MediaTek’s Octa-Core Processor Tested

MediaTek raised quite a few eyebrows earlier this year when it announced it would build the world’s first proper ARM octa-core, not a big.LITTLE design. The MT6592 has now popped up on a Chinese site, with the first Antutu results.

It scored 25,496, which places it behind the 1.7GHz Snapdragon in the HTC One, but it’s still a lot faster than the Nexus 4’s Qualcomm APQ8064, although throttling may have something to do with that. The score seems too high, but not long after the results emerged, a number of mobile sites started talking about disappointing results, claiming that MediaTek’s octa-core was somehow supposed to end up on a par with Samsung’s latest Exynos 5 big.LITTLE chip and the Qualcomm 800.

This of course is utter rubbish and FUD of the highest order.

The 28nm MT6592 is indeed an octa-core, but it has eight A7 cores, not a combo of A15 and A7 cores. The A7 is about one fifth of the die area of an A15 and according to ARM it consumes one quarter to one fifth of the power, making such comparisons asinine. In other words, MediaTek’s octa-core should end up a lot smaller and cheaper than a quad A15, maybe even a quad A12. That is why we find the 25,496 result hard to believe – it should be less, not more. For example, the Tegra 4 on Shield hits about 36,000, yet it’s a much bigger chip, on a device with more RAM.

The benchmarked chip ran at 1.7GHz, but MediaTek said the MT6592 should have no trouble hitting 2GHz, which could make it faster than a Snapdragon 600. What’s more, the tested device featured 1GB of RAM, 720p display and a Mali-450 GPU, so it is clearly not high-end.

However, the big problem for MediaTek’s curious new SoC is the sheer number of cores. Most apps simply can’t put them to good use and unless MediaTek has a clever trick up its sleeve, the chip might not be nearly as fast in real world applications. It does look promising in benchmarks, though.

Source

Does Wall Street Like Intel’s Mobile Plan?

In recent months Intel’s new CEO Brian Krzanich and President Renee James made several interesting statements, signalling to Wall Street that the chipmaker gets it – it has to do more in mobile.

With smartphone shipments expected to hit one billion per year as early as next year, Intel’s newfound love of mobile chips is hardly surprising. In recent months Intel told the world that it’s now treating Atom just like Core, which means Atom will no longer look like an unwanted stepchild. On the face of it this is good news for shareholders and investors, but scratch the surface it doesn’t look too encouraging.

As a result, most analysts expect Intel to post lacklustre results on Tuesday, which is hardly surprising given the state of the PC market, which is still the bulk of Intel’s core business. Analysts expect revenue of $13.47 billion, 0.1 percent higher year-on-year, but earnings per share are estimated at $0.53, or 8.6 percent down over last year. But negative EPS forecasts aren’t the biggest problem facing Intel. Most analysts agree that 2014 won’t be much better, but there are some factors that indicate even these bleak forecasts might be too optimistic.

The first Bay Trail products are starting to appear and initial performance reports are encouraging, but they are just that – encouraging rather than groundbreaking. Benchmarks seem to indicate that Bay Trail-T tablets end up marginally slower than Qualcomm 800 and Tegra 4 based devices, which are a bit older, too. With prices ranging from $32 to $37, the first batch of Bay Trail chips also cost a bit more than their ARM competitors, but a direct comparison is not possible as ARM players don’t disclose the unit prices of their chips.

Furthermore Intel still lacks integrated LTE support, which means Bay Trail isn’t going to score big phone design wins. Intel hopes to roll out its first LTE enabled products next year, but there’s still some ambiguity. For example, Intel discrete modems are still built on TSMC silicon and it could be a couple of years before they end up on the die of an Intel SoC built in an Intel fab. While Intel could roll out the first two-chip solution next year, it’s highly unlikely that it will have a proper integrated solution before 2015.

This is a bit of a problem for more reasons than one. Many analysts don’t dig deep enough, some of these technical issues go under the radar – so they stick to Intel’s promise of LTE in 2014. Quark is also being overhyped, although it won’t generate any significant revenue over the next few years. Many analysts also believe x86 support is still a big deal, and to some extent it is, but the relevance of x86 is often exaggerated and it is diminishing as we speak. That is why Intel is talking up hybrids, or 2-in-1s – because legacy x86 support is a lot more important for hybrids than regular tablets. In smartphones, x86 support is as useless as a Facebook share button on a porn site.

However, this is where it gets interesting, because Intel is also promising $99 Bay Trail tablets. Back at IDF, Krzanich said Intel’s new tablet platform would “go below $100 by Q4 2013,” giving the impression that Intel can do dirt cheap tablets as well. We are not sure that it can, not unless it subsidizes them with heaps of cash, and we all know how well that went with Ultrabooks.

As for phones, Intel is still dead in the water and this won’t change anytime soon. Apple is quite happy designing its own custom chips and having them built by the lowest bidder. Samsung is going for off-the-shelf IP and manufacturing its Exynos 5 chips in 28nm, and it will hit 20nm soon. Qualcomm dominates the market and Intel can’t erode its lead over the next couple of product cycles. Even if Intel comes up with competitive smartphone chips in a year or two, who will they be for? Apple won’t buy them, neither will Samsung. This would leave Intel in an awkward position of fighting over scraps with heavy hitters like Qualcomm and a range of smaller ARM players like Nvidia and MediaTek.

This is hardly a viable long-term mobile strategy. Intel is basically doing the only thing it can – and doing the only thing that can be done and calling it a strategy doesn’t really make for much of a strategy.

Source

Intel Goes AI

Intel has written a check for the Spanish artificial intelligence technology startup Indisys.

The outfit focuses on natural language recognition and the deal is worth $26 million. It follows Intel’s recent acquisition of Omek, an Israeli startup with specialties in gesture-based interfaces. Indisys employees have joined Intel already. Apparently the deal was signed on May 31 and the deal has been completed.

Intel would not confirm how they are using the tech: “Indisys has a deep background in computational linguistics, artificial intelligence, cognitive science, and machine learning. We are not disclosing any details about how Intel might use the Indisys technologies at this time.”

Source

NOAA Super Computer Goes Live

The National Oceanic and Atmospheric Administration has rolled out two new supercomputers that are expected to improve weather forecasts and perhaps help better prepare us for hurricanes.

The two IBM systems, which are identical clones, will be used by NOAA’s National Weather Service to produce forecast data that’s used in the U.S. and around the world.

One of the supercomputers is in Reston, Va.; the other is in Orlando. The NWS can switch between the two in about six minutes.

Each is a 213-teraflop system running a Linux operating system on Intel processors. The federal government is paying about $20 million a year to operate the leased systems.

“These are the systems that are the origin of all the weather forecasts you see,” said Ben Kyger, director of central operations at the National Centers for Environmental Prediction.

NOAA had previously used identical four-year-old 74-teraflop IBM supercomputers that ran on IBM’s AIX operating system and Power 6 chips.

Before it could activate the new systems, the NWS had to ensure that they produced scientifically accurate results. It had been running the old and new systems in parallel for months, comparing their output.

The NWS has a new hurricane model, which is 15% more accurate in day five of a forecast for a storm’s track and intensity. That model is now operational and running on the new systems. That’s important, because the U.S. is expecting a busy hurricane season.

Source

Intel Launches 530 Series SSD

Intel’s new SSD 530 series was detailed a while ago, but now it is finally official. Intel has formally announced the new drives in three form factors: mSATA, M.2 and 2.5-inch.

The drives will be available in a wide range of capacities, starting at 80GB, through 120GB, 180GB, 240GB, 360GB and 480GB for 2.5-inch drives. As for M.2 and mSATA drives, they will be available in 80GB, 120GB, 180GB and 360GB capacities.

Intel’s new 530 drives are based on 20nm MLC flash and the brains behind the brawn come from LSI, in the form of the SandForce SF-2281 controller. Although transfer speeds will vary depending on capacity, the fastest 530-series drives will deliver read speeds of up to 540MB/s and write at 490MB/s. As for random performance, they boast 41k IOPS in random read and 80k IOPS in random write.

Intel also says the 530-series is its most power efficient storage product to date, which is hardly surprising, but it is good news for notebook vendors who will use mSATA units.

Source

Will Arm/Atom CPUs Replace Xeon/Opteron?

Analyst are saying that smartphone chips could one day replace the Xeon and Opteron processors used in most of the world’s top supercomputers. In a paper in a paper titled “Are mobile processors ready for HPC?” researchers at the Barcelona Supercomputing Center wrote that less expensive chips bumping out faster but higher-priced processors in high-performance systems.

In 1993, the list of the world’s fastest supercomputers, known as the Top500, was dominated by systems based on vector processors. They were nudged out by less expensive RISC processors. RISC chips were eventually replaced by cheaper commodity processors like Intel’s Xeon and AMD Opteron and now mobile chips are likely to take over.

The transitions had a common thread, the researchers wrote: Microprocessors killed the vector supercomputers because they were “significantly cheaper and greener,” the report said. At the moment low-power chips based on designs ARM fit the bill, but Intel is likely to catch up so it is not likely to mean the death of x86.

The report compared Samsung’s 1.7GHz dual-core Exynos 5250, Nvidia’s 1.3GHz quad-core Tegra 3 and Intel’s 2.4GHz quad-core Core i7-2760QM – which is a desktop chip, rather than a server chip. The researchers said they found that ARM processors were more power-efficient on single-core performance than the Intel processor, and that ARM chips can scale effectively in HPC environments. On a multi-core basis, the ARM chips were as efficient as Intel x86 chips at the same clock frequency, but Intel was more efficient at the highest performance level, the researchers said.

Source

Intel Wins Patent Case

Intel has had a previous decision upheld by the US International Trade Commission (ITC), which decided it did not infringe several X2Y chip manufacturing patents.

X2Y had claimed that Intel’s chip manufacturing technology infringed five of its patents and requested an injunction to stop the chips from entering the US. Previously Judge David Shaw ruled that Intel did not infringe X2Y’s patents, a decision that has been upheld by the ITC, which said it will “terminate the investigation with a finding of no violation”.

X2Y’s patents relate to the shielding of electromagnetic interference that is a significant problem in electronics manufacture. The firm has sold licenses to other chipmakers including Samsung, however Intel decided to develop its own techniques, which are now deemed not to infringe X2Y’s patents.

Intel has had a previous decision upheld by the US International Trade Commission (ITC), which decided it did not infringe several X2Y chip manufacturing patents.

X2Y had claimed that Intel’s chip manufacturing technology infringed five of its patents and requested an injunction to stop the chips from entering the US. Previously Judge David Shaw ruled that Intel did not infringe X2Y’s patents, a decision that has been upheld by the ITC, which said it will “terminate the investigation with a finding of no violation”.

X2Y’s patents relate to the shielding of electromagnetic interference that is a significant problem in electronics manufacture. The firm has sold licenses to other chipmakers including Samsung, however Intel decided to develop its own techniques, which are now deemed not to infringe X2Y’s patents.

While Intel had the most to lose, X2Y also went after Apple and HP, two of Intel’s highest profile customers. Intel even argued that an injunction on its chips would hurt American jobs.

Not only did X2Y lose its patent battle with Intel, but two of its patents were ruled invalid. The company had filed two civil lawsuits that were put on hold pending the ITC’s decision, but given the ITC action it will be surprising if the firm continues in its legal campaign against Intel.

Source

Apple Squashes Rumors

Apple will not develop a new, inexpensive iPhone just for the sake of offering a cheaper alternative, Apple CEO Tim Cook said in a speech on Tuesday.

The company’s focus is on creating great products, and it will not make a smartphone that does not past the quality test, Cook said during a webcast from the Goldman Sachs Technology and Internet conference, which is being held in San Francisco.

“There are other companies that do that, that’s not who we are,” Cook said. “Our North Star is great products.”

Instead, the company is now dropping prices on the older iPhone models. That has been successful, and the demand for iPhone 4 models in December was greater than supply, Cook said.

“It surprised us as to the level of demand we have for it,” Cook said.

Lowering the price on older models is just one of the approaches Apple is taking to reach out to price-sensitive buyers. It’s not easy to balance quality and price, and that’s when innovation comes into play and new products could be created to meet consumer demand, Cook said.

“Sometimes you can take the issue … and you can solve it in different ways,” Cook said.

For example, the first iPod that shipped in 2001 was priced at $399, and now users can buy an iPod Shuffle for $49. There was also a big demand in the past to drop the price of Macs to under $500, and Apple tried and couldn’t do it, so it created the iPad tablet.

Source

No LTE On IBM’s SoC Until 2014

Intel is working on integrated LTE modems for its upcoming SoC designs, but CEO Paul Otellini claims they will not be ready for prime time until 2014.

In a recent conference call Otellini was directly asked about Intel’s plans for LTE integration and said “higher levels of integration” are expected next year. He went on to say that the first Intel-based phones with LTE should launch in early 2014, in time for the Mobile World Congress.

Otellini said Intel’s wireless team, formerly a business unit of Infineon, is making good progress in LTE.

“We believe we have a very competitive solution. The Infineon team is known for not necessarily being first to market, but being really good at engineering a very solid solution and being cost effective and cost competitive and I think that they are doing a very good job with respect to this product,” said Otellini.

Source….

Is Intel Really Catching ARM?

A new report suggests that Intel is close to matching ARM on power efficiency.

The study by Bernstein Research analysts said that the days of Intel being mocked because its power-hungry chips shortened the battery life of mobile devices could be over. Bernstein noted that the ARM camp has such a commanding lead in phones and tablets that Intel probably won’t make much of a dent in those markets for a couple of years — even with its energy-efficient chips.

But it said that both company’s chip types “are very close in terms of power efficiency and processing power.” It said that the fight between the ARM and Intel camps will heat up meaningfully as early as 2013, with likely damages on both sides and no winner. For its study, Bernstein compared Intel’s chip in a Motorola RAZR phone and a RAZR phone with an ARM chip. It also compared both chips in similar tablets outfitted with the Windows 8.

The bad news in the report for Intel was that ARM’s chips have become more powerful, making them “a very compelling choice” for consumers looking for low-end notebooks.

Source…

« Previous Page — Next Page »