Tag: amd

AMD might get lots of Chinese cash

 Photo of China from satellite - Wikimedia CommonsThe dark satanic rumour mill has manufactured a hell on earth yarn which suggests that AMD might be set to expand thanks to wodges of Chinese cash being thrown at it.

Since Intel paid AMD a billion for its anti-trust doings,  AMD’s bottom line has not been that good.

However it is still in a good position to churn out processors and video cards. This would make it a good deal for a buy out.  Some have suggested Samsung, but others Qualcomm.

But there is also one name which is cropping up on the rumour mill a lot more — a Chinese company called BLX IC Design Corp.

The sticking point to any buy out is that it would require the renegotiation of the licence with Intel over the x86 architecture, however an investment by a third company would work. Trade restrictions by the US government could prevent an outright purchase by an institution run by the Chinese government, but the US loves Chinese cash.

BLX has collaborated with AMD in the past, and does not need to buy the company to get what it wants.

The rumour, looked at  by Tom’s Hardware suggests that BLX IC Design could buy a share of AMD .  It controls the manufacturer of microprocessors Loongson Technology (MIPS architecture, family Godson), may make a strategic investment in technologies and products from AMD with the Chinese Academy of Sciences.

The Chinese chip designer could provide enough money to AMD to increase its capacity for research and development. The budget of AMD in research and development for this quarter will be about $ 200 million, well below the historical spending. As the company’s sales are shrinking, will not increase spending in this area, which could jeopardize its future and the long-term survival.

AMD bosses were in Beijing in late January and early February for meetings related to the possible deal.

Armed with Chinese cash, AMD could damage Intel with projects like Zen and K12 and put the fear of god into Nvidia. Its new partners could help it make inroads into the huge Chinese market and provide the Chinese semiconductor industry with much needed patents and R&D.

 

Amazon sells fake AMD CPUs

FireOnTheAmazonPosterAMD chip buyers using Amazon’s service have found themselves buying fake AMD A8-7600s.

When they receive the chips, the CPUs aren’t working and when they looked at the bottom of the chips they twigged that someone was taking older or less expensive CPUs and delidding them.

They were then taking newer CPUs that have more value, delidding those, and then putting the IHS from the new CPUs on the old CPUs. Unless you look at the pins before you buy, you would never spot the difference.

The cores being used are the AM2 Athlon X2 cores.

The scam is exclusive to the AMD A8-7600, and the situation appears to be limited to Amazon in the United Kingdom.

The danger is that this could spread over to other products if the scam is successful. c’t magazine said it would be a doddle to do with Haswell products, switching the IHS of a Celeron with a Core i7, because both products are LGA1150.

Amazon have not commented but AMD is furious and while refusing to confirm the scam it said:

“It is apparent that this isolated incident is not related in any way to AMD’s manufacturing or packaging, however AMD takes any reports of product tampering very seriously. As part of our ongoing efforts to help ensure consumers and businesses are sold only genuine AMD processors, we thoroughly investigate these extremely rare incidents in an effort to determine the source of the altered products, and consider all available legal remedies – including both civil and criminal prosecution – against persons found to have engaged in fraudulent actions affecting AMD products.”

AMD said it’s been on the blower to Amazon and the local enforcement authorities to fix this incident quickly and “ensure that the rigorous quality and reliability standards that AMD is known for are maintained.

“AMD already implements extensive security measures to ensure the authenticity of our products, we are currently evaluating further measures to implement additional security measures for maximum future support,” a spokesperson said.

AMD has a guide listed on its website to help verify the legitimacy of CPUs.

 

AMD rumoured to be up for sale

AMD, SunnyvaleFinancial analysts on Wall Street yesterday gave credence to rumours that Advanced Micro Devices (AMD) is up for sale.
The putative buyer is rumoured to be a Chinese company but there are caveats around such a deal.
For example, AMD’s licence agreement with chip behemoth Intel would fall into desuetude if such a sale was to go ahead.
But the company may be worth something even without that element, given that AMD’s X86 business isn’t that important any more.  Its graphics business continues to do well.
AMD’s share price closed yesterday at $2.70.
On the 20th of January, AMD released its fourth quarter results, showing revenues of $1.24 billion – down 22 percent compared to the same quarter the year before.
Newly sprung CEO Dr Lisa Su said then AMD had made some progress diversifying its business but admitted its PC business faced challenges.
It is facing “channel headwinds” in its computing and graphics segment, she said then.

 

Scots executive leaves AMD

Scottish senior executive John Byrne said he is to leave AMD.
He had been with AMD and formerly ATI for around eight years.
He made the announcement in a Facebook posting, saying: “After eight years there and after the birth of our two beautiful daughters, I just decided it was time to do something else”.

QNAP brings AMD NAS to market

QNAP-tvs873-frontFor a while now AMD has been largely ignored by the makers of NAS x86 gear – who have tended to favour Intel or, more lately,  ARM.

QNAP has become the first vendor to bring an AMD-based x86 NAS to the market and was showing off its wares at CES over the weekend.

Dubbed the TVS-x63 lineup has four, six and eight bay models. Each of them has a four or eight GB of RAM. The 8-bay model also comes with a ‘+’ SKU and a 10GBASE-T NIC pre-installed in the spare PCIe slot. The ‘+’ version comes with either 8 or 16 GB of RAM.

What makes this NAS different from many of the others out there is that QNAP specifies the CPU in the TVS-x63 models as ‘AMD quad-core 2.4 GHz with Radeon Graphics.’ This turns out to be AMD’s GX-424CC SoC.

This 4C/4T Steppe Eagle configuration is based on the Puma microarchitecture and has a TDP of 25 W. The L2 cache is 2 MB in size. The cores run at 2.4 GHz while the integrated Radeon GPU runs at 497 MHz.

This can support DDR3 memory at 1866 MHz. Puma supports out-of-order execution and is expected to turn out a performance similar to Silvermont in the Bay Trail SoCs.

The TVS-x63 has two HDMI outputs to handle multi-media. It supports true 4K output for the UI as well as QvPC. Video playback is restricted to 1080p and the the VCE engine is supported by the firmware, enabling hardware-accelerated transcoding similar to what we saw with the TS-x51 and TS-x53 Pro units that used Quick Sync.

The AMD offering by QNAP is a bit of a surprise and could force a price drop in similar NAS specs in the next few months. This is assuming that Intel’s vendors are beating a path to Intel’s door demanding either a product or a price which matches what AMD has done.

The TVS-863+-8G is expected to retail for $1400 which is really cheap if you take into account that the . 10G port is pre-installed. More basic models cost $1200 for the TVS-863-4G, $1000 for the TVS-663-4G and $800 for the TVS-463-4G.

 

AMD shuns the Internet of Things

1-AMD-s-New-Steamroller-Architecture-to-Bring-Significant-PerformanceWhile Intel is pinning its future on the Internet of Things (IoT), AMD appears to be spurning it as if it were a rabid dog.

Its senior vice president and general manager of the computing and graphics business group John Byrne thinks that it is much wiser to keep pushing into the PC market, which is still a $40-billion-a-year opportunity.

Talking to Venture Beat  he said that AMD has to execute on its upcoming Carrizo family of accelerated processing units (APUs), which will be focused on the mobile computing market. About 300 million PC processors and 90 million graphics chips are sold each year, and Byrne wants AMD to get its fair share of those sales.

Byrne thinks that setting up a chip making operation for the Internet of Things is just an invitation to lose money.

Byrne said while it concentrated on the IoT, Intel it was missing opportunities in the classic PC market.

“There’s still 300 million PCs, still 90 million graphics chips. If I look at Intel, Nvidia, and my revenue, that’s still a $40 billion market — even before you get to the IoT. If you look at the gross margin profile of that business, it’s still significantly more than AMD as a company’s average. There’s still significant market opportunities in the classic PC space,” Byrne said.

He said that AMD still had work to do in the PC chip market. It had to work on its x86 performance, ensuring that each product it bought to market is better x86. There needed to be improvements in graphics, notebooks needed to improve battery life.

Byrne said that it all meant that AMD could push into the commercial market a long more. He pointed out that AMD won the industry’s largest single tender in commercial 18 months ago in India and Elitebook with HP last year.

“Wait until you see the lineup of commercial platforms I have with Carrizo. It allows us to continue to attack that i3, attack that i5 consumer, and really get to penetrate the commercial market space. We’ll attack graphics. That’s going to be my strategy next, he said.

While he said that the Internet of Things is important there are two ways to make cash from it. Intel is concentrating on the silicon inside the wearable. However, that will cost under $10 and not make huge amounts of cash.

“You’re seeing that with Quark and some of the other investments our competitors are making. I’m not in business to lose money. Share and revenue is nice but so is profitability,” he said.

But all of those devices have to be connected and it is those higher end devices that AMD will be targeting.

AMD warns virtual reality is a long way away

 virtual-realityA true virtual reality machine is a lot longer away than many believe, warns AMD.

AMD’s chief gaming scientist Richard Huddy said that getting photorealism in games is impossible in the  current virtual reality hardware.

Talking to DevelopHuddy said that virtual reality was a staggeringly exciting field.

“But hardware companies need to produce something 100 or 200 times more powerful than current hardware if we’re going to get to the stage where we have complete photorealism in virtual reality headsets. It starts with the facts that, for a person with 20:20 vision, they will need a screen with a resolution of about 8k-by-6k to enjoy photorealism.”

This sort of statement flies in the face of many of the claims of the tech press who have been looking at the current generations of virtual headset.

Huddy predicts about 35 million pixels per eye should suffice for VR photorealism. That’s still 75 million pixels, taking us to a 35-fold increase compared to a standard 1080p monitor.

To get an indication about how far you have to go to get close to that Huddy you will need to get a 400x-to-1000x increase in horsepower to engender true, convincing VR photorealism that is indistinguishable from the real world.

 

Nvidia and AMD stuffed by TSMC’s Apple friendship

two-applesNvidia and AMD have had their move to 16nm and 20nm designs hampered by the limited capacity of both nodes at manufacturer TSMC.

According to WCCFTech.com,  AMD GPUs are made by TSMC as are Nvidia’s chips.  But it looks like all TSMC’s capacity has been sucked up by Apple and Samsung.

This is hard on Nvidia which already had to make the chips in its GTX 980 and 970 cards, using the 28nm process instead of the 20nm it wanted. Nvidia thought it was better to skip 20nm and go straight to 16nm for future designs.

AMD wanted to drop from 28nm to 20nm for its new GPUs but hit the same capacity issue which stuffed up the delivery of AMD’s 20nm R9 300 series graphics cards. We expected these in February and March of this year but now they are at least two months behind.

AMD’s Senior Vice President and Chief Technology Officer Mark Papermaster has warned that there would be 20nm and 28nm products in 2015 but no 14nm or 16nm products until 2016.

And the reason is because Apple’s 16nm A9 chip, which is being made by TSMC has priority and what is left is being taken by Samsung which outsources a lot of work to TSMC.

AMD and Nvidia are stuffed. The only other manufacturer with spare 14nm capacity is Intel and it is not very likely that Intel will sell capacity to its rivals.

What this means then is that the world is not getting cutting edge GPU technology from the two top vendors because Apple has a huge control of TSMC and Intel is Intel.

Windows 7 update malware spotted by AMD

Huntsman spider, Wikimedia CommonsSoftware giant Microsoft appears to have despatched an update which behaves like malware to its Windows 7 customers.

Microsoft has confirmed that a recent update, with the catchy title KB 3004394, is causing a range of serious problems and recommends removing it.

It was first flagged by AMD’s Robert Hallock who noticed that the update blocks the installation or update of graphics drivers such as AMD’s new Catalyst Omega. Nvidia users are also reporting difficulty installing GeForce drivers.

Hallock recommended manually uninstalling the update, advice now echoed officially by Microsoft.

However, the update does not just kill off graphics drivers. Microsoft’s Answer Forum has dark mutterings that USB 3.0 drivers are broken and User Account Control prompts have gone haywire. Microsoft has acknowledged that it even prevents the installation of future Windows Updates.

The Windows Defender service has been disabled by the update.

This is the third time in three years Microsoft has issued software and firmware updates to their Xbox platform which have “bricked” the consoles. In August 2014 and April 2013 PC updates caused widespread Blue Screens of Death.

AMD, Intel in tablet spat

tweedledummBitter semiconductor rivals Intel and AMD are set to up the stakes in 2015 with a fresh assault on the tablet market.

Both companies are often seen as the Tweedledum and Tweedledee of the chip market, continually foraying into battles where no one actually gets hurt.

But Digitimes, which is privy to both companies’ future roadmaps, said that they’ll foray out once more in an attempt to capture some slice of the tablet pie – estimated to represent 200 million units in 2015.

Intel has been forced to provide subsidies to companies in an attempt to bolster its rather feeble market share in the tablet and mobile markets.

It will kick off the show by releasing a system-on-a-chip (SoC) device codenamed Cherry Trail, which will be built using a 14 nanometre process and may be with the world as early as March next year.  The chip will have lower power consumption and support Windows and Android operating systems, said the wire.

But AMD is not going to stand still after receiving that SoC on the jaw.  It will introduce an X86 device dubbed Nolan, and an ARM device called Amur in the second half of next year.

Intel Braswell delays all about cost

979583-scroogeIntel has been doing its best to explain why its Braswell chip has been delayed.

For those who came in late, Braswell was supposed to be a 1Q 2015 launch, but it kept being delayed. The latest news is that it will not be seen in the shops until June and August 2015.

Kirk  Skaugen,  who heads up the company’s PC Client Group said the main reason was cost.

Intel has been having problems getting its 14 nanometer manufacturing technology to yield at economically acceptable levels. Although the company describes the current yield rate of its 14 nanometer technology as being in a “healthy range,” Intel indicated that the yields still are not where the prior generation technology was at this stage of its ramp.

Broadwell costs will actually remain higher than those for the 22 nanometer Haswell family of products until the third quarter of 2015.

This is all very tricky considering that Braswell is intended to be a very low-cost part for entry-level desktops and notebooks. While Intel can take a couple of quarters of elevated costs to get Broadwell right for the higher-value segments of the PC business, it has the luxury of waiting until Braswell’s manufacturing costs are lower than last year’s 22 nanometer Bay Trail’s costs for more cost-sensitive PCs.

Intel has been all about platform cost reduction with Bay Trail-M/D and has been working to reduce the platform bill of materials costs for its Bay Trail-M/D products. Since the low-end PC market is focused more on cost than on performance, these Bay Trail products may continue to hold their own against AMD’s newly announced Carrizo-L until Braswell arrives.

AMD executive jumps to Glofo

_66236467_ratroyrimmerJohn Docherty is departing his position as Senior Vice President (SVP) of Manufacturing Operations at AMD to join GlobalFoundries (GloFo) as SVP of Global Operations.

Docherty would appear to be a bit of a loss for AMD.  He has more than 35 years of semiconductor manufacturing experience and while he was at AMD he had hands-on experience with APU, CPU and GPU manufacturing. His CV includes senior positions at Motorola Semiconductor, LSI Agere Systems and ATI.

However his switch to GlobalFoundries might turn out to be good for both AMD and GlobalFoundries, as a former SVP at AMD he knows exactly what GlobalFoundries’ most prominent customer needs.

Docherty could assist with 14nm FinFET production to ensure it arrives at market sooner. AMD will be relying on 14nm silicon from GlobalFoundries in order to counter Nvidia’s upcoming GPU architecture, codenamed ‘Pascal’, based on TSMC’s 16nm process, which is said to be launching in 2016.

In fact, the whole move could be seen as a vertical integration attempt by AMD who founded GlobalFoundries back in 2009 as a separate company. GlobalFoundries announced a partnership with Samsung back in April to collaborate over the 14nm process.

Docherty would also be interested in taking advantage of the poor relations between Samsung and AMD’s archrival Nvidia, particularly if GloFlo and Samsung have any success with 14nm FinFET.

AMD’s R9 390X will put the wind up Nvidia

1-AMD-s-New-Steamroller-Architecture-to-Bring-Significant-PerformanceFigures leaked to the great unwashed by deep throats within AMD show that its next-gen flagship gaming graphics card will really put the wind up its envious competitor, Nvidia.

Dubbed the R9 390X, the card has numbers which make Nvidia’s Maxwell architecture-based GeForce GTX 980 and GTX 970 look a bit weak.

Let’s be clear, the GTX 980 and GTX 970 are damn fine cards – they are both faster and more power efficient than their predecessors, a tough act for AMD to follow. AMD was already behind in the power consumption stakes with its R9 290X performing well, but consuming much more power than the GeForce GTX 980.

It appears that AMD has cracked the high power consumption of its previous generation graphics cards.

Leaked benchmarks claim to show that a yet unknown graphics card is over 15 per cent faster than Nvidia’s GeForce GTX 980, yet consumes only 12W more power on average.

Several websites, including WCCFTech, claim that the style of the leaked slides is the same as those that appeared for several previous GPU launches too, so there is some credibility to the results as well.

This means that AMD could well be launching a stunning graphics card early in 2015 which will give Nvidia a good kicking.

It certainly needs to do something. Nvidia is in charge of the above $300 market, with its GeForce GTX 970 and GTX 980 out-performing AMD’s equivalents both in terms of speed and power efficiency. Nvidia is expected to launch its GTX 960 soon too, which will further cement its dominance a little lower down the price range.

 

ARM fails to dent X86 server market

intel_log_reversedBeancounters at mighty chip behemoth Intel can stop playing with their worry beads as it looks as though servers based on ARM technology are failing to dent X86 server business.

A report in Taiwanese wire Digitimes said that ARM has made serious attempts to invade the server business but hasn’t succeeding in storming the Intel fortress.

And with Intel having an 80 percent share in the PC market, shareholders in the chip giant believe that despite its appalling performance in the mobile space, it will continue to make high margins from its server chip offerings.

AMD is waiting in the wings but doesn’t have a great deal of traction in the server business,  the report claims.

Both ARM and Intel hope to make vast profits by being in the vanguard in offering products that will leverage the expected boom in the “internet of things”.

Intel and ARM are relying on cloud based apps to make everything work together.  These things are only a tiny fraction of the internet of things, however, and it’s hard to see either company having much of a share in the expected bonanza.

AMD introduces Carrizo SOCs

AMD's John ByrneIntel rival AMD said it has added two system on a chip (SoC) devices to its semiconductor roadmap.

The Carrizo and the Carrizo-L are being positioned as the firm’s answer to the mobile market.

The chips will support Microsoft DirectX 12, OpenCL 2.0, AMD’s Mantle and Freesync and support for Windows 10, when that emerges next year.

AMD senior VP John Byrne said his company is building on its existing intellectual property portfolio.

He said “our goal is to improve APU energy efficiency by a factor of 25 times by 2020”, and said the company would work with the latest industry standards.

The Carrizo microprocessor combines an X86 CPU core called Excavator with its next generation Radeon graphics in what AMD claims will be the world’s first heterogeneous system architecture compliant SoC.

The chips will ship in the first half of next year, Byrne said.