a computer equipped with an Intel i5 and one with a AMD FX what has not yet been launched, and the difference between the results offered by PCMark and SYSmark or their own software to AMD is immense. With this difference obtained at AMD accuses Intel and BAPco that cheat in results from benchmark software SYSmark analyzing CPU performance exclusively, while analyzing the video card including PCMark and other components of a computer.
The newest generation of Intel’s processors appears to have some problems, according to some information that appeared recently on the internet. The u.s. giant has admitted that recently that there is a problem they want to solve it as soon as possible. Above all, however, it is important to note that the only application of benchmarking to put such problems so far was Prime95.
As I write the folks at Slashgear, the issue is not one. users might have noticed a few months ago. Then, during the 786k tests, users would be discovered by the random errors of the software. Sometimes they occurred after only a few minutes, aletori, they took place after a couple of hours.
Thus, if you run Prime95 with the maximum number of threads available with the “CpuSupportsFMA3 = 0” to the 768 FFT, the system will give at some point an error. From what it seems, a solution to avoid such problems would be disabling the hyper-treading.
“Intel has identified a bug that can affect the sixth generation Intel Core processors. This problem only occurs under conditions of intensive activity, such as those encountered when you run Prime95 application. In those cases, the CPU might freeze or cause too unexpected behaviour of the system.
The problem is that the industry in the manufacture of transistors and semiconductor products reached the end of Moore’s law (named after Gordon e. Moore, one of the founders of Intel in the 1960s). To understand the phenomenon, as in a year reach the last density (10nm) which makes it possible to manufacture mass densities also processors. In the densest of 10nm chips appears the so-called tunnel effect which consists in conduct contrary to the law in the way atoms of classical mechanics, when a particle, in our case the electron, must pass through a potential barrier and to get out positive when it hit the barrier.