On 22/05/14 09:37, Anders.Montonen@kapsi.spam.stop.fi.invalid wrote:> Randy Yates <yates@digitalsignallabs.com> wrote: >> Paul Rubin <no.email@nospam.invalid> writes: >>> Ulf Samuelsson <ulf@notvalid.emagii.com> writes: >>>> If you want STM32 witrh M4F (FPU) core, then Digikey only sells >>>> maximum 256 kB versions. Atmel has 512kb/1MB versions. >>> I think that >>> http://www.digikey.com/product-detail/en/STM32F407VGT6/497-11605-ND >>> is the part on the STM M4 Discovery board. It has 1M flash, 192k ram, >>> and some crypto acceleration and secure boot IIRC. >> RAM FLASH Clock Temp. Price >> P/N (KB)(KB) MAC PHY Crypto UARTS CAN I2C ADC (MHz) Range (USD) >> http://tinyurl.com/oakwg9t MK60FN1M0VLQ12 256 1024 y y y 6 2 2 4 16-bit 120 -40 to 105 6.69 (10KU) >> >> http://www.newark.com/freescale-semiconductor/mk60fn1m0vlq12/mcu-32bit-cortex-m4-120mhz-lqfp/dp/89T6229 > > Might be worth pointing out that the Kinetis parts don't have > hardware floating point (ie. Cortex-M4 as opposed to Cortex-M4F). >Some of the Kinetis parts have floating point, some don't. As far as I can see, all the Kinetis at 120 MHz and above have floating point, while those at 100 MHz and below do not. In particular, the MK60FN1M0VLQ12 linked above has hardware floating point - that's what the "F" after the MK60 means.
ST ARM versus Atmel ARM?
Started by ●May 18, 2014
Reply by ●May 22, 20142014-05-22
Reply by ●May 22, 20142014-05-22
Op Tue, 20 May 2014 22:42:34 +0200 schreef John-Smith <noospam@noospam.com>:> "Boudewijn Dijkstra" <sp4mtr4p.boudewijn@indes.com> wrote > >>> IAR >>> compilers used to be good but they did contain some really bad details >>> e.g. an incredibly slow sscanf() where they were doing multiple >>> floating point ops for every single input digit. >> >> I'm curious which target processor and which version that was, but >> AFAICS >> recent libs don't use floating point math when dealing with integers and >> you can choose between different printf/scanf formatter features. > > It was their Z80/Z180 compiler. > > In that case we were using sscanf() to parse strings like 123.45 into > 32-bit floats. The code was profiled with a real hardware ICE (those > were the days!) and we found it was spending ~99% of its time in the > function, which was doing a float add and a float mult (by 10!) for > every digit.That sounds pretty horrible, but typical of the time. Nowadays multiple digits are collected in integers and then converted one word at a time. -- (Remove the obvious prefix to reply privately.) Gemaakt met Opera's e-mailprogramma: http://www.opera.com/mail/
Reply by ●May 22, 20142014-05-22
David Brown <david.brown@hesbynett.no> writes:> On 22/05/14 09:37, Anders.Montonen@kapsi.spam.stop.fi.invalid wrote: >> Randy Yates <yates@digitalsignallabs.com> wrote: >>> Paul Rubin <no.email@nospam.invalid> writes: >>>> Ulf Samuelsson <ulf@notvalid.emagii.com> writes: >>>>> If you want STM32 witrh M4F (FPU) core, then Digikey only sells >>>>> maximum 256 kB versions. Atmel has 512kb/1MB versions. >>>> I think that >>>> http://www.digikey.com/product-detail/en/STM32F407VGT6/497-11605-ND >>>> is the part on the STM M4 Discovery board. It has 1M flash, 192k ram, >>>> and some crypto acceleration and secure boot IIRC. >>> RAM FLASH Clock Temp. Price >>> P/N (KB)(KB) MAC PHY Crypto UARTS CAN I2C ADC (MHz) Range (USD) >>> http://tinyurl.com/oakwg9t MK60FN1M0VLQ12 256 1024 y y y 6 2 2 4 16-bit 120 -40 to 105 6.69 (10KU) >>> >>> http://www.newark.com/freescale-semiconductor/mk60fn1m0vlq12/mcu-32bit-cortex-m4-120mhz-lqfp/dp/89T6229 >> >> Might be worth pointing out that the Kinetis parts don't have >> hardware floating point (ie. Cortex-M4 as opposed to Cortex-M4F). >> > > Some of the Kinetis parts have floating point, some don't. As far as I > can see, all the Kinetis at 120 MHz and above have floating point, while > those at 100 MHz and below do not. In particular, the MK60FN1M0VLQ12 > linked above has hardware floating point - that's what the "F" after the > MK60 means.Right. This looks like a damn fine part to me - I recently recommended it in a proposal. It is cheaper than than TI's Sitara AM335x and has crypto. -- Randy Yates Digital Signal Labs http://www.digitalsignallabs.com
Reply by ●May 22, 20142014-05-22
Randy Yates <yates@digitalsignallabs.com> writes:> David Brown <david.brown@hesbynett.no> writes: > >> On 22/05/14 09:37, Anders.Montonen@kapsi.spam.stop.fi.invalid wrote: >>> Randy Yates <yates@digitalsignallabs.com> wrote: >>>> Paul Rubin <no.email@nospam.invalid> writes: >>>>> Ulf Samuelsson <ulf@notvalid.emagii.com> writes: >>>>>> If you want STM32 witrh M4F (FPU) core, then Digikey only sells >>>>>> maximum 256 kB versions. Atmel has 512kb/1MB versions. >>>>> I think that >>>>> http://www.digikey.com/product-detail/en/STM32F407VGT6/497-11605-ND >>>>> is the part on the STM M4 Discovery board. It has 1M flash, 192k ram, >>>>> and some crypto acceleration and secure boot IIRC. >>>> RAM FLASH Clock Temp. Price >>>> P/N (KB)(KB) MAC PHY Crypto UARTS CAN I2C ADC (MHz) Range (USD) >>>> http://tinyurl.com/oakwg9t MK60FN1M0VLQ12 256 1024 y y y 6 2 2 4 16-bit 120 -40 to 105 6.69 (10KU) >>>> >>>> http://www.newark.com/freescale-semiconductor/mk60fn1m0vlq12/mcu-32bit-cortex-m4-120mhz-lqfp/dp/89T6229 >>> >>> Might be worth pointing out that the Kinetis parts don't have >>> hardware floating point (ie. Cortex-M4 as opposed to Cortex-M4F). >>> >> >> Some of the Kinetis parts have floating point, some don't. As far as I >> can see, all the Kinetis at 120 MHz and above have floating point, while >> those at 100 MHz and below do not. In particular, the MK60FN1M0VLQ12 >> linked above has hardware floating point - that's what the "F" after the >> MK60 means. > > Right. This looks like a damn fine part to me - I recently recommended > it in a proposal. It is cheaper than than TI's Sitara AM335x and has > crypto.Doh! Sorry. It is cheaper than TI's Tiva TM4C1294NCPDT. -- Randy Yates Digital Signal Labs http://www.digitalsignallabs.com
Reply by ●May 22, 20142014-05-22
Den onsdag den 21. maj 2014 18.28.33 UTC+2 skrev Ulf Samuelsson:> 2014-05-20 22:45, John-Smith skrev: > > > What about code and fixed data security? > > > > > > On chip flash ought to be reasonably good for that. This is not a high > > > security application but it would be nice if I could protect the on > > > chip stuff from a casual code dump. > > > > > > > On chip flash is of course more secure than off chip flash. > > Then again, the extra security is going to be quite expensive. > > > > There are micros with AES/3DES crypto engines which can be > > used to encrypt external data. > > > > Some of those chips have export restriction, > > but the XMEGA does not. > > > > If you want STM32 witrh M4F (FPU) core, then Digikey only sells maximum > > 256 kB versions. Atmel has 512kb/1MB versions. >afaict all of the stm32f4xx/stm32f3xx family have FPU, regardless of whether they call them M4 or M4F digikey has plenty of stm32f4xx upto 2MB -Lasse
