Is RISC-V a risk to ARM $$.
Will any future generations of chips stay with ARM or move to RISC-V? ( S4 even)
will track along
Western Dig, Nvidia On Board with ‘RISC-V,’ So Pay Attention, Says Benchmark
By
Tiernan Ray
Updated Jan. 3, 2018 12:52 p.m. ET
If you like semiconductors, you should really check out this “RISC-V” thing, according to a missive today from Gary Mobley of The Benchmark Company.
RISC-V, in case you don’t know, is the latest incarnation of the “reduced instruction set computing” architecture, devised by Professor David Patterson of U.C. Berkeleyabout 40 years ago. I interviewed Patterson about RISC-V last summer for Barron’s print magazine.
The notion is that by making the “instruction-set architecture,” or ISA, of a chip more like open source, where it can be modified, it is possible to make chips that are tailored to a given need and therefore more efficient and less costly to produce. Some of these qualities are evident in the “Tensor Processing Unit,” or TPU, which Patterson helped devise for Alphabet’s (GOOGL) Google’s machine learning efforts.
Mobley, who plans to hold a conference call about the topic on January 22nd, with the executive director of the RISC-V Foundation, Rick O’Connor, today offered up a summary of the non-profit organization and also the support its gaining by chip makers.
He notes that hard drive and NAND flash chip maker Western Digital (WDC) recently gave its endorsement to RISC-V "as the company has pledged to transition its own consumption of processors to RISC-V."
"Western Digital ships over one billion cores per year, and plans to double that number,” notes Mobley.
"And if all goes according to plan, they will all be based on RISC-V, according to recent statement."
Mobley also points out that graphics chip titan Nvidia (NVDA) "quietly revealed that it’s going to build its next-generation GPU microcontroller on the RISC-V ISA,” which I had also learned independently from Patterson and from a company founded by one of his graduate students, SiFive.
Investors should pay attention, writes Mobley, because there could be “disruptive” effects upon traditional processors, which could ripple through the fortunes of many chip companies:
Any investor interested in learning how adoption of RISC-V stands to disrupt the CISC and RISC processor domains, including discrete processors and/or processor IP (cores and architectures) embedded within simple MCUs as well as advanced ASICs. Additionally, RISC-V stands to disrupt R&D development roadmaps for merchant and captive SoC companies. For example, if Western Digital truly intends to adopt RISC-V in storage products, Marvell (MRVL; Buy; $30 PT) will need to reconsider usage of Arm cores. This could lower the upfront licensing and royalty costs for Marvell; however, it may require a revamping of Marvell’s storage controller design flow. Processor IP companies such as Arm Holdings, Synopsys (SNPS; Hold), Cadence Design (CDNS; Hold), Imagination Tech (NR) and even CEVA, Inc. (CEVA; Buy; $55 PT) could see an impact.
Will track this along.
On the surface... Tensor -TPU is quite an endorsement?
Rick has said many makers want to get rid of MCUs.
RISC-V, from the little reading material has this
Another benefit of RISC-V is that it enables companies to develop a product that is tailored specifically to their workload, so they start with the RISC-V core and can add whatever it is they specifically need, saving both time and money.
Western digital moving a billion cores- or 2.
anybody in the business that can help in understanding please chime in.
Tnks in advance.
Brian Faith's sip of text...
http://blog.quicklogic.com/corporat...a-new-platform-paradigm/#sthash.nXkjQxTZ.dpbs
That high degree of flexibility really means that the SoC developer can create a platform rather than just a one-off product. And that fact has enormous consequences for the lifetime revenue and profitability of the SoC.
Let’s look at this another way. We would go so far as to say that given the high costs and long development times associated with SoCs today, nearly every developer should be thinking in terms of platforms rather than products. An open source ISA, a leading edge but broadly available process node, and embedded FPGA technology are the perfect combination for creating a flexible SoC platform. Once you have such a platform available, you can create many products, each fairly easily, spun out through a combination of software (implemented on the open source ISA) and hardware (implemented in the eFPGA technology on chip) changes. Make the big investment in the platform, and small incremental investments in delivering highly market-tuned products to your end customers. In most cases that will be the way to extract the most value from your company’s intellectual property and market knowledge.
Modifying mask sets to create new products is a time-consuming, expensive, and ultimately obsolete way of thinking. Welcome to the platform-based future that open source ISAs and eFPGA technology are helping to create.
In this blog item, the first time I read it I thought...
B Faith is more eloquent than anything else I have read by him.
SO please consider that we let some time go by, but
what he says would apple to any roadmap that QUIK
has for future next generation devices?
The S4 will be one of the first embodiments of this blog item?
Will track along
RISC-V its a small world...
devised by Professor David Patterson of U.C. Berkeleya
a company founded by one of his graduate students, SiFive.
SiFive Joins FDXcelerator™ Program to Bring RISC-V Core IP to GLOBALFOUNDRIES’ 22FDX® Process Technology
Tweet Share Share
SAN MATEO, Calif. – Nov. 28, 2017 – SiFive announced today that it has joined GLOBALFOUNDRIES’ FDXcelerator™ Partner Program, and will be making RISC-V CPU IP including SiFive’s E31 and E51 RISC-V cores available on GF’s 22FDX® process technology. Based on the open source RISC-V ISA, the SiFive E31 offers embedded chip designers new capabilities in high performance within strict area and power requirements, and the SiFive E51 offers a full 64-bit performance at 32-bit price, power and area.
“As the RISC-V ecosystem continues to grow, SiFive’s leading CPU IP is seeing increased adoption. Our partnership with GF is going to enable an even larger pool of system designers to build on an industry-leading process platform,” said Naveed Sherwani, CEO, SiFive. “SiFive has led the RISC-V ecosystem from early on and we are excited to continue extending RISC-V into new market segments.”
“As members of the RISC-V Foundation, we are excited to see more RISC-V IP offerings made available on our processes,” said Gregg Bartlett, senior vice president of CMOS business at GF. “SiFive’s wide range of cores makes them an ideal partner for our FDXcelerator program.”
GF’s FDXcelerator Program brings together select partners to integrate their products or services into validated, plug-and-play design solutions, giving customers access to a broad set of quality offerings specific to 22FDX technology. The program’s open framework enables members to minimize development time and cost while simultaneously leveraging the inherent power and performance advantages of FDX technology.
Glo Fo- 22FDX- cutting edge for many reasons.
QUIK has early access to this IP
It will have early access to the mRAM memory on the 22 FDX.
It has early access to potentially disruptive RISC-V core.
for its eFPGA or its own roadmap.
They can ALL go together on one SoC.
Its is doggone impressive?
'Hearables' could diagnose disease, if we let them
Technology that is always listening might sound scary, but it could save lives.
James Trew, @itstrew
03.14.18 in Wearables
Poppy Crum is the Chief Scientist at Dolby labs, and no stranger to the conference circuit. Her talk at this year's SXSW -- "A Hearable Future: Sound & Sensory Interface" -- promised to dive into the hidden possibilities that sound and the human ear have to offer technology. Unfortunately, and perhaps ironically, Crum's talk was plagued by audio problems throughout (through no fault of her own).
"The Ear is this incredible hub of insight between our internal state and the external world," Crum told the audience, before having to ask the technician to reduce the reverb on her microphone. Only moments earlier her laptop (and therefore presentation) had died thanks to a technician plugging it into the wrong outlet.
Crum handled the inconvenience deftly, taking a poignant question from an audience member asking if technology could offer her hope -- she was going to be fully deaf in 10 years. Crum said it could "we want to de-stigmatize wearing hearables, we want that" before explaining her goal of helping to democratize the hearing technology space, as six companies currently owning 98 percent of the market "that's not okay." Technical glitches and Crum's elegant handling of them had the audience cheering in support.
"The ear directs the eye" Crum added, talking about how situational awareness is often lead by our hearing, and not by our sight. Footsteps coming up behind us, or our ability to place a sound in 3D space long before we see what's causing it.
Current audio wearables, or, if we must, "hearables," are starting to take advantage of more than just delivering enhanced sound. Companies like Here, Nura Sound and Bragi (among others) have introduced sound augmentation with varied success. But most are still teetering on the edge of audio assistance -- reducing background noise, or adding to the sound we already hear. Crum thinks we can do much more, and with technology that already exists.
But that advantage comes at a price. "The power of the hearable is only realized if we let the device have access and process our personal data," Crum told Engadget after her talk. How tech firms have handled, or protected our personal data, hasn't exactly been a success so far, so we're forced to make the eternal choice between convenience/progress and privacy. "I think we know what to do to protect that data," Crum added, "but it's what we do with the understanding of that data [that's important]," hinting the pay off could be worth it, but it's a long road ahead.
Let's be clear, we're not just talking about better voice recognition, or knowing when to lower the music in our cars to calm our stressful drive. Using just our voices, scientists can predict the onset of multiple sclerosis, diabetes (through physiological changes that affect your vocal tract) and even psychosis (through vocal patterns).
No way? Seems Sci Fi like?
But do you want to let Amazon, Google or Apple be the ones to diagnose you? And have that data in its coffers? My guess is no.
"The ear is a very special place where we can gain some of the richest insight into our bodies and the external world."
Crum's definition of a "hearable" and the insights they can offer goes beyond wireless earbuds though. "Hearable are devices that listen, they don't even need a transducer, they could just listen to your body." This includes one key area of technology slowly but surely invading our living space: virtual assistants.
Right here, we have a technology that, if we let it, could listen to our daily lives, and offer up all sorts of insight: health issues, lifestyle assistance, entertainment recommendations (and enhancement) and more. But letting someone... some thing listen in on our daily lives is probably an adjustment that will take a while to earn the trust it deserves. "They might just know more about us than we know." Crum reminds us. And that's both horrifying and exciting at the same time.
Catch up on the latest news from SXSW 2018 right here.
Commentary, This is a rich audio company, who has no vested reason to say what it has regarding the ear. They have the vision, it is a source separate from QUIK that says the same things or even more....was a LOT of fun to read. How the heck could they diagnose these disease states like DM from voice changes...wow. THis Dolby talk suggests that Voice alone will not be enough- that means at least 2 of the QUIK cores.....BF says when 2 of the cores are needed they win about 75% of the time. We will have great things to read......
will track along.