While we want AMD to do well to balance the Intel and Nvidia empires, it does seem that the outfit cannot get a break. Today it announced that it is letting 500 staff go and will begin another wave of restructuring.
Of course, we predicted this would happen. The company is betting the farm on its coming Zen chip, but this will not appear until next year. Meanwhile it is facing shrinking sales and nearly impossible competition.
Under the restructuring AMD will outsource some IT and application development services. It will give 500 people, or five percent of its staff, their pink slips and P45s. The company will take a charge of $42 million, with $41 million of that recorded in the just-ended third quarter. AMD said it expected savings of about $58 million in 2016 from the restructuring plan.
This is about the same time AMD hopes to clean up with its Zen chips.
AMD said it will cut white-collar jobs and is not shutting or idling any fabricating operations. The jobs will be lost across AMD’s global operations, including Austin, Texas, and company headquarters in Sunnyvale, California. AMD only has 9,700 employees at the end of last year, so 500 is rather a chunk.
AMD reported its second-quarter revenue fell 35 percent from the year-earlier period, claiming that no one wanted to buy PCs.
The company has been shifting to gaming consoles and low-power servers, but it really has not moved fast enough or come up with the sort of “wow” technology which is needed to see off Intel.
MediaTek is quietly building an ecosystem to drive IoT strategy to push its System on Chip shipments across multiple devices.
The fabless chipmaker is signing partnerships with Amazon, Tinitell, Apple, and People Power.
MediaTek is starting to come out of the shadows in the West with its SoC designs. It sees the IoT as a way to push more of its chips.
It has put in a tender to buy power management outfit Richtek Technology to expand its leadership in Power Management Integrated Circuits (PMIC) to strengthen its overall capabilities for the IoT business model. The deal is expected to close in Q2 2016.
It has provided funding to People Power, a user engagement company providing apps, cloud and mobile services for IoT to further accelerate its penetration in the IoT market in both the U.S. and China, develop new IoT products based on its Fabrux and Influx software architecture
Release of two software development kits (SDKs) for Apple HomeKit, the framework in iOS 8 for communicating with and controlling connected accessories in a user’s home.
This is on top of its partnership with Amazon for the latest devices – Amazon Fire TV is powered by MediaTek’s MT8173, a 64-bit quad-core processor and the world’s first multimedia SoC with ARM’s Cortex-A72 cores; Fire HD 8 and Fire HD 10 tablets powered by MT8135, an up to 1.5 GHz quad-core processor, resulting in a fast and fluid user interface, and smooth running HD videos and high frame-rate games.
Chief Marketing Officer, Johan Lodenius said the company’s cunning plan was to innvovate widely available technology that provides integrated connectivity, while investing in and nurturing developers and the maker community to deliver practical yet innovative solutions.
Troubled chipmaker AMD’s has launched its Pro APUs quietly with just one major customer so far, the maker of expensive printer ink HP.
Based on the Godaveri and Carrizo chips, AMD adds its AMD Secure Processor for corporate peace of mind. The new Pro chips include the new AMD Pro A12 chip, which runs at 3.4GHz. All of the new Pro chips are APUs, which mean that they combine both graphics as well as the CPU core. The A12 integrates 12 compute cores (4 CPU cores and 8 GPU cores), based on the Radeon R7 graphics technology running at 800MHz.
What differentiates the new PRO chips from the more conventional models are what AMD calls the AMD Secure Processor, an embedded core that enables the ARM TrustZone secure environment to run on top of the chip. Theoretically, at least, the technology should supply an added layer of security to sensitive apps.
AMD PRO A-Series mobile processors (formerly codenamed “Carrizo PRO”) are aimed at the commercial laptop market. They were made in collaboration with HP, ExactTrak, and Qualcomm. HP is set to flog a few of them in its HP EliteBooks range.
David Bennett, corporate vice president and general manager, Commercial Products, AMD said the AMD PRO processors enable performance, reliability and opportunity for today’s businesses by giving customers choice and affordability to meet their specific business needs.
The AMD PRO A-Series processors are purpose-designed for business, offering long-term value commercial enterprises can depend on including a 24-month longevity commitment, 18-month image stability, commercial-grade quality assurance and available extended OEM warranty support for up to 36 months.
Protection against modern security threats with new enterprise-class security features including Device Guard, Enterprise Data Protection, and Windows Hello biometric authentication.
The AMD PRO A-Series processors are claimed to enable greater management flexibility in a multi-vendor client environment at what AMD calls a business-friendly price.
HP EliteBook G3 705 series pair the PROs with Qualcomm’s SnapdragonTM X5 LTE modem to provide 4G connectivity and location capabilities.
Fram Akiki, senior director of product management at Qualcomm Technologies said that the closer co-operation between AMD, HP, and Qualcomm on the HP EliteBook 705 G3 Series will benefit enterprise users.
The AMD PRO A-Series mobile processors are available today through online resellers and are currently offered on HP EliteBook 705 G3 Series PCs, including HP EliteBook 725, 745 and 755
The HP EliteBook 705 G3 series with the new Pro chips inside them. The business notebook weighs 2.78 pounds and includes 12.5-inch, 14.0-inch and 15.6-inch displays.
The new Pro chips also contain features that were launched with the earlier chips, such as Heterogenous Systems Architecture (HSA 1.0) compliance to allow programmers to more easily program the CPU, as well as an integrated HEVC video decoder.
Over the last few years, competitive gaming has made huge strides, building a massive fanbase, supporting the rise of entire genres of games and attracting vast prize pots for the discipline’s very best. Almost across the board, the phenomenon has also seen its revenues gaining, as new sponsors come on board, including some major household names. Sustaining the rapidity of the growth of eSports is going to be key to its long term success, maintaining momentum and pushing it ever further into the public consciousness.
In order to do that, according to Newzoo, eSports need to learn some lessons from their more traditional athletic counterparts. Right now, the research firm puts a pin in eSports revenues of $2.40 per enthusiast per year, a number which is expected to bring the total revenue for the industry to $275 million for 2015 – a 43 per cent increase on last year. By 2018, the firm expects that per user number to almost double, reaching $4.63.
That’s a decent number, representing very rapid growth, but it pales in comparison to Newzoo’s estimates on the average earning per fan for a sport like Basketball, which represents a $14 per fan revenue – rising to $19 where only the major league NBA is a factor. To catch up to numbers like this is going to take some time, but Newzoo’s research has listed five factors it considers vital to achieving that aim.
Right now, MOBAs are undeniably the king of the eSports scene, and one of the biggest genres in gaming. The king of MOBAs, League of Legends, is the highest earning game in the world, whilst others like Valve’s DOTA 2 are also represent huge audiences and revenues, including the prestigious annual International tournament. Shooters are also still big business here, with Activision Blizzard recently announcing the formation of a new Call of Duty League.
Nonetheless, MOBAs are still the mainstay and if you don’t like them, you’re not going to get too deeply into competitive gaming as a fan. Although their popularity with the athletes is going to make them a difficult genre to shift, Newzoo says that broadening the slate is a key factor to growth.
The major tournaments bring players, and audiences, from all over the world, but it’s often only the very top tier of players who can find themselves a foothold in regular competition. Major territories like the US, South Korea and Europe have some local structure, but again League of Legends stands almost alone in its provision of local infrastructure. By expanding a network of regular leagues and competitions to more countries, eSports stands a much better chance of building a grassroots movement and capturing more fans.
Already a problem very much on the radar of official bodies and players around the world, the introduction of regulation is always a tough transition for any industry. However, when you’re putting up millions of dollars in prize money, you can’t have any grey areas around doping, match fixing and player behaviour at events. These young players are frequently thrust into a very rapid acceleration of lifestyle, fame and responsibility – a heady mixture which can prove to be a damaging influence on many. Just like in other sports, stars need protecting and nurturing – and the competitions careful monitoring – in order for growth to occur without scandal and harm to its stars.
Dishing out the rights to broadcast, promote and profit from eSports is a complex issue. Whilst games like football are worldwide concerns, with media rights a hotly contested and constantly shifting field, nobody owns the games themselves. With eSports, every single aspect of the games being played is a trademark in itself, with its owners understandably keen to protect them. However, with fan promotion such a key part of the sport’s growth, and services like Twitch a massive factor in organic promotion, governing the rights of distribution is only going to become a murkier and more complex business as time goes on. With major TV networks, well used to exclusivity, now starting to show an interest, expect this to become a hot topic.
Conflict between new and old media
That clash of worlds, between the fresh and agile formats of digital user-sourced broadcasting and the old network model is also going to be source of many of its own problems. One or the other, or even both, is going to have to adapt fast for there to be a convivial agreement which betters the industry as a whole. There’s currently considerable pushback from established media against the idea of eSports becoming accepted as a mainstream activity, fuelled in no small part by their audiences themselves, so a lo of attitudes need to change. Add to that the links between these media giants and many of the world’s richest advertisers and you can start to see the problem.
MediaTek has revealed that its latest generation 10 core processor will be targeting neural networks and tge deep learning market.
Nvidia was one of the first to go after this area and Qualcomm is wants ”in” too. There will be a big scrap for what could be a huge market for all of these companies.
Kevin JouSr. Vice President & CTO of MediaTek said.
“Cloud-based computing provides big data for training a neural network, but on a device deep learning enables privacy, instantaneous usability of personalized databases. It can speed up the search for the picture you want. This speeds up the search of your personal data including payments, pictures and everything else that we don’t want to have in the cloud. You can just ask Jennifer Lawrence how smart it was to have the nude pictures in the iCloud.”
Kevin has confirmed that MediaTek is developing the deep learning SDK that will support multi-corps. We have seen that company’s Core Pilot 3.0 scheduler can enable the CPU, GPU, DSP and ISP to work together.
MediaTek’s Chairman and CEO Tsai Ming-kai said that the company has serious IoT and automotive aspirations. You need deep learning to teach a car the difference between a human printed on a piece of paper and the actual human on a street. This is a painful process, but when solved will enable self-driving cars that are promised to hit our streets by 2020, just five years from now.
TSMC has warned that its revenues for the fourth quarter will experience a sequential decline, however its bottom line has been saved by the fact that the US dollar is doing well against the Taiwan currency.
The company has announced that its third-quarter revenues will exceed its guidance given in mid-July, thanks to a more favorable US dollar exchange rate to the NT dollar. However, revenues for the third quarter will be about $6.42 billion. Gross margin and operating margin will still be within the previous guidance of 47-49 per cent.
TSMC’s revenues for the fourth quarter, however, will drop to between $6 billion with profit margin rates similar to the prior quarter’s levels, the company said.
For all of 2015, TSMC expects to post a double-digit increase in revenues as the company guided previously.
The company expected that it would gain more A9 processor orders from Apple Inc next quarter to offset customers’ inventory corrections. However one has to wonder if TSMC is now thinking that its Apple sales are set to take a hit. Its other smartphone chip customers like Qualcomm, MediaTek have been taking a hammering. TSMC will meet its full-year target, but only thanks to the currency move.
Global smartphone shipment growth is expected to slow this year and average 7.9 percent during the next five years, compared with 34 percent growth over the past five years. Sales of all types of devices are to decline by 1 percent this year, Gartner Inc said in a report yesterday.
Beancounters at Digitimes research claim that Mediatek’s moves to buy Richtek Technology will help the analog IC vendor to push its power management (PWM) gear to the smartphone and TV panel sectors.
Lately Richtek has been moving away from its notebook and motherboard segments and into PWM ICs for telecommunication and consumer applications. The idea was to avoid being damaged too much by the slump in the PC market.
Richtek has scored orders from Samsung Electronics for production of entry-level and mid-range smartphones and has also been ramping its PWM solutions to China’s LCD TV panel sector.
MediaTek has also announced its plans to acquire LCD driver IC maker Ilitech it means that the outfit will build up a comprehensive supply chain for the TV industry in China, Digitimes noted.
MediaTek could also use the planned 12-inch joint venture fab to be built by Powerchip Technology, which is the parent company of Ilitek, in Hefei, China.
The JV fab will provide foundry services for LCD driver ICs in 2018-2019.
AMD will expand its socket FM2+ chip lineup with three new parts – the A10-7890K and A8-7690K APUs, and the Athlon X4 880K CPU.
The new parts showed up on the compatibility list of socket FM2+ motherboards by BIOSTAR and it is not clear when they will be in the shops.
The architecture mentioned is “Kaveri,” but the silicon could be “Godavari” which is a Kaveri refresh.
The top of the range will be the A10-7890K, which has CPU clock speeds of 4.10 GHz out of the box. We do not know what the TurboCore frequency will be, but the current A10-7870K offers 3.90 GHz with 4.10 GHz TurboCore. The A8-7690K has a CPU clocks of 3.70 GHz. We are not sure what the iGPU clock speeds of the two chips.
The Athlon X4 880K is the most interesting. It has 4.00 GHz CPU clocks. The Athlon X4 FM2+ series lack integrated graphics that means that they are good for those who will buy discrete GPUs, on the FM2+ platform.
All three chips offer unlocked base-clock multipliers, enabling CPU overclocking.
MediaTek is planning to write a cheque for a 51 per cent stake in analogue ICs Richtek Technology and might even buy the whole company.
The company will offer US$5.94 for each Richtek common share. After completing the tender offer and going through relevant legal procedures, the company will move forward taking over the remaining shares of Richtek. The follow-up acquisition of Richtek shares is expected to complete in the second quarter of 2016.
Ming-Kai Tsai, MediaTek chairman and CEO said that Richtek was is a leader in analogue ICs and provides comprehensive power management solutions to satisfy various customer demand, backed by an experienced management and R&D team.
“We believe, through the deal, the competitive edges of both companies will be leveraged to maximize the platform synergy, strengthen MediaTek in Internet of Things segment and further enhance MediaTek’s competitiveness in the fast-changing and ever-competitive global semiconductor market,” he said.
Richtek chairman Kenneth Tai claimed the two outfits were complementary in power management IP and products which creates a leadership position in this field.
He said that by using MediaTek’s platform leadership, Richtek could optimize power management performance on the system level to enable competitive products for customers and further expand analogue IC offerings to propel the company into its next stage of growth.
The spec for the AMD Athlon X4 880K has leaked onto the internet. This AMD processor is based on Kaveri, but without a GPU.
The Athlon X4 880K Specs appeared in a compatibility listing from motherboard manufacturer Biostar.
The chip, which hasn’t been formally announced yet, will run at 4 GHz and will follow-up the 870K which is yet to be released.
There is no word on how much it will all cost yet, but the 860K has been seen for $100.00, which should give you some idea.
Contract chip-maker Taiwan Semiconductor Manufacturing Company (TSMC) is thinking of generating electricity in-house.
The cunning plan is to install electric generating equipment at its factories or even building its own power plant.
Apparently, the company’s electricity bill will go up by 50 per cent over the next ten years as it moves to more-advanced technologies.
Taiwan is already facing power shortage problems and TSMC is worried that its plans could be stuffed up.
TSMC has asked Taiwan’s Ministry of Economic Affairs (MOEA) and government-owned Taiwan Power Company (Taipower) about the feasibility of building its own power generators and related regulatory matters.
According to Digitimes companies can set up power generating equipment for use at their own factory sites, but the law has to be revised to allow TSMC to build its own power plant.
TSMC previously pointed out that it does not necessarily need nuclear power unless there is an alternative. We really hope that quote does not mean that TSMC is considering going nuclear.
The rumored Helio X30 is real and if you thought that X20 was not enough to see off Snapdragon 820, it looks like the Helio X30 has a much better chance.
All new Helio X20 deca-core has two A72 at 2.5GHz, four A53 at 2.0 and four A53 cores at 1.4 GHz. It has Core pilot 3.0 is a smart scheduler that decides which core gets what task.
This processor has every chance to be faster than Snapdragon 620 from Qualcomm. The Snapdragon 620 comes with four A72 cores at 1.8GHz and four A53 at 1.4 GHz but we are unsure how Helio X20 goes will match up against the Snapdragon 820 with its custom quad Krait cores.
But the the Helio X30 has four A72 cores at 2.5GHz, two A72 clocked at 2GHz, two Cortex A53 clocked at 1.5GHz and two low power A53 at 1GHz. A senior executive from MediaTek told us that not all cores were created equal.
Despite the fact that the word “A53″ on the box looks like “A53″ on the other box, one is optimized for performance and the other for low power. If it is unclear if the A53 based cluster from MediaTek is the same as A53 cluster from Qualcomm.
As you can read at Fudzilla we spent quite some time learning about the potential gains of having three clusters. The X20 can have 30 to 40 percent less power consumption, simply by being smart how it uses all ten cores / three clusters.
With Helio X30 you will gain more performance with six out of ten cores being based on the A72 core. Having ten cores in four clusters raises another question, how efficient will the four cluster approach be versus the three cluster approach?
MediaTek has not officially confirmed or launched the Helio X30, but we expect that this will happen soon. The X30 should be shipping in devices in early 2016. at least this is what we would expect to place it well against the Snapdragon 820.
An open saucy general-purpose graphics processor (GPGPU) has been unveiled at the Hot Chips event.
The GPGPU is relatively crude and is part of another piece of an emerging open-source hardware platform called MIAOW.
Karu Sankaralingam, an associate professor of computer science at the University of Wisconsin-Madison said that an open source hardware platform is emerging that has inherent value
He said that big companies will someday be built using open source hardware, just as multi-billion-dollar Web giants owe their existence to open source software.
He said more people needed to contribute to open source hardware to improve the platform layer so there’s enough for entrepreneurs to build from it.
A 12-person team developed the MIAOW core in 36 months. Their goal was simply to create a functional GPGPU without setting any specific area, frequency, power or performance goals.
The resulting GPGPU uses just 95 instructions and 32 compute units in its current design. It only supports single-precision operations. Students are now adding a graphics pipeline to the design, a job expected to take about six months.
MIAOW compares favourably on several benchmarks to AMD’s latest high-end chip, Tahiti. However, it also falls far short on other benchmarks. Apparently AMD had a quick look at it and said that the designers were not doing anything “too crazy”.
However quite how MIAOW will navigate through the shark infested patent sea is anyone’s guess.
AMD and Nvidia both appear to be certain to get their “14 nm” out next year.
According to TweakTown Nvidia is apparently dotting the “I” and working out where to put in the semi-colons for its Pascal GPU using TSMC’s 16nm FinFet node. AMD rumored has been wining and dining its old chums at GlobalFoundries to use its 14nm process for its Greenland GPU.
Although these sound like different technologies the “14nm and 16nm” is difference how you measure a transistor. The outcome of both 14 and 16 should be a fairly same sized transistor with similar power features. TSMC calls its process 16nm FinFet, while Samsung and GloFo insist on calling it 14nm FinFet.
The dark satanic rumor mill suggests that the Greenland GPU, which has new Arctic Islands family micro-architecture, will have HBM2 memory. There will be up to 32GB of memory available for enthusiast and professional users. Consumer-oriented cards will have eight to 16GB of HBM2 memory. It will also have a new ISA (instruction set architecture).
It makes sense, AMD moved to HBM with its Fury line this year. Nvidia is expected to follow suit in 2016 with cards offering up to 32GB HBM2 as well.
Both Nvidia and AMD are drawn to FinFET which offers 90 percent more density than 28nm. Both will boost the transistors on offer with their next-generation GPUs, with 17 to 18 billion transistors currently being rumored.
AMD is continuing to lose market share to Nvidia, despite the fact that its new best video card, the Fury is out.
AMD always had a get out of jail card when the last GPU market share numbers were out on the basis of it not having released anything. At the time NVidia had 76% of the discrete GPU market. This was when Nvidia’s best card was the GeForce GTX 980.
A lot happened in that time. There was the release of the Titan X in March, and before the GTX 980 Ti in June. AMD had its Hawaii architecture inside of the R9 290X, and the dual-GPU in the form of the R9 295X2. It was expected that the R9 390X might turn AMD’s luck around but that turned out to be another rebrand. Then there was the arrival of the R9 Fury X.
AMD has new products on the market: the R9 Fury X, R9 Fury, R9 390X and a bunch of rebranded 300 series video cards. But according to Mercury Research’s latest data, NVIDIA has jumped from 76% of the discrete GPU market in Q4 2014 to 82 per cent in Q2 2015.
AMD has 18 per cent of the dGPU market share, even after the release of multiple new products.
It is not that the Fury X isn’t selling well, but because of yield problems there will only 30,000 units made over the entire of the year.
AMD also rebranded nearly its entire product stack thus making no reason to buy a R9 390X if you own an R9 290X.
Sure there is 8GB of GDDR5 on board compared to the 4GB offered on most R9 290X cards, but that’s not enough to push someone to upgrade their card.
Tweaktown noted that there was a big issue of the HBM-powered R9 Fury X not really offering any form of performance benefits over the GDDR5-powered GeForce GTX 980 Ti from NVIDIA. The 980 Ti beating the Fury X in some tests which it should not have.
Nvidia has plenty of GM200 GPUs to go around, with countless GTX 980 Ti models from a bunch of AIB partners. There is absolutely no shortage of GTX 980 Ti cards. Even if you wanted to get your paws on a Fury X, AMD has made it difficult.
Now it seems that next year could be a lot worse for AMD. Nvidia will have its GP100 and GP104 out next year powered by Pascal. This will cane AMD’s Fiji architecture. Then Nvidia will swap to 16nm process when its Maxwell architecture is already power efficient. Then there is the move HBM2, where be should see around 1TB/sec memory bandwidth.
All up the future does not look that great for AMD.