Our well-placed industry sources have told us that we should not expect to see the HMB 2.0 based GPUs shipping anytime soon. Nvidia Pascal and AMD Polaris 10 / 11 will stick with GDDR5 memory for the time being.
The 2nd generation High Bandwidth Memory (HBM 2.0) for high-end GPUs might happen in very late Q4 2016 but realistically it probably won’t ship until 2017 in any volume.
The first card that we expect supporting this feature might be the Greenland, a card that AMD might end up calling Vega. Even according Radeon Technology Group’s official GPU roadmap, Vega / Greenland now look like a 2017 product, or at very best, late 2016 card. Nvidia might make the HBM 2.0 version of the Titan card, but we don’t expect to see a Geforce GTX based on Pascal GPU and HBM 2.0 coming to the market this year.
We managed to talk to some of the memory manufactures and they told us that HBM 2.0 is very limited in supply, and limited supply makes things expensive.
It seems that GPUs of 2016, including the new AMD Polaris and the new Geforce, will be stuck with GDDR5 and in best case scenario with GDDR5X from Micron. The word on the street is that both Geforce GTX based on Pascal and AMD/RTG’s Polaris 10 / Ellesmere and Polaris 11 / Baffin might launch at Computex during last days of May or early June 2016.
On Thursday Taiwan Semiconductor Manufacturing Company announced an 18 percent quarterly revenue decline for Q1 2016 from the same timeframe a year ago in Q1 2015. The chip manufacturing giant also announced Q1 2016 net profit of $2 billion USD ($64.78 billion TWD), representing an 8.3 percent quarterly profit decline from the same timeframe a year ago in Q1 2015.
For TSMC, Q1 2016 was marked by a reduction of demand for high-end smartphones, while smartphone demand in China and emerging markets had upward momentum. Beginning Q2 2016 and onward, the company expect to get back onto a growth trajectory and is projected to hit a 5 to 10 percent growth rate in 2016.
“Our 10-nanometer technology development is on track,” said company president and co-CEO Mark Liu during the company’s Q4 2015 earnings call. “We are currently in intensive yield learning mode in our technology development. Our 256-megabit SRAM is yielding well. We expect to complete process and product qualification and begin customer product tape-outs this quarter.”
“Our 7-nanometer technology development progress is on schedule as well. TSMC’s 7 nanometer technology development leverage our 10-nanometer development very effectively. At the same time, TSMC’s 7-nanometer offers a substantial density improvement, performance improvement and power reduction from 10-nanometer.
These two technologies, 10-nanometer and 7-nanometer, will cover a very wide range of applications, including application processors for smartphone, high-end networking, advanced graphics, field-programmable gate arrays, game consoles, wearables and other consumer products.”
In Q1 2016, TSMC reached a gross margin of 44.9 percent, an operating margin of 34.6 percent and a net profit margin of 31.8 percent respectively. Going forward into Q2 2016, the company is expecting revenue between ~$6.65 billion and ~$6.74 billion USD, gross margins between 49 and 51 percent, and operating profit margins between 38.5 and 40.5 percent, respectively.
Chips used for communications and industrial uses represented over 80 percent of TSMC’s revenue in FY 2015. The company was also able to improve its margins by increasing 16-nanometer production, and like many other semiconductor companies, is preparing for an expected upswing sometime in 2017.
In February, a 6.4-magnitude earthquake struck southern Taiwan where TSMC’s 12-inch Fab 14 is located, a current site of 16-nanometer production. The company expected to have a manufacturing impact above 1 percent in the region with a slight reduction in wafer shipments for the quarter.
“Although the February 6 earthquake caused some delay in wafer shipments in the first quarter, we saw business upside resulting from demand increases in mid- and low-end smartphone segments and customer inventory restocking,” said Lora Ho, Senior Vice President and Chief Financial Officer of TSMC.
“We expect our business in the second quarter will benefit from continued inventory restocking and recovery of the delayed shipments from the earthquake.”
In fiscal year 2016, the company will spend between $9 and $10 billion on ramping up the 16-nanometer process node, constructing Fab 15 for 12-inch wafers in Nanjing, China, and beginning commercial production of the 10-nanometer FinFET process at this new facility. Samsung and Intel are also expected to start mass production of 10-nanometer products by the end of 2016.
During its Q4 2015 earnings call, company president and co-CEO Mark Liu stated the company is currently preparing and working on a 7-nanometer process node and plans to begin volume production sometime in 2018. Meanwhile, since January 2015, a separate research and development team at TSMC has been laying the groundwork for a 5-nanometer process which the company expects to bring into commercial production sometime in 1H 2020.
So far in Q1 2016, shipments of 16 and 20-nanometer wafers have accounted for around 23 percent of the company’s total wafer revenues.
Comments Off on Is Tesla Poaching nVidia’s Engineers?
Tesla Motors,’ which has been poaching engineers from Apple and AMD, could be causing a few headaches for Nvidia.
MKM analyst Ian Ing pointed out that Nvidia and Tesla have partnered in machine-learning which is the key to autonomous driving. Nvidia’s own automotive segment grew 80 per cent to $320 million in revenue.
It had been known that Tesla is swiping Apple and AMD engineers, but the difficulty is that it also needs staff from its old chum Nvidia. Ing said that Apple and AMD staff are not as steeped in graphics processing units and machine learning as Nvidia’s staff.
“Although there are widely reportedly headlines that Tesla has been hiring chip architects from Apple and AMD, we note that expertise has been focused more on multi-purpose application processors vs. the GPU accelerators necessary for machine learning,” Ing wrote.
This could either pressure Nvidia to work more closely with Tesla, or it too might lose staff to the carmarker. However that might be a small headache for Nvidia which is doing obscenely well, according to Ing. He is suggesting everyone should buy Nvidia shares.
Comments Off on iPhone SE Goes With Qualcomm Inside
Contrary to our previous reports we got a tip that iPhone SE will continue using Qualcomm modems and not change to Intel.
The tear downs will start happening soon but our sources very close to the matter said with high certainly that all iPhone SE come with an updated Qualcomm modem.
Intel is still in the run but apparently Apple still felt confident to continue using Qualcomm even for this generation of the phone. A few analysts did suggested that iPhone 7 and beyond might get Intel LTE hardware, but not with iPhone SE.
Back in December, when we originally wrote that Intel got the iPhone SE deal, our sources did suggest that Apple can still change its mind if it doesn’t feel that Intel modem is ready. This might be the case, but in the future, we are quite confident that Apple will get a second LTE supplier at some point, just as it did with different manufacturing fabs.
Having two suppliers will drive the cost down, and for Apple every dollar or cent they save of components means millions more in its pocket. Apple claims “LTE up to 50 percent faster than iPhone 5s,” but it doesn’t give a real number. The iPhone 5S uses MDM9615 that was first introduced in 2011. This modem is at the technology range of Cat 4, X5 modem that Qualcomm ships in its entry level SoCs or as an external component.
We will have to wait for the first teardowns to appear as it is not easy to get to “ LTE up to 50 percent faster than iPhone 5s.” You would need a modem that is capable of 225 Mbps and the next of potential candidates for the iPhone SE is the MDM 20nm 9×35. Qualcomm calls this modem X7 these days, it use to call it Gobi back in late 2014 and this is a Cat 6, 300 Mbit per second download and 50 Mbit per second upload capable chip.
The fact that Apple continues the exclusive deal with Qualcomm is bad news for Intel, but we are sure that the team blue will keep working on getting inside of iPhone.
Comments Off on Is The Smartwatch Boom Really A Bust?
The bottom is dropping out of the smart watch industry as VC’s start to realise that the Apple dream is not making many people much dosh.
This week smartwatch maker Pebble CEO Eric Migicovsky blamed VCs for not giving him all the money he needs and laid off a quarter of its workforce.
Only a few years ago, Pebble was the darling of the crowdfunding crowd, having raised over $30 million on Kickstarter. This was when Apple was rumoured to be making one and the Tame Apple Press was claiming they were going to be the next big thing,
When Migicovsky confirmed the layoffs. He implied that VCs are now less keen on funding the dream.
Now Apple, which was said to be the market leader of smartwatches, has dropped the price of the Apple Watch by $50. It is probably not going to upgrade the next one with any serious bells and whistles. It looks like the only people who bought one were Apple’s hard core of fanboys who buy everything that Jobs’ Mob makes regardless of whether they need it.
The IDC sees wearable devices reaching 110 million by the end of 2016 which should be 38.2 percent growth. But it seems that this is not enough.
Fitbit was initially championed as an industry leader but this year saw its stock has been battered in 2016. It appears that Smartwatches haven’t set the market alight. Pebble’s rivals are Apple, Samsung, Motorola, LG and others. It also does not have any other businesses to fall back on.
During Samsung’s 2016 SSD Forum in Japan, the company took the wraps off its first ever ball-grid array (BGA) solid state disk for mobile devices, the PM971. This particular SSD aims to replace module-based M.2 drives in the 2-in-1 hybrid PC market. The company is claiming it will offer improved thermals, up to 10-percent more battery life and a reduction in vertical storage height for OEMs, product designers and system manufacturers.
The Samsung PM971 built using the company’s Photon controller and runs MLC 3D V-NAND (contrary to the picture above, PC Watch claims it is actually 3-bits per cell). The drive will be available in 128GB, 256GB and 512GB storage capacities and will feature sequential reads up to 1,500MB/s, sequential writes up to 600MB/s, random reads up to 190,000 IOPS and random writes up to 150,000 IOPS.In general, SSDs with BGA packaging are considerably smaller than those using the M.2 form factor, and Intel has claimed that using a PCI-E BGA SSD could allow an increase in battery size by around 10-percent compared to using an M.2 2260 SSD (with GPIO using 1.8v power rail instead of 3.3v), lower thermals than M.2 (from BGA ball conduction to motherboard instead of through M.2 mounting screws), and a vertical height savings of 0.5mm to 1.5mm in notebook devices.
The nice thing about BGA SSDs is that they are “complete” storage solutions and integrate NAND flash memory, the NAND controller and DRAM all into a single package. Currently, there are several BGA M.2 form factors being proposed that will make single-chip SSDs a reality sooner than later as the result of a collaboration between HP, Intel, Lenovo, Micron, SanDisk, Seagate and Toshiba. The four BGA SSD packages proposed are Type 1620, Type 2024, Type 2228 and Type 2828, ranging anywhere between 16 x 20 millimeters and 28 x 28 millimeters with up to 2-millimeter vertical height. It is currently unknown whether the Samsung PM971 adopts any of these proposed BGA M.2 standards.
Based on the demonstration at the 2016 Samsung SSD Forum in Japan, the PM971 offers decent performance thanks to a PCI-E 3.0 x4 interface and the company’s new Photon controller. According to the PC Watch website, the drive is physically smaller than an SD card and Samsung expects device manufacturers and OEMs to begin adoption in the second half of 2016 or the first half of 2017.
Comments Off on Will Razer’s External Graphics Box Fail?
We first saw the Razer Core, an external graphics box that connects to a notebook via Thunderbolt 3 port, back at CES 2016 in January, and today, Razer has finally unveiled a bit more details including the price, availability date and compatibility details.
At the GDC 2016 show in San Francisco, Razer has announced that the Core will be ready in April and have a price of US $499. As expected, it has been only validated on Razer Blade Stealth and the newly introduced Razer Blade 2016 Edition notebooks but as it uses Thunderbolt 3 interface, it should be compatible with any other notebook, as long as manufacturer wants it.
With dimensions set at 105 x 353 x 220mm, the Razer Core is reasonably portable. It comes with a 500W PSU and features four USB 3.0 ports, Gigabit Ethernet and Thunderbolt 3 port which is used to connect it to a notebook.
As far as graphics cards support is concerned, Razer says that the Core will work with any AMD Radeon graphics card since Radeon 290 series, including the latest R9 Fury, R9 Nano and Radeon 300 series, as well as pretty much all Nvidia Maxwell GPU based graphics cards since Geforce GTX 750/750 Ti, although we are not sure why would you pair up a US $500 priced box with a US $130 priced graphics cards. The maximum TDP for the graphics card is set at 375W, which means that all dual-GPU solutions are out of the picture, so it will go as far as R9 Fury X or the GTX Titan X.
There aren’t many notebooks that feature a Thunderbolt 3 ports and we have heard before that Thunderbolt 3 might have certain issues with latency, which is probably why other manufacturers like MSI and Alienware, went on with their own proprietary connectors. Of course, Razer probably did the math but we will surely keep a closer eye on it when it ships in April. Both AMD and Nvidia are tweaking their drivers and already have support for external graphics, so it probably will not matter which graphics card you pick.
According to Razer, the Razer Core will be available in April and priced at US $499. Razer is already started taking pre-orders for the Razer Core and offers a US $100 discount in case you buy it with one of their notebooks, Razer Blade 2016 or Blade Stealth.
MediaTek has told Fudzilla that the Helio X25 SoC is not only real, but that it is a “turbo” version of the Helio X20.
Meizu is expected to be one of the first companies to use the X25. Last year it was also the first to use MTK 6795T for its Meizu MX5 phone. In that case the “T” suffix stood for Turbo. This phone was 200 MHz faster than the standard Helio X10 “non T” version.
In 2016 is that MediaTek decided to use the new Helio X25 name because of a commercial arrangement. MediaTek didn’t mention any of the partners, but confirmed that the CPU and GPU will be faster. They did not mention specific clock speeds. Below is a diagram of the Helio X20, and we assume that the first “eXtreme performance” cluster will get a frequency boost, as well as the GPU.
The Helio X25 will not have any architectural changes, it is just a faster version of X20, just like MTK 6795T was faster version of MTK 6795. According to the company, the Helio X25 will be available in May.
This three cluster Helio X25 SoC has real potential and should be one of the most advanced mobile solutions when it hits the market.The first leaked scores of the Helio X20 suggest great performance, but the X25 should have even better scores. There should be a dozen design wins with Helio X20/ X25 and most of them are yet to be announced. There should be a few announcements for the Helio X25 soon, but at least we do know that now there will be a even faster version of three cluster processor.
The global GPU market has fallen by 20 per cent over the last year.
According to Digitimes it fell to less than 30 million units in 2015 and the outfit suffering most was AMD. The largest graphics card player Palit Microsystems, which has several brands including Palit and Galaxy, shipped 6.9-7.1 million graphics cards in 2015, down 10 per cent on year. Asustek Computer shipped 4.5-4.7 million units in 2015, while Colorful shipped 3.9-4.1 million units, and is aiming to raise its shipments by 10 per cent on year in 2016.
Micro-Star International (MSI) enjoyed healthy graphics card shipments at 3.45-3.55 million in 2015, up 15 per cent on year, and EVGA, which has tight partnerships with Nvidia, also saw a significant shipment growth, while Gigabyte suffered from a slight drop on year. Sapphire and PowerColor suffered dramatic drops in shipments in 2015.
There are fears that several of the smaller GPU makers could be forced out of the market after AMD gets its act together with the arrival of Zen and Nvidia’s next-generation GPU architectures launch later in 2016.
Cisco has patched high-impact vulnerabilities in several of its cable modem and residential gateway devices which are popular among those distributed by ISPs to their customers.
The embedded Web server in the Cisco Cable Modem with Digital Voice models DPC2203 and EPC2203 contains a buffer overflow vulnerability that can be exploited remotely without authentication. Apparently all you need to do is send a crafted HTTP requests to the Web server and you could see some arbitrary code execution.
Cisco said that its customers should contact their service providers to ensure that the software version installed on their devices includes the patch for this issue.
The Web-based administration interfaces of the Cisco DPC3941 Wireless Residential Gateway with Digital Voice and Cisco DPC3939B Wireless Residential Voice Gateway are affected by a vulnerability that could lead to information disclosure. An unauthenticated, remote attacker could exploit the flaw by sending a specially crafted HTTP request to an affected device in order to obtain sensitive information from it.
The Cisco Model DPQ3925 8×4 DOCSIS 3.0 Wireless Residential Gateway with EDVA is affected by a separate vulnerability, also triggered by malicious HTTP requests, that could lead to a denial-of-service attack.
Hackers have been hitting modems, routers and other gateway devices, hard lately – especially those distributed by ISPs to their customers. By compromising such devices, attackers can snoop on, hijack or disrupt network traffic or can attack other devices inside local networks.