- GeForce GTX 1080 Ti at the frequencies of 2544-12208 MHz set 3DMark03 record - Taiwanese vendors will need the courage to create original products on Volta and Vega - Imagination gets rid of the loss-making unit for the development of SoC - AMD will not artificially reduce the capabilities of junior models EPYC -

	xtreview is your : Video card - cpu - memory - Hard drive - power supply unit source

POSTER: computer news   ||   GEFORCE GTX 1080 TI AT THE FREQUENCIES OF 2544-12208 MHZ SET 3DMARK03 RECORD	DATE:2017-05-21
Asustek Computer recently opened the "Pandora's Box", giving all interested enthusiasts a special BIOS for GeForce GTX 1080 Ti graphics cards, which allows to achieve higher frequencies in overclocking - at least, if we are talking about a graphics processor.  The Greek enthusiast $ @ 39 @ got the first place in 3DMark03 now corresponds to the result of 355 858 points. The central processor Core i7-7700K for this was overclocked under liquid nitrogen to 7038 MHz with Hyper-Threading disabled in the motherboard Asus Maximus IX Apex, and the frequency of the video card with a similar cooling method was raised to 2544/12208 MHz.  This is the new absolute world record 3DMark03, it even surpasses the achievements of the more expensive NVIDIA Titan Xp graphics card. Select Rating : Currently 10.00/10 1 2 3 4 5 6 7 8 9 10 48730 Rating: 10.0/10 (18 votes cast) GeForce GTX 1080 Ti at the frequencies of 2544-12208 MHz set 3DMark03 record	Add Comment ( 1 ) GEFORCE GTX 1080 TI AT THE FREQUENCIES OF 2544-12208 MHZ SET 3DMARK03 RECORD

POSTER: computer news   ||    TAIWANESE VENDORS WILL NEED THE COURAGE TO CREATE ORIGINAL PRODUCTS ON VOLTA AND VEGA	DATE:2017-05-21
Computer journalism veteran Nebojsa Novakovic, who has connected with Taiwan and China for the last 25 years, published a article about the prospects for independent Taiwanese video card manufacturers and manufacturers in the light of the release of highly integrated AMD and NVIDIA GPUs with HBM / HBM2 memory.  NVIDIA has not yet pleased fans with video game cards with HBM memory, but it is certain that in 2018 it will release a GeForce gaming adapter on the Volta GPU with HBM memory.  By that time, AMD will be selling graphics cards with GPU Vega and HBM2, which will be the second generation after the GPU Fiji generation of AMD graphics processors with HBM memory on the substrate. Experience with video cards on the Fiji GPU showed that AMD partners can pay attention to cooling systems of adapters.  They can not do anything with the GPU and HBM memory, but they can show the talent for creating original heat removal systems, which, in particular, allow to release a full-length adapter (not shortened) with dimensions for installation in one slot.  This opportunity, by the way, is also open to the Taiwanese developers of cooling systems.   Another possibility for the creativity of AMD and NVIDIA partners can be the creation of gaming dual-processors and even three-processor video cards.  Independent companies on the wave of high integration of the GPU with HBM are quite capable of creating such products before the emergence of the reference solutions of NVIDIA and AMD.  This will save resources for GPU developers and add credibility to Taiwanese vendors.  By the way, the author advises paying attention to the NVIDIA Tesla DGX products.  These are the boxes, which are integrated from the four NVIDIA P100 and V100 GPUs.  Similar in desktop performance and at a more humane price than NVIDIA wants (up to $ 69,000), it is quite possible to find its buyer from among the fans of virtual reality and video with a resolution of 4K / 8K.   Select Rating : Currently 10.00/10 1 2 3 4 5 6 7 8 9 10 48729 Rating: 10.0/10 (15 votes cast) Taiwanese vendors will need the courage to create original products on Volta and Vega	Add Comment ( 1 ) TAIWANESE VENDORS WILL NEED THE COURAGE TO CREATE ORIGINAL PRODUCTS ON VOLTA AND VEGA

POSTER: computer news   ||    IMAGINATION GETS RID OF THE LOSS-MAKING UNIT FOR THE DEVELOPMENT OF SOC	DATE:2017-05-21
In the spring of this year, Apple forced the company Imagination Technologies to ask the classic question: "To be or not to be?".  In two  years, Apple plans to stop paying royalties to Imagination.  Now the developer of gadgets from Cupertino pays Imagination deductions for graphic technologies PowerVR.  This brings Imagination to half of the consolidated revenue and the lion's share of the revenue for the PowerVR cores. If this money stops coming, Imagination will lose money to develop new products. To prepare for this moment, Imagination Technologies has decided to focus on the architecture and technologies of PowerVR and sell everything else, including the MIPS, DSP and special Ensigma computing units, as well as the IMGworks group to develop single-chip highly integrated solutions on their own architectures.  IMGworks, as it became known a few days ago, was the first to go under the knife of optimization. The British company Sondrel has agreed to buy IMGworks for an unnamed sum.  It is an independent developer of custom LSI and other highly integrated solutions.  In the process of absorption, IMGworks engineers will go to work in Sondrel and, thus, increase the staff of this company to more than 250 employees. According to the Sondrel management, the experience of IMGworks engineers will help the company create solutions with advanced integrated graphics.  This is all the more important because graphical cores are beginning to be widely used for solving non-graphical problems.  These are technologies related to artificial intelligence, including autopilots, supercomputing and machine learning.  Also, the demand for products with Internet connection is growing, which allow to more fully recruit the experience of Imagination specialists.   Select Rating : Currently 10.00/10 1 2 3 4 5 6 7 8 9 10 48728 Rating: 10.0/10 (14 votes cast) Imagination gets rid of the loss-making unit for the development of SoC	Add Comment ( 1 ) IMAGINATION GETS RID OF THE LOSS-MAKING UNIT FOR THE DEVELOPMENT OF SOC

POSTER: computer news   ||    AMD WILL NOT ARTIFICIALLY REDUCE THE CAPABILITIES OF JUNIOR MODELS EPYC	DATE:2017-05-21
The main details about the recently announced EPYC server processors (Naples) will be reported next month, but for now we know about their support for eight-channel DDR4 memory with a total capacity of up to 4 TB and 128 lines of PCI Express 3.0.  The functions of information security and reliability enhancement are also given increased attention, because we are talking about server solutions. At the event, for the analysts of the AMD EPYC processors, a separate report was dedicated, and this is a good reason to extend the "AMD days" until Sunday inclusive.  Representatives of the company explained that it is not going to artificially limit the functionality of the younger EPYC processors, as Intel does with inexpensive Xeon models.  The approach somewhat resembles the situation with Ryzen - all processors of this family have a free multiplier, which facilitates their overclocking.  Only in the case of server processors, instead of a free multiplier, all models will be offered an equal number of supported memory channels and PCI Express, as well as equal opportunities in the part of auxiliary functions. Preparing for the announcement of EPYC, which is scheduled for June this year, AMD distributed to customers and partners over five thousand engineering samples of Naples.  The EPYC processors are still refusing to talk about the price policy, but they hint that they will be very competitive.   Select Rating : Currently 10.00/10 1 2 3 4 5 6 7 8 9 10 48727 Rating: 10.0/10 (5 votes cast) AMD will not artificially reduce the capabilities of junior models EPYC	Add Comment ( 1 ) AMD WILL NOT ARTIFICIALLY REDUCE THE CAPABILITIES OF JUNIOR MODELS EPYC

	xtreview is your : Video card - cpu - memory - Hard drive - power supply unit source

computer news computer parts review OLD Forum Downloads New Forum Login Join Articles terms Hardware blog Sitemap myspace proxy Get Freebies