| Author |
Message |
|
|
 Posted: May 10, 2012 - 12:01 AM |
|
Joined: Aug 05, 2005
Posts: 265
Location: Tacoma, WA
|
|
| I am cribbing from a design that uses a PIC, I would like to substitute the PIC16F684 with an AVR. Any opinions on which AVR is likely to be the closest in terms of capabilities, pins, ports, power consumption? |
|
|
| |
|
|
|
|
|
|
Posted: May 10, 2012 - 12:12 AM |
|
Joined: Sep 04, 2002
Posts: 21248
Location: Orlando Florida
|
|
| Well, how many pins and ports does it have, and whats its power consumption? Seems like speed, ram, rom and price would be more important. |
_________________
Imagecraft compiler user
|
| |
|
|
|
|
|
|
Posted: May 10, 2012 - 12:40 AM |
|
Joined: Aug 05, 2005
Posts: 265
Location: Tacoma, WA
|
|
OK I admit I didn't want to look at the Atmel web site. I hoped someone might know.
220uA at 4Mhz, 14pin package, 2048flash, 128 SRAM, 256 EEPROM, about a buck and 1/2. |
|
|
| |
|
|
|
|
|
|
Posted: May 10, 2012 - 01:29 AM |
|
Joined: Nov 22, 2002
Posts: 12031
Location: Tangent, OR, USA
|
|
Does it need to be DIP?
If you need minimum physical size (14 pin DIP), it will have to be a Tiny.
Jim |
_________________
Jim Wagner
Oregon Research Electronics, Consulting Div.
Tangent, OR, USA
"The only thing standing between us and victory is defeat" P.G.Wodhouse in Wooster & Jeeves series
|
| |
|
|
|
|
|
|
Posted: May 10, 2012 - 02:00 AM |
|
Joined: Sep 04, 2002
Posts: 21248
Location: Orlando Florida
|
|
| Now you have to put an importance coefficient on pins, speed, flash, ram, power, price. What if there is an AVR with 20 pins, but has more flash, more speed, more ram, more price? If the pic has most of the 2k flash used, a 4k AVR might be better. At this level, we need to look at stuff like timers (8 bit? 16 bit), uart, spi, twi etc. Pic guys used to do a lot of bitbanged spi and uarts. Price is everything in the 100K quantities I guess. |
_________________
Imagecraft compiler user
|
| |
|
|
|
|
|
|
Posted: May 10, 2012 - 04:04 AM |
|
Joined: Mar 19, 2003
Posts: 736
|
|
Your closest match from Atmel is an ATTINY24V. Both have 2K flash and 128 bytes SRAM. The PIC has 256 bytes of EEPROM while the TINY24 has 128 bytes of EEPROM.
Both have timers, 10 bit A/D converter, and analog comparator. Also, they have the same number of I/O. The AVR has a Universal Serial Interface whereas the PIC has no serial communications module.
Power usage: Be careful reading the "tag lines"
------------------------------------------------
Conditions for both processors:
Temp = 25C
Vdd = 3V
"CORE SPEED" (not clock speed) = 1MHZ
Clock Source = Internal Oscillator
All peripherals ON
PIC16F684 = 580uA (See fig 16-12 of datasheet)
ATTINY24V = 504uA (Fully compensated value, see below)
(For Atmel, See fig 21-1 of datasheet. Note, this chart is for all peripherals turned OFF. Must add in voltage compensated current draws from Table 21-1 when peripherals are turned ON. Must also adjust Core current draw from fig 21-1 to compensate for 3V Vdd)
Notice "CORE SPEED" parameter above. To get a 1MHZ core speed from the PIC, it must be clocked at 4MHZ since each instruction cycle is 4 oscillator clock cycles. For the Atmel AVR, each oscillator clock cycle is an instruction clock cycle. Therefore, to get a 1MHZ core speed from the AVR, the oscillator is run at 1MHZ.
Cost from Digi-Key @ 100pc Qty
SOIC14 package for both
------------------------------
PIC16F684-I/SL $1.13
ATTINY24V-10SSU $1.34
Hope this helps in your evaluation. |
|
|
| |
|
|
|
|
|
|
Posted: May 10, 2012 - 05:08 AM |
|
Joined: Aug 05, 2005
Posts: 265
Location: Tacoma, WA
|
|
It did. Thanks a lot.
Jim |
|
|
| |
|
|
|
|
|
|
Posted: May 10, 2012 - 05:35 AM |
|
Joined: Dec 18, 2001
Posts: 4696
|
|
|
gahelton wrote:
Notice "CORE SPEED" parameter above. To get a 1MHZ core speed from the PIC, it must be clocked at 4MHZ since each instruction cycle is 4 oscillator clock cycles. For the Atmel AVR, each oscillator clock cycle is an instruction clock cycle. Therefore, to get a 1MHZ core speed from the AVR, the oscillator is run at 1MHZ.
Noteworthy - as power consumption is governed by the clocking!!
PIC's architecture is really old. |
|
|
| |
|
|
|
|
|
|
Posted: May 10, 2012 - 02:05 PM |
|
Joined: Feb 19, 2001
Posts: 25880
Location: Wisconsin USA
|
|
| If you are only doing a few and size isn't paramount (how could it be if a DIP design?), then I'd recommend the Mega88 family. Mega48 is about $1.25 at 100 pieces. |
|
|
| |
|
|
|
|
|
|
Posted: May 11, 2012 - 07:13 PM |
|
Joined: Aug 05, 2005
Posts: 265
Location: Tacoma, WA
|
|
I have had a chance to review the way the PIC was used in the design, and I find that the tiny24A is a more than acceptable replacement. It turns out that the PIC was using a bit banged i2c. The SPI in the tiny will do that job nicely. The price is good too...$0.61 for the tiny24A. Price is a big driver, but I can get the tiny84A for $.89
As far as footprint goes, I really need the small size so the QFN package is perfect. |
|
|
| |
|
|
|
|
|
|
Posted: May 11, 2012 - 07:14 PM |
|
Joined: Sep 04, 2002
Posts: 21248
Location: Orlando Florida
|
|
| Yo Jim... did you mean TWI instead of SPI? |
_________________
Imagecraft compiler user
|
| |
|
|
|
|
|
|
Posted: May 11, 2012 - 07:25 PM |
|
Joined: Dec 06, 2007
Posts: 2512
Location: Redmond, WA USA
|
|
Yo Bob... did you mean USI instead of TWI?  |
_________________
Larry
Those afraid to embrace the future will quickly fade into the past. - larryvc
|
| |
|
|
|
|
|
|
Posted: May 11, 2012 - 07:45 PM |
|
Joined: Sep 04, 2002
Posts: 21248
Location: Orlando Florida
|
|
| I guess. I didn't realize that was 'either' till recently. |
_________________
Imagecraft compiler user
|
| |
|
|
|
|
|
|
Posted: May 11, 2012 - 11:25 PM |
|
Joined: Aug 05, 2005
Posts: 265
Location: Tacoma, WA
|
|
This is a problem, though. I too realized a while ago that the SPI was really a USI. But, I really need the full functionality of the SPI. So I am back to looking at the mega48 or back to bit banging something. I really didn't want to do that. The M48 will work (it has to work with a Nordic nRF2560, which can accommodate either a SPI or an i2c). The problem is price. The M48 is 1.26 and the Tiny is .61 USD. So, to do it my way, it looks like I need the M48. I guess I now see why the guys at Nordic Semi chose to bit bang the PIC.
Power wise, with the M48, it looks like I can keep the running current at 1MHz to 400-500uA. |
|
|
| |
|
|
|
|
|
|
Posted: May 12, 2012 - 12:16 AM |
|
Joined: Sep 04, 2002
Posts: 21248
Location: Orlando Florida
|
|
| I guess I called it right. See my msg from May 9 where I said there are a lot of bitbanged peripherals in pic programs? Whats the full IO count? Does the avr version have to be faster? cheaper? more features? All three? I think you can do a bitbanged spi and a bit banged twi on an avr at 20mhz a lot faster than on a pic at 4mhz. I have a program here that measures bit banged spi thruput and it is about half of the double speed hw spi thruput for writes, so in the case of writing tp the spi, the hw doesnt give you much help. You either run thru the loop toggling the data and clk 8 times, or you sit and wait for the spif flag. Both take about a microsecond. 8megabits, 1megabyte thruput. |
_________________
Imagecraft compiler user
|
| |
|
|
|
|
|
|
Posted: May 12, 2012 - 03:13 AM |
|
Joined: Dec 18, 2001
Posts: 4696
|
|
| Why bit-bang SPI when all but the crude micros have hardware SPI? |
|
|
| |
|
|
|
|
|
|
Posted: May 12, 2012 - 03:41 AM |
|
Joined: Dec 06, 2007
Posts: 2512
Location: Redmond, WA USA
|
|
|
bobgardner wrote:
I guess I called it right. See my msg from May 9 where I said there are a lot of bitbanged peripherals in pic programs? Whats the full IO count? Does the avr version have to be faster? cheaper? more features? All three? I think you can do a bitbanged spi and a bit banged twi on an avr at 20mhz a lot faster than on a pic at 4mhz. I have a program here that measures bit banged spi thruput and it is about half of the double speed hw spi thruput for writes, so in the case of writing tp the spi, the hw doesnt give you much help. You either run thru the loop toggling the data and clk 8 times, or you sit and wait for the spif flag. Both take about a microsecond. 8megabits, 1megabyte thruput.
Bob, have you an "Enter" key on your keyboard? |
_________________
Larry
Those afraid to embrace the future will quickly fade into the past. - larryvc
|
| |
|
|
|
|
|
|
Posted: May 12, 2012 - 04:39 AM |
|
Joined: Jul 02, 2005
Posts: 5912
Location: Melbourne, Australia
|
|
| What? ... and break his train of thought ... |
_________________
Ross McKenzie
ValuSoft
Melbourne Australia
|
| |
|
|
|
|
|
|
Posted: May 12, 2012 - 05:13 AM |
|
Joined: Sep 04, 2002
Posts: 21248
Location: Orlando Florida
|
|
| I used to log on to the TBBS bulletin boards in Orlando in the early 80s. Whenever I entered a hard carriage return in a message the sysop would chew me out for messing up the TBBS message formatting software, which was formatting the line width to match the display. Since I am an Only Child, I probably have a desire to Not Make Waves and do what my elders tell me, so I learned to let the forum software wrap the lines. This seems to work correctly in this forum software also. Are you now telling me that the carbon based units can format the messages better than the forum software writers? Should the moderators issue a warning that all messages that dont have a carriage return at 80 chars will have their login disabled? What if you have one of those newfangled movie screens with a 16x9 aspect ratio and the 80 column messages dont get all the way to the right side of the screen? Would you bitch about that too? Just get a life how bout? |
_________________
Imagecraft compiler user
|
| |
|
|
|
|
|
|
Posted: May 12, 2012 - 06:06 AM |
|
Joined: Dec 06, 2007
Posts: 2512
Location: Redmond, WA USA
|
|
Yes. No. What if. Yes. I have one. The only reason that I mentioned it is that you, as well as me, are getting older and it has become, for me anyway, harder to read print that has been written as though there is not enough bandwidth on the internet to handle the extra white space that could, and would, make it so much easier to figure out what it is this time that you are trying say. It seems that you have this preconceived notion that things are like they were 30 years ago. It is time to embrace this century with an open mind, don't you think? This has nothing to do with the forum software formatting, or wrapping, the text that you have entered. It seems more like a lazy attitude that you have become accustomed to and it would be nice to see a few carraige returns in your posts. You have a lot of experience to share and I appreciate your input on this forum and I have learned from it. It is just so hard to read.
I feel better now. I hope you do too.  |
_________________
Larry
Those afraid to embrace the future will quickly fade into the past. - larryvc
|
| |
|
|
|
|
|
|
Posted: May 12, 2012 - 02:43 PM |
|
Joined: Mar 19, 2003
Posts: 736
|
|
The ATTINY2313 comes real close to meeting your needs for hardware I2C(TWI) and SPI - also very low cost. However, even though this processor supports both of these functions with separate sets of hardware using the USI and UART respectively, they share the same pins. This is unfortunate.
The next stop for hardware TWI and SPI is the ATMEGA48 which you believe to be too costly.
Looks like bit banging is in your future. However, don't fret. Bit banging an SPI in master mode is extremely simple and fast. |
|
|
| |
|
|
|
|
|
|
Posted: May 12, 2012 - 02:56 PM |
|
Joined: Sep 04, 2002
Posts: 21248
Location: Orlando Florida
|
|
| Finally. For once, someone else on this stoopid forum has the sense to just type in the message to fill the screen with automattically wrapped text. It looks so warm and familiar. It made my morning. |
_________________
Imagecraft compiler user
|
| |
|
|
|
|
|
|
Posted: May 12, 2012 - 04:57 PM |
|
Joined: Dec 06, 2007
Posts: 2512
Location: Redmond, WA USA
|
|
| Oh, thank you. I thought you would like that. |
_________________
Larry
Those afraid to embrace the future will quickly fade into the past. - larryvc
|
| |
|
|
|
|
|
|
Posted: May 13, 2012 - 04:32 AM |
|
Joined: Dec 06, 2007
Posts: 2512
Location: Redmond, WA USA
|
|
Jim,
Im sure there are many other demos for SPI and usi but I came across this one on the net. I tried the Master-Slave code and it works great. I plan on trying the ATtiny2313 USI as Master code next.
http://www.jaxcoder.com/Projects.aspx?id=280385176 |
_________________
Larry
Those afraid to embrace the future will quickly fade into the past. - larryvc
|
| |
|
|
|
|
|
|
Posted: May 14, 2012 - 12:34 PM |
|
Joined: Sep 29, 2009
Posts: 16
|
|
Jimlake,
Not only is there a core speed difference between the PIC & Atmel products but the words are different bit lengths.
PIC16F684 has 2k words flash, which is actually 3.5k bytes because of a 14-bit word length.
Atmel's Attiny processors have a 16-bit word length so the attiny24 & attiny2313 processors will have only 1k words flash (2kB), the attiny44 2k words flash (4kB) and the attiny84 4k words flash (8kB).
Flenser |
|
|
| |
|
|
|
|
|
|
Posted: May 15, 2012 - 03:02 AM |
|
Joined: Aug 05, 2005
Posts: 265
Location: Tacoma, WA
|
|
To address Steve's question, it appears that most of the small 8 bitters do not have a SPI, the AVR's, at least, have a USI. None of them have a i2c.
It looks to me like the difference between the SPI and the USI is the lack of MISO. This can be a bit of a pita if the other device is expecting a full HW SPI. I have decided to resolve this by using a Mega48. In the VQFN package, it is very small, but it does cost more than the Tiny. |
|
|
| |
|
|
|
|
|
|
Posted: May 15, 2012 - 06:35 AM |
|
Joined: Jan 03, 2006
Posts: 4410
Location: Hemel Hemsptead, UK
|
|
As a master, SPI is easier to generate than I2C which is basically an abomination of a protocol. Having to change the direction of both the clock and data line in I2C to manage the three-state and clock stretching parts is a pain, and the PIC's inability properly to implement read/modify/write when a port is driving a capacitive load is extremely annoying.
There are PICS in the 18F series (next size up basically - twenty-eight pins or more) that have UART, SPI, and I2C but the implementation on the ones I examined meant that you could use only one of the three at a time, and the distribution of pins was such that even the inputs and outputs didn't match so it would be very tricky to change between them programatically. |
_________________
Neil Barnes
www.nailed-barnacle.co.uk
|
| |
|
|
|
|
|
|
Posted: May 15, 2012 - 06:36 AM |
|
Joined: Dec 18, 2001
Posts: 4696
|
|
All relative I suppose. To me, the mega32 is "small", in that RAM size tends to limit what you can do.
indeed, there are many 2KB code apps. I just don't often do that kind. |
|
|
| |
|
|
|
|
|
|
Posted: May 15, 2012 - 08:51 AM |
|
Joined: Jan 08, 2009
Posts: 1148
Location: Lund, Sweden
|
|
The tiny48/88 do have both SPI and I2C (but no usart  ). |
|
|
| |
|
|
|
|
|
|
Posted: May 15, 2012 - 09:15 AM |
|
Joined: Nov 02, 2009
Posts: 3239
Location: Zelenograd, Russia
|
|
| The Tiny1634 has 2 UARTS, 2 I2C, SPI (via UART/USI), ADC etc. in a 20 pin package. |
|
|
| |
|
|
|
|
|
|
Posted: May 15, 2012 - 12:43 PM |
|
Joined: Dec 08, 2004
Posts: 4719
Location: Nova Scotia, Canada
|
|
|
MBedder wrote:
The Tiny1634 has 2 UARTS, 2 I2C, SPI (via UART/USI), ADC etc. in a 20 pin package.
And no inventory from any of Atmel's preferred distributors. |
|
|
| |
|
|
|
|
|
|
Posted: May 15, 2012 - 12:53 PM |
|
Joined: Dec 08, 2004
Posts: 4719
Location: Nova Scotia, Canada
|
|
|
jimlake wrote:
To address Steve's question, it appears that most of the small 8 bitters do not have a SPI, the AVR's, at least, have a USI. None of them have a i2c.
It looks to me like the difference between the SPI and the USI is the lack of MISO. This can be a bit of a pita if the other device is expecting a full HW SPI. I have decided to resolve this by using a Mega48. In the VQFN package, it is very small, but it does cost more than the Tiny.
What exactly do you mean by "no MISO"?
The USI uses DI (MISO for Master systems, MOSI for Slave systems), DO (MOSI for Master systems, MISO for Slave systems), and USCK (SCLK output for Master systems, SCLK input for Slave systems).
It does not have hardware slave select - but for a SPI master, that's a fairly easy limitation to work around.
It is designed to do SPI modes 0 and 1; apparently it's not suited to modes 2 or 3.
Section 14.3.2 of the ATtiny24 datasheet shows an assembly function for USI operating in 3-wire Master mode, that accepts the desired output byte to be sent to the slave via DO as an argument, and returns the read-back value from DI.
Section 14.3.3 demonstrates using the USI in 3-wire Slave mode. |
|
|
| |
|
|
|
|
|
|
Posted: May 15, 2012 - 02:43 PM |
|
Joined: Feb 19, 2001
Posts: 25880
Location: Wisconsin USA
|
|
|
Quote:
The Tiny1634 has 2 UARTS, 2 I2C, SPI (via UART/USI), ADC etc. in a 20 pin package.
Wow--where did THAT model come from? |
|
|
| |
|
|
|
|
|
|
Posted: May 15, 2012 - 02:45 PM |
|
Joined: Jul 18, 2005
Posts: 62203
Location: (using avr-gcc in) Finchingfield, Essex, England
|
|
|
Quote:
Wow--where did THAT model come from?
+1
Is there anything else "lurking"? I imagine the device support list in the recently released final AS6 might be a good place to look for "new meat"? |
_________________
|
| |
|
|
|
|
|
|
Posted: May 15, 2012 - 02:48 PM |
|
Joined: Feb 19, 2001
Posts: 25880
Location: Wisconsin USA
|
|
|
Quote:
And no inventory from any of Atmel's preferred distributors.
How can you say that, when Atmel's press release from March says
Quote:
All AVR ATtiny1634 devices are available now...
 |
|
|
| |
|
|
|
|
|
|
Posted: May 15, 2012 - 06:49 PM |
|
Joined: Jan 09, 2007
Posts: 1848
Location: Arlington, Texas, U.S.A.
|
|
|
theusch wrote:
Wow--where did THAT model come from?
What's Changed.
Thank you Atmel! |
|
|
| |
|
|
|
|
|
|
Posted: May 15, 2012 - 06:54 PM |
|
Joined: Nov 02, 2009
Posts: 3239
Location: Zelenograd, Russia
|
|
|
gchapman wrote:
Thank you Atmel!
...for yet another vaporware chip  |
|
|
| |
|
|
|
|
|
|
Posted: May 16, 2012 - 12:00 AM |
|
Joined: Mar 19, 2003
Posts: 736
|
|
| Mouser is showing a delivery date of 7/6/2012 for the ATTINY1634. This looks like a really nice part for less than $1.00. |
|
|
| |
|
|
|
|
|
|
Posted: May 16, 2012 - 12:30 AM |
|
Joined: Sep 04, 2002
Posts: 21248
Location: Orlando Florida
|
|
| So someone wanted an avr with those specs, and placed an advance order for a couple million or whatever the R&D deal was, and the distributers get what overflows into the supply chain after the initial run to the initial customer is fulfilled? How does the distribution model work? If they can deliver in 2 months after the product announcement, they must be building up stock. Any old semiconductor retiress out there that want to give some insight how it really works? |
_________________
Imagecraft compiler user
|
| |
|
|
|
|
|
|
Posted: May 16, 2012 - 12:35 AM |
|
Joined: Sep 04, 2002
Posts: 21248
Location: Orlando Florida
|
|
| On a similar theoretical semiconductor question... would an AVR with dual CPUs get anyone excited? Seems like if you consider yourself an Expert AVR programmer, the idea of being able to write two programs/tasks and get double thruput is a win. Digital filters, etc. If anyone from Atmel reads this and forwards it to the VP of Good Ideas Submitted from the AVRfreaks, just add a PS that this one is a stockholder. |
_________________
Imagecraft compiler user
|
| |
|
|
|
|
|
|
Posted: May 16, 2012 - 01:36 AM |
|
Joined: Jun 19, 2002
Posts: 950
Location: SF Bay area
|
|
I don't think that the idea of multi-core microcontrollers makes much sense, outside of particular assymetric applications that amount to "smart peripherals" (some of the XMEGA stuff, Freescales "Time Processing Unit", etc.) Multicore is where you go after you hit the limits of process speed and power dissipation, and microcontrollers are almost by definition limited by other factors.
You can always stick two AVR on the same PCB.  |
|
|
| |
|
|
|
|
|
|
Posted: May 16, 2012 - 02:56 AM |
|
Joined: Dec 06, 2007
Posts: 2512
Location: Redmond, WA USA
|
|
|
|
|
|
|
|
Posted: May 16, 2012 - 04:23 AM |
|
Joined: Jul 21, 2005
Posts: 1373
|
|
Regarding multicore AVR, AVR fills a niche I think that niche is smooth and easy development it certainly is not computational horsepower and anything that makes the development cycle harder (I am guessing having another core would increase the complexity) would be detrimental to AVR.
OP if you can afford it, stick with the 48/88/168/xx8 series. It is the best value AVR on the market, you get alot more peripherals for the extra money.
If money is the primary factor then freescale has 8 bitters priced around 50 cents in 100pc quantity.
Sticking with PIC is not a horrible option, their plain vanilla 8 bit stuff is atleast cheapish though is quite slow when it comes to the core. The redeeming thing about PIC is that they have lots of peripheral options, and arguably they have a wider upgrade path than AVR. |
|
|
| |
|
|
|
|
|
|
Posted: May 16, 2012 - 04:28 AM |
|
Joined: Sep 04, 2002
Posts: 21248
Location: Orlando Florida
|
|
| I'm sort of in Awe of the consultant level dudes that can switch from one computer and language to another. I'm so old and stupid that I want to stick with What I Know to try and get stuff done. If the only tool you have is a hammer, every job looks like a nail. My tool is an AVR, and I'll try and use that unless it so glaringly obvious that it is about a factor of 10 too slow. However, I have duly taken notice of your out of hand Dissing of my dual core AVR idea. Wont you be embarassed if some Atmel guy thinks its a great idea? (I've already bought my Lotto tickets for Wednesday's drawing). |
_________________
Imagecraft compiler user
|
| |
|
|
|
|
|
|
Posted: May 16, 2012 - 05:24 AM |
|
Joined: Jul 21, 2005
Posts: 1373
|
|
| I meant no disrespect bob, it is just my opinion. |
|
|
| |
|
|
|
|
|
|
Posted: May 16, 2012 - 05:25 AM |
|
Joined: Dec 06, 2007
Posts: 2512
Location: Redmond, WA USA
|
|
I was never "Dissing" your idea. I actually like trying new chips and a dual core AVR would be cool.
I really like the ATtiny1634 specs. Have to get out my electric frying pan for that one.
Funny, I bought lotto tickets too. |
_________________
Larry
Those afraid to embrace the future will quickly fade into the past. - larryvc
|
| |
|
|
|
|
|
|
Posted: May 16, 2012 - 06:10 AM |
|
Joined: Sep 04, 2002
Posts: 21248
Location: Orlando Florida
|
|
| Well, I was kidding about being DisRespected. I'm glad you guys considered my idea. What if a couple of the several 1000s of AVRfreaks really came up with a good product suggestion/idea for Atmel? What if one of the Atmel Staff Engineers comes up with a New Product idea? How do the ideas get considered? So guy with a 6 figure salary just goes with his gut feel, or is an idea quantifiable somehow? (Darn marketing dudes outght to earn their money somehow). |
_________________
Imagecraft compiler user
|
| |
|
|
|
|
|
|
Posted: May 16, 2012 - 07:49 AM |
|
Joined: Dec 16, 2005
Posts: 3086
Location: Bratislava, Slovakia
|
|
The road to the silicon hell is paved with chips which came out of the best ideas.
FPSLIC, anyone?
JW |
|
|
| |
|
|
|
|
|
|
Posted: May 16, 2012 - 08:22 AM |
|
Joined: Dec 06, 2007
Posts: 2512
Location: Redmond, WA USA
|
|
|
|
|
|
|
|
Posted: May 16, 2012 - 10:00 AM |
|
Joined: Jul 18, 2005
Posts: 62203
Location: (using avr-gcc in) Finchingfield, Essex, England
|
|
|
Quote:
and get double thruput is a win
Why wouldn't you just get a 50Mhz or 100MHz or 120MHz (or whatever) ARM? Multi-CPU are only needed at the very top end where you reach the limit of the silicon and still need to go faster. Or perhaps when you have a specific job that lends itself to parallelism - like pixel processing in a GPU.
Quote:
What's Changed.
Thank you Atmel!
Interesting page. I read that tiny43U model number and thought "interesting - small package USB chips now" but as far as I can see the U (unlike most other AVR model numbers) does not have any relation to USB. Nothing like keeping the numbers consistent! |
_________________
|
| |
|
|
|
|
|
|
Posted: May 16, 2012 - 10:07 AM |
|
Joined: Dec 16, 2005
Posts: 3086
Location: Bratislava, Slovakia
|
|
|
|
|
|
|
FAQ
Search
Register
Log in to check your private messages
Log in
Maximize
RSS
