Slashdot Mirror

Operating frequency, ie 3ghz is not the same thing as faster. Processor speed at completing tasks is a complicated mix of frequency, IPC and several other factors. Even at 200mhz faster the Qualacomm chip could still be significantly faster.

ARM collaborates not only with Samsung.

From https://www.design-reuse.com/n...

"ARM and TSMC announced a multi-year agreement to collaborate on a 7nm FinFET process technology which includes a design solution for future low-power, high-performance compute SoCs"

And on the frequency side of things ...

From https://www.anandtech.com/show...

"With the CLN5, TSMC will also offer an Extremely Low Threshold Voltage (ELTV) option that will enable its clients to increase frequencies of their chips by 25% "

Taking 2.8GHz as baseline, a 25% increase of frequency would land the new chip at 3.5GHz.

While it is true that the new chip might not be running 25% faster, it still able to run at 3.5GHz, rather than 2.8GHz.

And, once these multi rate gigabit get cheaper, cat5e should be good up to 5gig at 100m, 10G short run. :D

Unless they get really serious about pushing VP9 in hardware and very soon, history will repeat itself with HEVC/VP9.

VP8 didn't have a chance, because H.264 was long established and popular YEARS before VP8/WebM was even released. H.264 had 7 years of no competition, so yeah, it caught on, and VP8 didn't.

This time, though, VP9 was released head-to-head with H.265. Things could be very different. While companies start coming up with H.265 decoders, VP9 will be right there next to it, completely free to implement without paying the fees they do for h.265, and with a big company like Google encouraging them, and even offering to help.

Google, and On2 before them, have always encouraged development of hardware VPx decoders. And VP9 is no exception:

http://www.design-reuse.com/ne...

Like that design, I wouldn't be surprised at all to see most H.265 chips just incidentally also supporting VP9 on equal footing. After all, why NOT provide both, when the latter is free?

There's no guarantee that VP9 will catch on, just as there's no guarantee that anybody will use H.265... it might get basically passed-over by everyone, like H.263 did... but this time around, VP9 has a solid chance to actually compete, rather than being too late to the party.

Clock mesh technology produces a much lower clock skew compared to a conventional clock tree and, more importantly, is inherently OCV tolerant. On-chip variations (OCV) derated clock mesh designs generally have both lower skew and higher performance than clock tree designs. source: http://www.design-reuse.com/articles/21019/clock-mesh-benefits-analysis.html

Many thanks! Great article! Looks like getting the clock signal across the chip is an understood problem.

Which leads me to a question I'm not sure how to phrase, so please bear with me. At a high level, there's "data" and there's "processing". It seems to me that as we crank up the clock rate, latencies in accessing the data can be problematic unless the data is "nearby" I recall that Intel was working on Knight's (bridge? landing? something-or-other) which had lots of smaller cores on one die. IOW, it gets the processing and the data closer together. Is that the shape of things to come? What other technologies, if any, are available to keep the flow of processing at its peak?

Thanks again for the clock mesh info!

The below processor looks like the current transistor king to me, way beyond the scope of the discussions on moores law here, sometime you need to think outside the mainstream box more than double Intels best


Virtex-7 2000T FPGA Device First to Use 2.5-D IC Stacked Silicon Interconnect Technology to Deliver More than Moore and 6.8 Billion Transistors, 2X the Size of Competing Devices SAN-JOSE, Calif., Oct. 25, 2011-- Xilinx, Inc. (Nasdaq: XLNX) today announced first shipments of its Virtex®-7 2000T Field Programmable Gate Array (FPGA), the world's highest-capacity programmable logic device built using 6.8 billion transistors, providing customers access to an unprecedented 2 million logic cells, equivalent to 20 million ASIC gates, for system integration, ASIC replacement, and ASIC prototyping and emulation. This capacity is made possible by Xilinx's Stacked Silicon Interconnect technology, the first application of 2.5-D IC stacking that gives customers twice the capacity of competing devices and leaping ahead of what Moore's Law could otherwise offer in a monolithic 28-nanometer (nm) FPGA.

FPSLIC was an 8bit AVR with FPGA up to 2300 cells. All the AVR did was manage the reconfig process. It doesn't look like an app on the AVR could actively invoke logic on the FPGA. The Zynq boots into Linux and leaves the FPGA as a peripheral for circuits to deliver data or interface with external HW. It can run Linux and its app processes on one core, leaving the other core "raw" for running processing directly, or embedded in the FPGA to optimize circuits factoring out ARM instructions from FPGA.

You think $15 is expensive for a dual-core ARM-A9 and a fat Artix FPGA, integrated in a fast AMBA bus? An ARM-A9 that can run Linux 2.6, and access the FPGA as a peripheral? In 28nm, with low power consumption? Even at $25 for hundreds-count it seems a great deal for the performance.

That's not very precise analogy IMHO; what Apple needs in SoC is pretty damn close to what other mobile phones need (plus...A4 is also for tablets; seems it could be not-so-optimal after all for one of the products in which it is used, don't you think?)

Also, from TFA: "What we found was an APL0398 chip, presumably the next-generation processor from the APL0298 that we found in the iPhone 3GS."
And Samsung has the expertise, that's almost as good already. Or, here, grab the tools & data used with GPL CPU design mentioned in this article and play with creating your own chip.

It's unlikely the CPU core was modified much, they probably used some more efficient in comparison to what they had DSPs/etc., or throttling methods of those; so A8 part doesn't really come into consideration (and even if - then Apple has it just in time for A9 SoCs showing up, for example)

Oh, and you overestimate how designing SoC can often look nowadays... (screenshot; yes, even basically point'n'click CPU customisation)

Funny is how everything changes after 2012, they will have a different type of transistors. Maybe the guy really thinks things won't matter after 2012 - nut-case.

Just in case, I ask you to hold them to their other words too:

http://www.design-reuse.com/news/4850/intel-building-blocks-10-ghz-processors.html

Next year we are going to see 10GHz processors, this is going to be an interesting exercise.
Maybe Tom's Hardware or some other brave soul will manage.

I like the idea, I like the idea a lot, but the fact that they opted for a simple but slow topology doesn't fill me with hope. Especially as they suggest running SMP over it. Processors close to the centre of the "mesh" will be resource-starved.

These chips seem to be designed for specific applications, not as a general purpose CPU, especially in the DSP and digital video markets. I found this and this .

I don't see these chips as being that revolutionary or anything. Yes, they are similar to the transputer, and somewhere I read where they have a "revolutionary" shared L3 cache, which AMD ships today, and Intel either has them or are shipping soon as well.

These things seem pretty cool, but I see these as having a limited range of usage.

But, they run Linux!

Since I am a green horn at writing drivers, I would not start with Networking... since they seemed pretty covered-- for now.

However if they were USB or other dongle related, then this is where the need seems to be more now (with Linux).

Also, the article/thread seems to head in the very same direction that many other 'beginer' or 'how to's' are... which is.. either the writer asumes an awfule lot on what the reader knows already.. or it cover things the average programmer already knew since he fell out of his crib.

http://www.linuxjournal.com/article/4786 is an attempt at helping those interested in working with developing USB drivers and it's associated aspects.

And here http://www.us.design-reuse.com/news/news8597.html is interesting tech that helps simplify the device driving process from the start, by helping programmers develop the hardware, automatically after putting in merely the algorithms.

Although this process is not directly related, the point is we have to get away from re-inventing the wheel, by using more automatic code-writing, and into the realm of letting the computer do what it does best (repetition & long calcs).. and let developers do what they do and love best... *designing*.

It's amazing that this thread, which could have been the driest all week, turned out to be the one more interesting & funniest =:]

---NO!! I am NOT trying to be a 'good boy'.. It's just that sometimes I can't help myself :)

There's an inbetween.

A three week public flame war? Incidents like this are why I have a hard time selling Linux as something other than a "high school kid" operating system. Regardless of the reasons behind it, there are much better ways to handle it than a flame war, let alone one that lasts three weeks.

Actually, there is a good discussion about Bluetooth available here and basically what it has to say about licensing costs is that there are reasonable licensing plans for virtually any size production (from just a few to millions of products).

This is utter bullshit.

Really, you should have a clue before posting. According to this press release, the ARM processor consumes around 8mW, though I can't dig out any real-life figures atm. It really is an order of magnitude more power-efficient than anything x86. If you do some basic micro-processor studies you will find that most of the Intel chips are taken up with trying to translate the cludgy CISC x86 instruction set into the 'RISC-like' opcodes Intel uses internally. There is no avoiding the fact this wastes power. Here is an ARM device at random that consumes
less less than 450mW.

Phillip.

One article from 8 years ago is hardly the most convincing rebuttal.

• 18 November 1998: IBM kills study on software patents by the Whitehouse
• 1 July 1998: IBM manages to push through patents on "computer programs that have a further technical effect beyond the normal physical interaction between a computer and a computer program" at the EPO (represented by Fritz Teufel, their head patent lawyer in Europe). An example of such a further technical effect is reducing the number of mouse clicks (as in Amazon One Click).
• 8 September 2000: Pension Benefits case: IBM manages to push through "program claims" (claims on computer programs on a carrier) at the EPO.
• 30 October 2000: Fritz Teufel, IBM head patent lawyer, takes part in a German debate representing the pro software patents side (German article)
• IBM: $1.6 billion revenue from patent licensing in 2000.

Ok, I should have seen this coming. Fujitsu sells RC extraction tools, has licensed an ARC core and they make SPARC processors. Obviously PARC was the missing link!

In retrospect I wish USB had never happened, too, at least not the way that it did. This is just the latest straw to break the back of a camel that is already saddled with $15 cables that can only be 16.5 feet long and have four different ends for maximum incompatibility.

Firewire is all well and good, but how about... ethernet?

Maybe it's time for openhardware to look into developing a standard for isocronous cooperative use of FastEthernet and GigabitEthernet the way we now use USB... where any number of devices connected with a normal $15 ethernet switch could exchange data as long as they followed a timesliced multiple-acces protocol.

Then get it to work on making a very cheap SoCwhich botique manufacturers can slide into their products in place of the USB system. Not to mention that hardware hackers would have a grand old time using them for whatever purpose they can imagine. The right choice could even be universal, handling any additional worthy layer1 like Firewire.

Then any device "trusting" could be built around open standards, not corporate back-room boondoggles like the first two USB standards.

It would great if the Open Source community could gain some foothold in determining what gets bolted
into consumer devices. We are far from that. But a next step might be to have an opensource-friendly component-of-choice for embedded systems. Could having the design and base OS for a SoC lincensed free of charge give such a thing a competitive edge against proprietary applications?

Maybe it is cause ARM does not really shove itself down people's throats. Their business practices help set them apart.

No they don't, they throw their weight around just like any other monopolist. See this project.

And see it die at the hands of ARM's lawyers.

ARM is no saint. Their strong arm tactics just haven't been noticed yet.

In 2001 a student produced an open source microprocessor implementing a cut down version of the ARM instruction set, However not long after, ARM pressured OpenCores to remove the it from their website, and nnARM disappeared.

Maybe the reason people like ARM is that at the moment, most of their competition is from big companies and not open source. If projects like OpenCores catch on and FPGAs become cheaper then maybe open source can perform as well in that region as it does in software. Then I think people would not be happy with ARM taking down compatible products, just as people would not be happy if Microsoft went after WINE.