Intel had been working to bake in security into the chip, but it seems that effort has drawn to a close with the selloff of its security division into a revamped McAfee company. Now AMD appears to be taking up the idea.
AMD has a cunning plan to push its Zen chips into the Enterprise market on the back of its new Secure Memory Encryption (SME) and Secure Encrypted Virtualisation (SEV) security features.
These new functions will help enterprises protect their databases that run on Zen servers and this could be just the edge required to get AMD back onto the corporate buy list.
This sort of tech is really useful on virtualised servers which are used through cloud hosts. This makes them affordable and flexible compared to hosting on a physical server. The virtual servers adjust accordingly the load it receives and no bandwidth is wasted.
Normally virtual servers are insecure because the data can be hacked, but the SME and SEV features will help servers protect the data.
So far Intel has not come up with any of this sort of function for its processors, despite the fact that was predicted when it wrote a big cheque for McAfee. What we are still waiting for is the information as to how the Zen chips will help consumer gamers who are leaning on discrete GPUs.
Chief technology officer at Oculus, John Carmack, says mobile VR is currently “coasting on novelty” and developers need to be harder on themselves.
According to CNET, Carmack told the assorted throngs at the Oculus Connect event that developers need to pull their socks up and create experiences on par with non-VR applications and games.
“We are coasting on novelty, and the initial wonder of being something people have never seen before But we need to start judging ourselves. Not on a curve, but in an absolute sense. Can you do something in VR that has the same value, or more value, than what these other [non-VR] things have done?”
Carmack moaned about the higher loading times in mobile VR games as a key area in need of improvement. Users should not have to sit through 30 seconds given the brevity of most currently available VR experiences. Although he has clearly never tried to play Total War II whose screens take ages to load.
Still Carmack said that 30 second loads are acceptable if you’re going to sit down and play for an hour “…but in VR initial startup time really is poisonous. If your phone took 30 seconds to unlock every time you wanted to use it. You’d use it a lot less.”
That is true, the daft screen saver/adware/alleged resource saving software makes turning on my wife’s phone an exercise in futility.
He added: “There are apps that I wanted to play, that I thought looked great, that I stopped playing because they had too long of a load time. I would say 20 seconds should be an absolute limit on load times, and even then I’m pushing people to get it much, much lower.”
Memory-chip maker Micron Technology reported better-than-expected fourth quarter revenue, which saw its smallest decline in a year, as pricing improves and the personal computer market starts to make its big comeback.
Micron ‘s accountants said that there would be a first-quarter adjusted profit of 13-21 cents per share, while analysts were expecting a profit of only nine cents. This means that the company ahs made its first profit in three quarters.
The maker of DRAM chips used in PCs and NAND flash memory chips has enjoyed a recovery in the last six months as the PC market picks up.Micron CEO Mark Durcan said:
“We are seeing improving market conditions in terms of both slowing supply growth and improving demand across a number of key segments,”
With DRAM prices rebounding to seven month highs, Micron is seeing the supply glut in the market has dried up following aggressive cut backs in production amid signs of a bounce back in demand.
Prices of both DRAM and NAND chips are expected to rise in the fourth quarter. The latest quarter included a $58 million charge related to a restructuring program the company announced in the third quarter. Micron’s net sales fell 10.6 percent to $3.22 billion in the fourth quarter. Analysts on average were expecting revenue of $3.15 billion.
MediaTek already spilled the beans about the Helio P20 at the Mobile World Congress in February and revealed a few key details. Helio P20 is an octa-core FinFET processor manufactured using the 16nm TSMC process. It has eight Cortex A53 cores clocked up to 2.3GHz. It supports a 24-megapixel camera, full HD resolution 1920×1080 screen and a Cat 6 LTE modem with 2×20 carrier aggregation at 300/50Mbps data speed. We are sure that Helip P20 will put a lot of pressure on the Snapdragon 652 or its sucessor whenever Qualcomm announces it.
The new Mali T880 graphics unit is clocked at a speedy 900MHz and the new SoC might be the world’s first to use the LPDDR4X. MediaTek invested a lot of time improving its MediaTek Imagiq Image Signal Processor (ISP) by adding advanced 12bit Dual ISP supports Bayer and Mono sensors. These elevates picture quality by reducing noise and capturing three times more light than conventional Bayer + Bayer sensors. It also has dual phase-detection autofocus which achieves real-time auto focus that is four times faster than traditional autofocus systems. The 3A HW engine upgrades for more natural, responsive and detailed photographs and powerful multi-scale temporal de-noising technologies which renders videos and photography more accurately, with less noise, even in low light.
The Helio P25 is faster because of its Cortex A53 cores clocked to 2.5GHz. It might have a faster GPU too but this was not revealed. Helio P25.
Helio P10 secured lots of design wins and the Helio P20 / P25 should continue that tradition when it is launched later this year. MediaTek promised Helio P20 in the second half of 2016 and so far, the only phone we could find announced and listed is the Elephone P20. This phone didn’t ship despite the fact we have entered the last quarter of this year.
Manufacturing a phone takes quite some time despite the fact that MediaTek is shipping Helio P20 to manufacturers as we speak.
Intel has announced that it is making 14nm application processors for Spreadtrum and the pair might have Samsung as a client.
Spreadtrum thinks its Intel chips will just be what Samsung wants for its mid-range smartphone series for 2017.
According to Digitimes, Spreadtrum is also contracting TSMC to build chips using the foundry’s 16nm process technology. LG has signed up Intel to make its 10nm process technology for its ARM-based mobile SoCs. LGE is also TSMC’s 16nm and 28nm clients.
Intel only has a workforce of 2,000 people, but it is targeting fabless customers based mainly in China, South Korea, Japan and other countries of Asia. In addition, a team of Intel employees stationed in Shanghai is working with local Chinese manufacturers.
Nvidia has already announced that it will have some major announcements at the CES 2017 and it appears that the GP102-based GTX 1080 Ti could be a part of that keynote.
Earlier in September, Nvidia announced that its co-founder and CEO, Jen-Hsun Huang, will be making a per-show keynote address to talk about AI, self-driving cards, VR and gaming and according to the latest rumor, the launch of new high-end Geforce GTX 1080 Ti could be a part of that keynote.
The latest report comes from Chinese site Zol.com.cn, which suggest that the Nvidia could use that same Consumer Electronics Show 2017 (CES 2017), which kicks off on 5th of January, to launch its newest high-end Geforce GTX 1080 Ti graphics card.
The source also confirms earlier reports that the Geforce GTX 1080 Ti could be based on a cut-down version of the Pascal GP102 GPU, same one behind the Titan X Pascal graphics card. The cut-down version could have 26 enabled SMs (Stream Multiprocessors), which means that the GTX 1080 Ti should end up with 3,328 CUDA cores, 208 TMUs and 96 ROPs.
The GTX 1080 Ti could also be equipped with 12GB of GDDR5X memory, paired up with 384-bit memory interface, hitting 480GB/s of memory bandwidth.
The GPU could be clocked at 1503MHz for the base and 1623MHz for the GPU Boost clock and it will probably end up with the same blower-style cooler as the GTX 1080 graphics card.
There is still no word regarding the price or the availability date but we suspect that Nvidia could have at least a limited supply after the CES 2017 show, while the price could easily reach US $999, unless Nvidia decides to drop the US $1199 Titan X price at the same time.
We will probably see more details and leaks as we draw closer to the CES 2017 show.
TSMC has said it will prop up Moore’s Law and will innovate to keep it going.
TSMC co-CEO Mark Liu told the assorted throngs at a recent event in Hsinchu that his outfit has been mass producing 16nm chips, and is looking to enter volume production of 10nm chips by the end of 2016.
The next plan is to start risk production of 7nm chips in early 2017, and meanwhile has been engaged in the development of 5nm process technology he said.
While TSMC’s R&D for 5nm process continues, a team of the foundry’s 300-400 engineers has already been involved in R&D for 3nm process, Liu claimed.
The company is apparently chatting to academics about how it can develop 2nm process technology. If all comes to fruition, said Liu Moore’s Law will continue to be relevant.
TMSC said that its cunning plan includes smartphones being its main growth momentum. He thinks that smartphones will arrive with more innovative features and with more ICs and sensors than you can point a stick at.
Chipmakers will have to constantly upgrade their technologies for the manufacture of advanced ICs and sensors for phones, Liu said.
High-performance computing (HPC), Internet of Things (IoT) and system integration are also the applications that will drive TSMC’s future growth, Liu added.
Nvidia CEO Jen-Hsun Huang has announced a 2017 processor s codenamed Xavier to succeed Parker.
This is a 16 FinFET SoC with an 8 core custom ARM64 CPU and the next generation 512 Core Volta GPU. Nvidia hinted before that Volta would stick to 16nm manufacturing.
Nvidia plans that Xavier will replace the dual-core Drive PX and can do the same sorts of things with a fraction of the power. It should work with a 20 W TDP envelope.
This new computer vision accelerated SoC will speed up self-driving and can process Dual 8K HDR video streams which will be pretty crucial for the future use planned for the chip.
Design for the ASIL C Fictional safety will sample next year but this is as much as Jen-Hsun wanted to share now It should be able to reach 20 bps DL and 160 SpecINT. With 20W TDP it won’t be heading for any tablet or, god forbid, a phone.
The numbers of VR-enabled smartphones and tablets, as well as shipments of VR devices bundled with gaming consoles or PCs will grow like topsy in the fourth quarter.
Beancounters at Digitimes Research have added up some numbers and divided by their shoe size and reached the conclusion that we should see some significant changes in the VR market soon.
Shipments of VR video-enabled smartphones and VR devices bundled with consoles will be higher compared to other devices. Vendors of VR-enabled tablets and VR headset bundled PCs which niche markets initially before they make headways by coming out with products with reduced prices and enriched content, should do rather well, the Digitimes Report claim.
Gaming and video are still the dominant VR applications in 2016. The successful launch of VR video-enabled flagship smartphones by Samsung Electronics in the first half of 2016 will encourage other vendors to follow suit.
Google and ARM updates to their VR video applications with reduced algorithm requirements in the fourth quarter of 2016 will help develop more VR video-enabled mobile devices.
Shipments of VR video-enabled smartphones are expected to reach 70 million units in 2016, accounting for 5 per cent of global smartphone shipments, Digitimes Research thinks.
Sony is expected to ship over three million PlayStation VR devices in the quarter, far higher than rival vendors.
While Intel has admitted it can’t build a 10nm chip, Mediatek is planning to release two of them using TSMC’s process.
According to the Economic Daily News MediaTek is considering rolling out two versions of its 10nm chips, the Helio X30 for high-end smartphones and the X35 for the lower-end segment.
It said that it will start volume production for the Helio X30-series SoCs as scheduled between the end of 2016 and early-2017. It is also thinking of having another 10nm series designed for mid- and high-end but not necessarily flagship smartphones.
The Helio X35 chips from MediaTek will also be built by TSMC using a lower-spec variant of the foundry’s 10nm processes. It is the first of TSMC’s first group of customers to adopt its 10nm process technology. The other is Apple.
TSMC said that its 10nm process has received product tape-outs from three clients, and will start generating revenues in the first quarter of 2017.
A few of you might remember that we exclusively posted the news that AMD is working on a 7nm CPU codenamed Starship. The 7nm APU is codenamed Gray Hawk and it aims to attain lower TDPs.
The AMD Starship X86 CPU is a 7nm unit with up to 48 cores and 96 threads and this definitely targets the high end server market as well as performance desktop computers. These CPUs will have a range of TDP values from 35W all the way to 180W. It is safe to assume that the version with 35W TDP ends up with much less than 48 cores.
Now AMD plans to launch its first 7nm and target some embedded markets. Of course, there will be a notebook version of a Gray Hawk, possibly with a different codename but AMD plans to use the 7nm quad core with eight threads, in 7nm for casino gaming machines, arcade gaming, industrial control and automation, retail signage, HMI and security machines. It will also fit into the highly profitable medical imaging market, premium thin clients and communication infrastructure.
We already said with that the APU that joins Polaris GPU architecture and 14nm FinFET Zen core is coming in the second half of 2017, and the Gray Hawk is the successor to that.
There is a big chance that this APU will mix with the Navi architecture that is also expected to launch in 7nm. This product is scheduled for a 2019 launch, so we have quite some time before it happens, but it is good to know that AMD is planning far ahead.
The lowest TPD parts will get to 10W, which sounds quite amazing considering what kind of specification that APU might end up having.
The middle of next year is when we expect to see the Zen / Polaris APUs in notebooks and a bit later in embedded systems. AMD’s Lisa Su was clear at Computex earlier this year. She said that the company plans to launch the desktop first, following with server then notebook and last of all t will be a unit aimed at the embedded market.
Bear in mind that these products should still be considered as concepts and they are subject to change. AMD first needs to master a 14nm FinFET low TDP notebook and embedded Zen based parts before it can more to the very exciting 7nm.
AMD Chief Technology Officer Mark Papermaster has told the world that AMD will become the top manufacturer when it comes to PCs and servers.
According to IDG, Papermaster said that the outfit will be making Vega 10 GPU available by first half of 2017. He added that AMD plans to release high-end PCs and servers which will be powered by the new Zen chip and the first Vega 10 GPU.
He thinks that this will gain market share in the gaming, virtual reality, other desktop applications, which will require high-performance GPUs. AMD is going to pitch Zen and Vega 10 GPU (possibly AMD Radeon GTX 490) as being the best of the PC generation. Apparently that positive attitude will give Nvidia and Intel a good kicking.
AMD’s next GPU architecture powered by HBM2, which is proven to increase performance significantly while maintaining power efficiency. HBM2 is also reported to provide maximum throughput of up to 256GBps, thus it is capable of carrying out all existing powerful apps such as virtual reality, 3D rendering and many more.
This leaves the budget and mid-level PCs running Polaris.
Basically this means that AMD is carrying on the same business model it always has done – compete on cost against Nvidia and Intel. That does not mean that the quality is noticeably different, but it does mean that it will always be cheaper.
A Wall Street analyst, with no thought to his personal safety, has dared to question what AMD fans have been telling us for ages – Zen will bring about peace on earth, cure cancer and above all give Intel a good kicking.
However, Christopher Danely with Citigroup has bravely claimed that with its “Kaby Lake” family of processors, Intel remains a “step-function ahead of AMD Zen when Zen chips are released in Q416.”
He also doubts the benchmark stats that AMD presented to promote Zen’s capabilities relative to Intel’s microprocessors claiming that the AMD controlled benchmark compared an engineering sample of a Zen processor that has not been released yet against a three-month old Intel processor, with both chips clocked at 3.0 Ghz.
“We note the maximum speed for the Intel chip is 3.2 GHz. The result showed AMD completing the benchmark 2 per cent faster than Intel, implying higher CPU efficiency on a “clock for clock” basis. AMD kept both 8 core chips at the same clock speed of 3.0 Ghz, below the native clock speed of the Intel chip. The benchmark result showed the Zen Summit Ridge processor completing the Blender rendering benchmark in 48.07 seconds, 2 per cent ahead of Intel Broadwell-E chip’s time of 49.06 seconds. We note this is only one benchmark using a custom workload performed at an AMD event under controlled conditions, and therefore cannot be verified by third parties and does not represent expected results under all workloads, Danely said.
Instead he thinks that Chipzilla will benefit from its process technology lead while AMD’s manufacturing partner, Global Foundries, which has a “spotty track record.”
“After several of delays and eventually failing to develop 20nm and 14nm on its own, GlobalFoundries entered into a partnership agreement with Samsung in April 2014 to adopt Samsung’s 14nm FinFET process. Despite using the same tools, recipes, and materials as Samsung’s 14nm process, products built on GlobalFoundries’ 14nm did not appear until earlier this year, roughly a year after Samsung released its Exynos 7420 SoC built on its 14nm process,” Danely pointed out.
Since the partnership agreement with Samsung does not include 10nm or lower nodes, we think the technology gap between AMD and Intel will widen again once Intel migrates to 10nm next year.
Meanwhile Intel released its new Kaby Lake chips on an improved 14+nm process this month, featuring a 15 per cent performance improvement over its Skylake chips. Kaby Lake chips deliver up to 12 per cent faster productivity performance and 19 per cent faster web performance over comparable Skylake chips
“We expect independent benchmarks to show Intel’ performance is a step function ahead of AMD Zen when Zen chips are released in 4Q16,” he said.
Below you will find lots of rantings from Intel and AMD fanboys and we expect the language to be colourful. Those of a sensitive disposition might want to look away now.
Those who thought that VR would be a cure for cancer will be disappointed to know that sales of the hardware have stalled already.
Earlier stock problems for the Oculus Rift and HTC Vive are now over but no one is buying according to a Steam survey.
We had already suggested that the Rift and Vive would not have mass market appeal, mostly because the hardware requirements were too great. It looks like betting the farm on the tech was not a good idea.
With the Oculus Rift priced at $550 and the HTC Vive at $800, the price is obviously the main problem, with a Steam survey showing that only 0.18 per cent of users own an HTC Vive and only 0.10 per cent have an Oculus Rift.
HTC Vive sales grew only 0.3 per cent in July and were completely flat in August. Oculus Rift sales grew by the same amount in July and just 0.1 per cent in August.
To be fair the Oculus Rift has not officially launched in the UK until September 20, but it is hard to see how it will be more popular in the EU.
This does not mean that the technology is stuffed. It just means that the small number of early adopters who wanted to play with it already have one. Interest from the great unwashed is not happening. This is mostly because there are no games or anything interesting to play with.
This might change with the arrival of the PlayStation Neo console,that is specifically designed with the PlayStation VR headset in mind. Not only will it be cheaper, at £350 it also has significant software support from the rest of the games industry. Still it is an Atlas-like task for Sony to convince the world that VR is a good idea.
AMD has been on the blower to point out that figures from Mercury and Jon Peddie Research, show that it has been growing market share for the fourth consecutive quarter.
A spokesman for AMD said that for the last nine months, AMD has got its mojo back through its Radeon Technologies Group. During that time, the company has made significant investments in hardware, marketing, and software for the graphics line-up leading to four straight quarters of market share growth.
Mercury Research said that AMD gained three points of unit volume share in Q1 2016. The Mercury Research and Jon Peddie Research market share data for Q2 2016 shows AMD seeing its fourth consecutive quarter of desktop discrete GPU share growth, driven by AMD’s strongest quarter of channel GPU sales since 2015 and the commencement of shipping of the next generation Polaris GPUs.
In total discrete graphics, AMD gained 4.8 share points to 34.2 per cent of market by unit volume (based on Mercury Research). In desktop discrete sector, AMD saw a 7.3 share point increase, rising to 29.9 per cent market share.
“This is another positive testament AMD’s strategy is working as the company drives forward towards “Vega” offerings for the enthusiast GPU market, which AMD expects to bring to market in 2017 to complement our current generation of “Polaris” products,” the spokesperson said.
“AMD believes it is well positioned to continue this trend in market share gains with the recently launched Radeon RX 480, 470, and 460 GPUs that bring leadership performance and features to the nearly 85 per cent of enthusiasts who buy a GPU priced between $100 and $300,” she added.