Computing and Graphics segment revenue of $472 million increased 9 percent sequentially and 11 percent from Q3 2015. The sequential and year-over-year increases were driven primarily by increased sales of GPUs, offset by decreased sales of client desktop processors and chipsets. The year-over-year increase was also driven by increased sales of client mobile processors.
Operating loss was $66 million, compared with an operating loss of $81 million in Q2 2016 and an operating loss of $181 million in Q3 2015. The sequential improvement was driven primarily by higher GPU revenue, and the year-over-year improvement was primarily due to higher GPU revenue and the absence of an inventory write-down charge.
Client average selling price (ASP) decreased sequentially driven by lower mobile and desktop processor ASPs and was flat year-over-year.
GPU ASP increased sequentially and year-over-year driven by higher channel and professional graphics ASPs.
Enterprise, Embedded and Semi-Custom segment revenue of $835 million increased 41 percent sequentially and 31 percent year-over-year due to higher sales of semi-custom SoCs.
revenue to decrease 18 percent sequentially, plus or minus 3 percent. The midpoint of guidance would result in Q4 2016 revenue increasing approximately 12 percent year-over-year and 2016 revenue increasing 6 percent from 2015.
Operating income was $136 million compared with $84 million in Q2 2016 and $84 million in Q3 2015. The sequential improvement was primarily due to higher revenue and the year-over-year improvement was primarily due to higher revenue and a $24 million IP licensing gain.
All Other category operating loss was $363 million in Q3 2016 compared with $11 million in Q2 2016 and $61 million in Q3 2015. The sequential increase was primarily driven by a $340 million WSA charge.
Or did you miss your own word "BK"?
YBF.... u know what? Get ready for the Big One (Q3)!
Will you be here next week? Let us know!
Along with the release of numerous high-profile game titles, gaming consoles, desktops and laptops, there also comes the battle of the graphics cards. This time, enthusiasts are raving about which graphics chip wins — whether it is Advanced Micro Devices (AMD) Radeon RX 490 or Nvidia's GTX 1080 Ti.
It is important to note that both graphics cards are not officially released yet. Their manufacturers have both been very quiet about the cards' development; however, there have been info leaks that gave tech freaks a peek at their respective specs.
In earlier teasers, AMD has been addressing the Radeon RX 490 as a 4K gaming type of graphics processing unit (GPU). AMD has repeatedly listed, probably by accident, the RX 490 on their own website. This gave their followers a hint that the GPU is actually a project slated to be released in 2016.
On the other hand, there have been few spills too that give a sneak peek to Nvidia's GTX 1080 Ti. in an earlier report, the upcoming Nvidia graphics card specs were leaked by someone who goes by the username Excalibur50 through an Overclock3D forum.
According to the mentioned spill, the Nvidia GTX 1080 Ti comes with a base clock speed of 1,503-megahertz (MHz) and boost clock speed of 1,623 MHz. Meanwhile, the AMD Radeon RX 490 is speculated by Videocardz to have 1,200 MHz base clock speed. The latter's boost clock is unknown. If these are correct, Nvidia wins the clock speed round.
However, AMD evens out with its tad more powerful architecture with 12 teraflops while Nvidia's entry only has 10.
In the battle of memory bandwidths, GTX 1080 Ti is reported to sport a 384 GB per second while RX 490 will have 716.8 GB per second. With several hundreds GB more on its pocket, AMD clearly wins this round.
More on the memory department, GTX 1080 Ti has 12 GB GDDR5, or graphics double data rate type five synchronous random access memory, on a 320-bit bus while RX 490 is expected to have 16 GB HBM2, or high bandwidth memory 2, on a 4,906-bit bus.
Meanwhile, on the thermal design power, Nvidia's GTX 1080 Ti is expected to require 250-watts while the Radeon RX 490 might need a lower 225-watts graphics card power.
As several reports point out, AMD's RX 490 is set to be released in the fourth quarter of 2016, which means anytime soon. Meanwhile, Nvidia's GTX 1080 Ti might have an early 2017 release.
Previously in this series, we saw that Intel (INTC) is moving quickly in the mobile space and is playing its strength—manufacturing. Intel’s Mobile business forms the part of the company’s CCG (Client Computing Group), which comprises processors for PCs and mobile phones. CCG accounted for 54% of Intel’s revenue in fiscal 2Q16.
Intel is a leader in the PC processor space and has 80% market share. Advanced Micro Devices (AMD) is another major player in this market.
As shown in the above graph, Intel’s CCG witnessed revenue declines in five of the past six quarters after its Mobile business was integrated into the segment and PC sales slowed. However, the company should see some improvement in fiscal 3Q16. The segment may report revenue growth due to the replenishment of PC inventory. The company earlier expected a high-single-digit decline in CCG revenue in fiscal 3Q16.
Moreover, the PC market witnessed stabilization in 2Q16 and demand improved in 3Q16 as PC makers stock up inventory for the holiday season.
Shift in technology road map
Other than a slowdown in PC sales, there was one miss from Intel. Although Intel upgrades its processor every year, the company missed an upgrade in its desktop processor between mid-2014 and mid-2015. Kirk Skaugen, former Intel CCG head, confirmed that the company skipped the upgrade to save research and development expense.
This decision delayed the upgrade of Apple’s (AAPL) MacBooks and iMacs, which use Intel’s processors. However, Intel launched the SkyLake desktop processor in August 2015, but it failed to encourage PC sales. In fiscal 3Q16, Intel announced KabyLake, the third processor on the 14nm (nanometer) node.
The Broadwell processor was not built for the desktop, so CCG revenue fell in fiscal 3Q15 despite a seasonally strong quarter.
Intel delays processor launch
However, Intel has delayed the launch of KabyLake from 4Q16 to 1Q17. Digitimes reported that Intel has also delayed the launch of Coffee Lake f
"But, right now, with amd starting to collapse...."
Did it collapse many years ago? LoL
How long does it take for a comany to collapse by itself? 40 yrs? LoL
Do you mean THE BIG ONE?
..."the consensus EPS forecast for the quarter is $-0.02. The reported EPS for the same quarter last year was $-0.19."
...17 percent faster than the equivalent Intel Core i5 chip?
...88% more graphics performance?
Wow! Pay less, get more!!!!
Advanced Micro Devices, Inc. (NASDAQ:AMD) plans to introduce the next-generation Vega architecture before the end of this year, if a report from Fudzilla is to be believed. According to sources, Vega 10 will be announced later this year with High-Bandwidth Memory 2 on board.
A recent leak earlier this month revealed details and specifications of the Vega architecture lineup. At its fullest, the Vega 10 has 64 Compute Units, 16GB HBM2, 512GB/s memory bandwidth, and 12 TeraFLOPs performance. AMD is likely to introduce Vega 10 for the professional market first. A leaked server map showed Vega 10 and dual Vega 10 as solutions for 2017 for professional segments such as HPC, Deep Learning, and Inference.
Vega 10 will be accompanied by Vega 11 in 2017. Details about Vega 11 are not available but it will serve as the big chip in the lineup, that is until Vega 20 is introduced later in 2018. Vega 20 will reportedly feature up to 32GB HBM2, 1TB/s memory bandwidth and 64 Compute Units. The most interesting thing about Vega 20 is that its built on 7nm FinFET but based on Vega architecture. A leak also revealed that Vega 20 is actually Navi’s part.
NVIDIA currently has the HPC market with TITAN X (Pascal). AMD would definitely want to shake up the ground by announcing Vega 10 with 16 HBM2 later this year, even though the actual release be made until the first quarter of 2017, as confirmed by its official roadmap.