Duane Ellis a écrit : > **THIS*IS*OFF*LIST** > >>> The context switching in the NEC devices has no match. > > Really, I'd like to see how you do this, in a pre-empt context switch > there are lots of registers to save, do you have a *specific* NEC > support group you work with? I'd like a name or two, local guys here > are clue-less. Factory guy came in 5 months ago, gave us a training > session about PM Plus not able to answer questions I had. > This is not something a salesman is telling me but actual data from NEC where they compare the AR7TDMI, Cortex-M3 and V850ES According to their chart, an interrupt takes 24 cycles for ARM7TDMI, 12 for Cortex-M3 and 4 for their new V850ES.
I would assume that it has to be hardware contex switching. I would agree that many salesman are full of something, Freescale and Infineon to lead the pack in that. The funny part was with Freescale who told me that ST would not be allowed to sell us PowerPC parts for 3 years. ST was furious. Freescale being pissed that I would not consider their 16 bit part to replace an ARM device tried to go over my head. NEC loan me some boards and software, I will make my own evaluation. I accepted the Infineon board for courtesy but it doesn't look to me like it can drive my stepper motors. The device to replace is basically a cluster device, not too many on the market. There is TI, Samsung, NEC, Freescale, Fujitsu., two of them being 16 bits. I can connect 6 steppers directly on the NEC, TI or Samsung devices, four with Fujitsu. > FYI - I'm an os-type person, very familiar with *other* CPUs, wrote a > few kernels, etc, but I am very new to the V850ES parts, and support I > see from NEC here is best described as a simple sales man. Writing a > simple task switcher from a data sheet - is sort of difficult. > >>> Find me a 32 bits processor that takes 4 cycles to get to an interrupt. > > But then you have a lot to save right? > This would imply hardware switching like on the V25. As I said earlier I will be able to differentiate between sales crap and reality as I play with it. What gives me the impression that he might be correct is my past experience with the V25 where the switching was very quick. Each interrupt had its own set of registers. I did a light control system for the Montréal Casino with a V25 and I used the hardware interrupt switching. That was something considering that with the regular interrupt structure there are tons of registers to push on the stack. With NEC interrupt on the V25 I didn't need to push anything on the stack. I don't remember the number of cycles but that was quite a difference with the standard 80188 hardware. In the kind of system I work on interrupts are as short as possible. I don't do OS type stuff, I have my own small kernel with its tasks scheduler. On a project I did recently I had to bit bang 35 LEDs at different duty cycles for dimming. I used a Cortex-M3 from ST. That thing is impressive but so far ST doesn't have anything that would suit our requirement so far. > ***** > > Also what tools do you use? The NEC PM+, IAR, or GreenHills, or > something else? I have a NEC ICE-CUBE, and a few eval boards.. but > without *real* technical people answering questions I find it hard if > not impossible to get information. > Either NEC or IAR, depending on how much my boss wants to spend. We do have a few good NEC technical people in Michigan. They have been very helpfull so far. I used Greenhills for a few weeks and hated it. Never got any real help from them. IAR are very helpfull. Micronas was a pain, one thing that was amusing was when I asked why a certain pin had to be pulled to ground. They refused to give the reason saying that it must be pulled to ground. Debugging thru JTAG was marginal at best in the devices we were using following their advice. One day I had a board that accidentally erased itself. I got my answer but Micronas should have been honest. The morons have a Spansion flash internally, they chose the biggest Flash to be at the boot. This is ridiculous since my CAN bootloader is about 8k. Cypress was pushing some new devices and the salesman had a funny look when I asked if the new devices had the self erase feature. Under certain conditions their flash would self erase. This can be problematic in a vehicle when a gauge has its flash erased. They fixed that bug, the reminder didn't make him too happy. Michel -- Tired of Microsoft's rebootive multitasking? then it's time to upgrade to Linux. http://home.comcast.net/~mcatudal _______________________________________________ Openocd-development mailing list [email protected] https://lists.berlios.de/mailman/listinfo/openocd-development
