Slashdot Mirror
Texas Instruments Embedding Linux
darthcamaro writes "Looks like pretty soon Linux will truly become ubiquitous thanks to Texas Instruments new DaVinci System on a chip DSP. The new consumer electronics chip aimed at capturing the Digital Video market is powered by MontaVista Linux. 'TI understands that there is a larger number of Linux programmers than there are DSP programmers,' Huy Pham, DSP System-on-Chip product marketing manager at TI, told internetnews.com. 'What [DaVinci] does in partnership with MontaVista is enables the Linux developer to use the DSP without needing to understand the complexity of programming the DSP.'"
As a veteran of digital signal processing I believe this is a major step forward. The roots of DSP are different from traditional Von Neumann computing and yet there is some overlap. For 50 years we have basically had two streams of thinking, two paradigms to cope with. Algorithms for filters, impulse systems, convolution, table mapping, oscilators
and so forth have a mathematical basis that generally incorporates time implicitly although at the most basic level they are matrix computations. Expressing these in orthodox code can be frustrating, counterintuitive and inefficient. For example compare side by side the solutions to an audio filter written in Nyquist (scheme), vanilla C code, 56 series DSP
assembly or as a process flow in Max/PureData. They are barely recognisable as the same equation and converting between one paradigm and another (say functional to procedural) is so difficult one may as well start from scratch with the original equation. The idea of microcode/silicon instruction sets combined with the abstraction of a familiar kernel and realtime operating system as the starting point is going to be immensely empowering for the next generation of DSP programmers. Indeed I expect that in 10 years time we will no longer consider the two as distinct disciplines at all.
The next step in microprocessor evolution is, in accordance with the 'great wheel', for these entire architectures and their instruction sets to be incorporated back into the mainstream CPU core along with language constructs for dealing with them, such that one day it will be no more unusual express a closed form cosine sum than it is to write a for loop today.
Gee, maybe you should go to the manufacturer's site for that kind of detail, instead of whining that a news article didn't publish tech specs.
Kudos to the Linux Penguin crowd on this one:
I say that, because for one thing, this shows how flexible Linux can be - this is 1 thing it has going for it by all means vs. Microsoft, including more multi-platform capability/portability & better clustering abilities... so far @ least on the latter (more on that upcoming).
As far as "multi-platform" capability? Well, Microsoft used to have builds of NT 3.x-3.5x that ran on diff. processor platforms (e.g.-> MIPS), but has since centered on MOSTLY Intel compatible x86 code for their OS, at least as far as personal computing.
This use of Linux imo is a good thing that shows where Linux has a great deal of flexibility, & even a "Pro-Microsoft person" like myself has to admit & admire it.
(Remember - & this is coming from one of the most "Pro-MS" people here most likely @ slashdot (making me a member of a HUGE 'minority' here @ slashdot, no doubt)).
APK
P.S.=> Linux is making strides, there's no doubt about it, but then, so is Microsoft!
(I think VISTA will be a HELL of an OS, especially since it is being built off the Windows Server 2003 99.999% uptime rated core which is far more secure by default than is Windows NT/2000/XP & which I have been running for 2++ years now @ home solid & stable as titanium steel in this capacity (default workstation mode install) + Windows STILL LEADS in device driver/hardware support)
Also, I think the Windows CCS (compute cluster server) will be another 'surprise' out of Microsoft, especially after acquiring personell from Cray in regard to SuperComputing ready platforms)... but, we'll have to wait & see on both accounts! Time will tell & all that!
However, imo - The best part is, we as the end-users/consumers of BOTH OS families get the benefits! To quote a tune from "The Beatles":
"It's getting better all the time..." apk
TI wants to push its crappy DSP because they are getting ousted by power and arm processors.
But what I got from the article (reads like a press release, actually) is that it is designed to interface with arm, so they are using the DSP for what the DSP is good for (filter) and using a generic processor for the rest...
A house divided against itself cannot stand.
I'm sure it's just ignorance on the part of the salesidiot... In another matter, I asked a simple "yes/no" question which resulted in a 3 day scheduling negotiation to speak with one of their lead engineers. Once we got the conference call going, I was asked to pose my question. I posed it and the lead engineer said "err, Yes" with a barely detectable tone that said "I'm going to kill this salesperson". In the end; MV was far more expensive. $18k USD a seat for GCC plus $5k/year for business-hours support, maximum 5 'incidents'. 2 days of effort produced for me a fully functional cross-compile toolchain on which I was able to develop a prototype of our production product as proof of concept. So we punted on MV. We're not idiots. We can support our own toolchain. I suspect most embedded software shops have people who possess reasonable clue and don't need someone like MV to hoover money out of their wallets to answer gcc questions. Non-clueful people are probably not doing embedded software. So just who is MV's target audience? PHB's.