Category: Chips

Good chip engineers are hard to find

FUTURE HORIZONS ARMENIA 2016 Mentor Graphics has offices all over the world but we talked to the head of its Armenian office and she had plenty to say about finding top engineers.

Irina Dumanyan said that that she looks after 150 people in Armenia and also runs an internship programme.

Mentor has a set of different projects which are worked on by people in its offices all around the world.

But its Armenia office was never in the business of hiring cheap labour.

“Engineers are hard to find, that’s why we established the internship programme,” she said. The company puts its internees on live projects.

She said universities needed to get more savvy about what graduates will actually need in the real world.

“The university mentality should be changed and lots more investment is required,” she said.  It’s hard for Mentor to cooperate with the Armenian state university.

Chip industry still suffering from economic crunch

snail-8296a552f7bd1064368205306ff8a3c7c7bdc7c4-s900-c85The chip industry is still in the doldrums and that isn’t going to change anytime soon.

That’s the verdict of Malcolm Penn, CEO and chairman of UK semiconductor analysts Future Horizons.

At a semiconductor conference in London, Penn said that the chip industry is driven by four factors: the economy, fab capacity, unit demand and average selling prices (ASPs).

Penn said: “The economy is in a mess and it’s not getting any better. It’s the ‘wait one more quarter’ syndrome. Nobody reacts when the data is good because they don’t believe it.”

He said that 2016 seemed to be an almost exact replay of 2015. Nobody believes in the numbers any more, he said.

“No-one is spending money. There are no new killer products on the horizon. There is nothing, nothing at all.”

Killer apps can’t be predicted and have always been a surprise, Penn said. The outlook is somewhat grim.

“No one knows how to restart the engines.

Regionally, he said, the picture is also pessimistic. “Japan is a complete disaster. China is on a downward trend. Russia has shot itself in the foot,” he said. “Newly industrialised Asian countries have run out of steam. The overall trend is decidedly bad.”

So what of the future?

Historically, integrated circuits (ICs) have shown a 10 percent growth but, he said, the current trend is low with only a six percent figure in 2016. He said: “PC and smartphone IC shipments are still relatively very small. Unit demand is driven by the economy, and the PC market is as dead as a dodo.”

The figures over the last four years show a steady decline, 9.5 percent growth in 2013, 8.3 percent in 2014, 5.1 percent in 2015 and six percent this year.

He said: “Fab capacity is in the hands of the few. No there is no excess capacity. It takes a year to add new capacity and the lead time has never changed.”

As far as capital expenditure (CapEx) goes, Intel, Samsung and TSMC rule, he said at 60 percent CapEx. That’s not true for Global Foundries (GloFo) which is only nine percent.

He said: “A capacity shortage is waiting to happen. At some point of time there will be a shortage and it will catch everyone by surprise.” And most of the bigger fabs are in earthquake zones.

He’s gloomy about next year too. In 2017 he expects a weak PC and smartphone market.

“The economy is still horrid. There’s no life in the semiconductor business.”

Samsung dumps ASL stake

holland1Chipmaker Samsung is reducing its investment in Dutch chip-equipment supplier ASML.

Samsung acquired three percent of Dutch chip-equipment supplier ASML in 2012 with the idea that the money spent would help the company speed up its R&D.int extreme ultraviolet lithography.

However other companies who made significant investments in ASML for the same reason, including Intel and TSMC and Inte have started walking away. In fact TSMC sold its entire stake last year. Samsung has now confirmed that it’s dumping half of its stake in ASML via a private placement that’s valued at around $681 million.

This means selling off 6.3 million shares in ASML which will account for almost half of its 3 percent stake in the company. It will remain heavily invested in the outfit and there does not appear to be any bad blood between the two companies. It was reported in May Samsung has signed a deal with ASML to purchase the latest extreme ultraviolet equipment for mass production of its 7-nanometer process.

The company is expected to complete installation of this advanced machinery by the first half of 2017, making it the first time Samsung deploys EUV equipment in its chip-making process.

Why Kaby Lake and Zen is Windows 10 only

Windows 10Microsoft raised a few eyebrows when it announced that only Windows 10 will support Intel’s and AMD’s next-generation processor microarchitectures – codenamed Kaby Lake and Zen.

It appears that there are a few features on Kaby Lake and  Zen that require significant updates to Windows 10 to optimally function.

Kaby Lake uses Intel’s Speed Shift technology that make it possible to change power states more quickly than Skylake. Because Kaby Lake can make Speed Shift transitions faster, 7th Gen Core processors based on the architecture can increase and decrease clocks quickly. Speed Shift is hardware enabled but it uses the OS to function properly.

Intel’s Turbo Boost Technology 3.0  with Turbo Boost Max 3.0 technology improves single-thread performance by identifying the fastest core on a particular processor die and prioritising critical workloads for that core. This pushes up the processor’s frequency when needed and  workloads are also directed to the fastest possible core available. Support for that technology needs to be in the operating system.

AMD’s Zen-based processors have fine-grained clock gating with multi-level regions throughout the chip. Zen will bring in newer Simultaneous Multi-Threading technology for AMD chips. Microsoft will  have to make updates to the Windows kernel and system scheduler, which is more involved than a driver update. Vole did something similar to add proper support for Bulldozer-based processors with Windows 7.

So as far as AMD, Microsoft and Intel are concerned getting rid of support for older systems makes perfect sense. You can’t lock these chips into something which was released seven years ago. Windows 8 is similar to Windows 10 but about as popular as the Boston Strangler it is just not worth trying to update.

While corporate customers might like to remain on Windows 7 and incorporate next-gen hardware into their infrastructure, there will not be many of them. Older versions of Windows and alternative operating systems will still install and run on Kaby Lake and Zen, but they just won’t do the cool stuff.

Princeton boffins come up with open source super-chip

mad scientistPrinceton University researchers have emerged from their smoke filled labs with a new open source computer chip that promises to boost the performance of data centres.

Dubbed “Piton” after the metal spikes driven by rock climbers into mountainsides to aid in their ascent the chip was shown off at the Hot Chips conference.

The Princeton researchers designed their chip specifically for massive computing systems. Piton could substantially increase processing speed while slashing energy usage. The chip architecture is scalable — designs can be built that go from a dozen to several thousand cores.

The architecture enables thousands of chips to be connected into a single system containing millions of cores.

David Wentzlaff, a Princeton assistant professor of electrical engineering and associated faculty in the Department of Computer Science said that Piton was based on a new thinking about computer architecture.  It was built specifically for data centers and the cloud.

“The chip we’ve made is among the largest chips ever built in academia and it shows how servers could run far more efficiently and cheaply.”

The current version of the Piton chip measures six millimetres by six millimetres and has 460 million transistors, each of which are as small as 32 nanometres.

The bulk of these transistors are contained in 25 cores. Most personal computer chips have four or eight cores.

In recent years companies and academic institutions have produced chips with many dozens of cores — but the readily scalable architecture of Piton can enable thousands of cores on a single chip with half a billion cores in the data centre, Wentzlaff said.

“What we have with Piton is really a prototype for future commercial server systems that could take advantage of a tremendous number of cores to speed up processing,” Wentzlaff said.

At a data centre, multiple users often run programs that rely on similar operations at the processor level. The Piton chip’s cores can recognise these instances and execute identical instructions consecutively, so that they flow one after another. Doing so can increase energy efficiency by about 20 percent compared to a standard core, the researchers said.

Piton chip parcels out when competing programs access computer memory that exists off of the chip so they do not clog the system. This approach can yield an 18 percent increase in performance compared to conventional means of allocation.

The Piton chip also gains efficiency by its cache memory management. In most designs, cache memory is shared across all of the chip’s cores. But when multiple cores access and modify the cache memory it is less efficient. Piton assigns areas of the cache and specific cores to dedicated applications. The researchers say the system can increase efficiency by 29 percent per chip.

Wentzlaff said. “We’re also happy to give out our design to the world as open source, which has long been commonplace for software, but is almost never done for hardware.”

Arista Networks loses appeal against Cisco’s trade ban

the Cisco kidThe U.S. International Trade Commission (ITC)  has upheld an import ban on Arista Networks ethernet switches which it thinks infringe Cisco’s patents.

The decision follows a complaint from  Cisco filed in December 2014 about the switches, which are used in computer data centres and servers.

In a blog post on the company’s website on Monday, Cisco general counsel Mark Chandler said the import ban was to start on Tuesday. The ITC said Arista infringed three Cisco patents relating to managing and securing communications networks. The ruling excludes the import of Arista’s network devices, including its 7000 series of switches, which generates most of that company’s product revenue. It also prevents the sale of domestic supplies of the imported products.

Arista’s general counsel, Marc Taxay, said the company has redesigned the software in its switches and believes it is in “full compliance” with the ITC’s orders.

“Our primary focus remains the continued supply of non-infringing products to our customers,” he said.

Arista also said it would appeal to the U.S. Court of Appeals for the Federal Circuit.  Arista has not received approval from the ITC for the redesigned products.

 

ARM has a new chip for data centres

Eniac-USarmyPhoto700A few weeks after Japanese company Softbank said it would buy ARM for $32 billion. ARM is expected to announce a design to crash into the server and Internet of Things market.

The chip design is being detailed at the Hot Chips conference in Cupertino, California today.

It looks like ARM has got the nod to build a new supercomputer based on the new chip design which will be installed in Japan. The Post-K supercomputer will be developed by Fujitsu, which dropped SPARC architecture for ARM for high-performance computers. Fujitsu helped ARM develop the new chip.

Post-K will be 50 to 100 times speedier than its predecessor, the K Computer, which is currently the fifth fastest computer in the world.

The new ARM processor design will be based on the 64-bit ARM-v8A architecture and have vector processing extensions called Scalable Vector Extension. Vector processors drove early supercomputers, which then shifted over to less expensive IBM RISC chips in the early 1990s, and on to general-purpose x86 processors, which are in most high-performance servers today.

Goodbye Intel.

Intel reaches Nervana for its intelligence

In_Utero_(Nirvana)_album_coverChipzilla has decided that AI smells a bit like teen spirit and is going to write a cheque for Nervana Systems.

Intel is tsarting to see AI as the next big thing and it is buying in expertise. Nervana which has not been the same since its lead singer killed himself appears to have got itself involved with cutting edge AI reseach [are you sure that is the name Nervana? Ed].

Intel wrote in its bog that the buy out will help develop Intel’s artificial intelligence portfolio and enhance the deep learning performance of Intel Xeon and Intel Xeon Phi processors.

The company is led by former Qualcomm researcher Naveen Rao and has raised $25 million in venture funding and also has a contract to work with In-Q-tel, the U.S. intelligence community’s venture arm.

It has not said how much it paid for the company but investors in Nervana include Global Playground, CME Ventures, Lux Capital, Allen & Co and AME Cloud Ventures so it looks like a large amount of money must have changed hands. It has been rumoured that Intel had to pay $350 million,

 

 

Flash Memory Summit 2016 – Consolidation?

FMS2016_BannerAd_300x250The Flash Memory Summit 16 will be convening at the Santa Clara Convention Center over August 9 -11, 2016. Flash memory is now established as a key technology enabling new designs for many products in the consumer, computer and enterprise markets.

Storage Crossover

The industry is at a critical juncture where the total cost of ownership for flash based SSD’s achieved crossover with hard disk drive equivalents last September as the enterprise storage medium of choice.

The fact that the number of producers is limited has altered the landscape of consumption with some analysts indicating that serious shortages will exist for some time to come. An interesting, but mitigating fact is that most of the analysts are not technical – the ones that we’ve talked to that have a technical bent are not so sanguine about the availability mix. One item that stands in the road to profits is the need for this next generation storage device to not only retain data but do so interactively without losing bits. The unrecoverable bit boogie man is now staring the industry down. The ability to store immense amounts of “ready data” for execution now depends on the technologies ability to reliably retain data.

All Flash Array producers are now entering the “really big data storage array” market – the battle has dropped down to the cost of storage per dollar creating a whole new category of marketing lows. 3D Flash is now so dense that failure modes are now dependent upon being aware of “how and when” the bits were used during the entire lifetime of the device.

Cork, Ireland NVMdurance was the first to understand this phenomenon and is now firmly embedded in their first customer Altera (now Intel). Pure and Nimble Storage are offering their services for their AFAs – seems that leasing AFA memory is a probable in the future of solid state storage. We’re still left reading the indemnification clauses of their contracts.

Poison Pill

Micron Technology filed with the SEC a poison pill last Friday. The buzz is that the company is once again in play. The likely suitor is none other than Intel according to the lead rumor. We will be talking with Micron and Intel at FMS 16 and although they’ll not say anything about what’s going on we’ll at the very least get to look into their pupils while they’re telling us…,

Nvidia releases answer to AMD’s Radeon RX 480

nvidiaNvidia just launched its Pascalish answer to AMD’s Radeon RX 480 mainstream card.

The GeForce GTX 1060 has about half of the resources of Nvidia’s super expensive GeForce GTX 1080 and the outfit claims it’s on par with a previous generation high-end GeForce GTX 980.

It runs on 120W and is a mix of low-power and high-performance. The new GeForce GTX 1060 features a new Pascal derivative GPU that’s somewhat smaller, called the GP106. It has 10 streaming multiprocessors (SM) with a total of 1280, single-precision CUDA cores and eight texture units.

The GeForce GTX 1060 also features six 32-bit memory controllers, for 192-bits in total. GeForce GTX 1060 cards with either 6GB or 3GB of GDDR5 memory will be available and offered performance that just misses the mark set by the pricier AMD Radeon R9 Nano.

The GeForce GTX 1060 has the largest leads over the Radeon RX 480 in the DirectX 11 tests, though the Radeon had a clear edge in OpenCL and managed to pull ahead in  some DirectX 12 tests.

The GeForce GTX 1060, however, consumes significantly less power than the Radeon RX 480 and is quieter too.

All up it means that Nvidia and AMD are squaring up with different offerings for a similar price