Author:
Subject: Processors
Manufacturer: ARM

Cortex-A12 Optimized!

ARM is an interesting little company.  Years ago people would have no idea who you are talking about, but now there is a much greater appreciation for the company.  Their PR group is really starting to get the hang of getting their name out.  One thing that ARM does that is significantly different from what other companies do is announce products far in advance of when they will actually be seeing the light of day.  Today they are announcing the Cortex-A17 IP that will ship in 2015.
 
arm_01.jpg
 
ARM really does not have much of a choice in how they announce their technology, primarily because they rely on 3rd parties to actually ship products.  ARM licenses their IP to guys like Samsung, Qualcomm, Ti, NVIDIA, etc. and then wait for them to actually build and ship product.  I guess part of pre-announcing these bits of IP provides a greater push for their partners to actually license that specific IP due to end users and handset makers showing interest?  Whatever the case, it is interesting to see where ARM is heading with their technology.
 
The Cortex-A17 can be viewed as a more supercharged version of the Cortex-A12, but with features missing from that particular product.  The big advancement over the A12 is that the A17 can be utilized in a big.LITTLE configuration with Cortex-A7 IP.  The A17 is more power optimized as well so it can go into a sleep state faster than the A12, and it also features more memory controller tweaks to improve performance while again lowering power consumption.
 
arm_02.jpg
 
In terms of overall performance it gets a pretty big boost as compared to the very latest Cortex-A9r4 designs (such as the Tegra 4i).  Numbers bandied about by ARM show that the A17 is around 60% faster than the A9, and around 40% faster than the A12.  These numbers may or may not jive with real-world experience due to differences in handset and tablet designs, but theoretically speaking they look to be in the ballpark.  The A17 should be close in overall performance to A15 based SOCs.  A15s are shipping now, but they are not as power efficient as what ARM is promising with the A17.
 

Nvidia's renamed Tegra K1 SoC uses Denver and Kepler

Subject: General Tech | January 6, 2014 - 11:08 AM |
Tagged: tegra k1, tegra, SoC, nvidia, kepler, k1, cortex a15, CES, arm, A15

Project X Logan K1 is the first big news out of CES from NVIDIA and represents a bit of a change from what we were expecting.  The current belief was that the SoC would have four 28nm Cortex A15 processors but that will only be one flavour of K1, a Denver based dual core version will also be released.  Those ARMv8 64-bit processors will natively handle 64 bit applications while the A15 version that The Tech Report had taken pictures of will be limited to 32 bit applications, though that will not matter in many mobile applications.   You should also check out Ryan's deep dive into the new Denver and Kepler version here.

die.jpg

"In early 2011, during a CES press event, Nvidia revealed its Project Denver CPU initiative. On Sunday evening, at another CES press conference, the company provided a glimpse of the first Denver-based processor: the Tegra K1. This next-generation SoC features dual Denver CPU cores clocked at up to 2.5GHz. The cores were designed by Nvidia, and they're compatible with the 64-bit ARMv8 instruction set. They have a seven-way superscalar pipeline and a hefty 192KB of L1 cache."

Here is some more Tech News from around the web:

Tech Talk

 

Coverage of CES 2014 is brought to you by AMD!

PC Perspective's CES 2014 coverage is sponsored by AMD.

Follow all of our coverage of the show at http://pcper.com/ces!

Author:
Subject: Mobile
Manufacturer: NVIDIA

Once known as Logan, now known as K1

NVIDIA has bet big on Tegra.  Since the introduction of the SoC's first iteration, that much was clear.  With the industry push to mobile computing and the decreased importance of the classic PC design, developing and gaining traction with a mobile processor was not only an expansion of the company’s portfolio but a critical shift in the mindset of a graphics giant. 

The problem thus far is that while NVIDIA continues to enjoy success in the markets of workstation and consumer discrete graphics, the Tegra line of silicon-on-chip processors has faltered.  Design wins have been tough to come by. Other companies with feet already firmly planted on this side of the hardware fence continue to innovate and seal deals with customers.  Qualcomm is the dominant player for mobile processors with Samsung, MediaTek, and others all fighting for the same customers NVIDIA needs.  While press conferences and releases have been all smiles and sunshine since day one, the truth is that Tegra hasn’t grown at the rate NVIDIA had hoped.

Solid products based on NVIDIA Tegra processors have been released.  The first Google Nexus 7 used the Tegra 3 processor, and was considered the best Android tablet on the market by most, until it was succeeded by the 2013 iteration of the Nexus 7 this year.  Tegra 4 slipped backwards, though – the NVIDIA SHIELD mobile gaming device was the answer for a company eager to show the market they built compelling and relevant hardware.  It has only partially succeeded in that task.

denver2.jpg

With today’s announcement of the Tegra K1, previously known as Logan or Tegra 5, NVIDIA hopes to once again spark a fire under partners and developers, showing them that NVIDIA’s dominance in the graphics fields of the PC has clear benefits to the mobile segment as well.  During a meeting with NVIDIA about Tegra K1, Dan Vivoli, Senior VP of marketing and a 16 year employee, equated the release of the K1 to the original GeForce GPU.  That is a lofty ambition and puts of a lot pressure on the entire Tegra team, not to mention the K1 product itself, to live up to.

Tegra K1 Overview

What we previously knew as Logan or Tegra 5 (and actually it was called Tegra 5 until just a couple of days ago), is now being released as the Tegra K1.  The ‘K’ designation indicated the graphics architecture that powers the SoC, in this case Kepler.  Also, it’s the first one.  So, K1.

The processor of the Tegra K1 look very familiar and include four ARM Cortex-A15 “r3” cores and 2MB of L2 cache with a fifth A15 core used for lower power situations.  This 4+1 design is the same that was introduced with the Tegra 4 processor last year and allows NVIDIA to implement a style of “big.LITTLE” design that is unique.  Some slight modifications to the cores are included with Tegra K1 that improve performance and efficiency, but not by much – the main CPU is very similar to the Tegra 4.

NVIDIA also unveiled late last night that another version of the Tegra K1 that replaces the quad A15 cores with two of the company's custom designs Denver CPU cores.  Project Denver, announced in early 2011, is NVIDIA's attempt at building its own core design based on the ARMv8 64-bit ISA.  This puts this iteration of Tegra K1 on the same level as Apple's A7 and Qualcomm's Krait processors.  When these are finally available in the wild it will be incredibly intriguing to see how well NVIDIA's architects did in the first true CPU design from the GPU giant.

Continue reading about NVIDIA's new Tegra K1 SoC with Kepler-based graphics!

Coverage of CES 2014 is brought to you by AMD!

PC Perspective's CES 2014 coverage is sponsored by AMD.

Follow all of our coverage of the show at http://pcper.com/ces!

Author:
Subject: Processors, Mobile
Manufacturer: ARM
Tagged: t622, mali, cortex, arm, A9, A15, a12

Cortex-A12 fills a gap

Starting off Computex with an interesting announcement, ARM is talking about a new Cortex-A12 core that will attempt to address a performance gap in the SoC ecosystem between the A9 and A15.  In the battle to compete with Krait and Intel's Silvermont architecture due in late 2013, ARM definitely needed to address the separation in performance and efficiency of the A9 and A15. 

arm1.jpg

Source: ARM.  Top to bottom: Cortex-A15, A12, A9 die size estimate

Targeted at mid-range devices that tend to be more cost (and thus die-size) limited, the Cortex-A12 will ship in late 2014 for product sampling and you should begin seeing hardware for sale in early 2015.

arm3.jpg

Architecturally, the changes for the upcoming A12 core revolve around a move to fully out of order dual-issue design including the integrated floating point units.  The execution units are faster and the memory design has been improved but ARM wasn't ready to talk about specifics with me yet; expect that later in the year. 

arm6.jpg

ARM claims this results in a 40% performance gain for the Cortex-A12 over the Cortex-A9, tested in SPECint.  Because product won't even start sampling until late in 2014 we have no way to verify this data yet or to evaluate efficiency claims.  That time lag between announcement and release will also give competitors like Intel, AMD and even Qualcomm time to answer back with potential earlier availability.

Continue reading our overview of the newly announced ARM Cortex-A12 and Mali-T622!!

MWC 12: TI OMAP5 will beat your A9 with its own two cores

Subject: General Tech, Processors, Mobile, Shows and Expos | February 25, 2012 - 04:06 PM |
Tagged: texas instruments, MWC 12, arm, A9, A15

Texas Instruments could not wait until Mobile World Congress to start throwing punches. Despite their recent financial problems resulting in the closure of two fabrication plants TI believes that their product should speak for itself. Texas Instruments recently released a video showing their dual-core OMAP5 processor based on the ARM Cortex-A15 besting a quad-core ARM Cortex-A9 in rendering websites.

Chuck Norris joke.

Even with being at a two core disadvantage the 800 MHz OMAP5 processor was clocked 40 percent slower than the 1.3 GHz Cortex A9. The OMAP5 is said to be able to reach 2.5 GHz if necessary when released commercially.

Certain portions of the video did look a bit fishy however. Firstly, CNet actually loaded quicker on the A9 processor but it idled a bit before advancing to the second page. The A9 could have been stuck loading an object that the OMAP 5 did not have an issue with, but it does seem a bit weird.

About the fishiest part of the video is that the Quad-Core A9, which we assume to be a Tegra 3, is running on Honeycomb where the OMAP5 is running Ice Cream Sandwich.  Ice Cream Sandwich has been much enhanced for performance over Honeycomb.

We have no doubt that the ARM Cortex-A15 will be much improved over the current A9. The issue here is that TI cannot successfully prove that with this demonstration.