waltsailing: From what I understand, the fifo in PropII can be used by the spin interpreter as its stacks. I am not sure if variables could also be placed here. However, the stack would provide a huge performance boost on it's own. I gained a huge performance leap by placing the decoder in hub ram (it could be in hub rom on the propII). So I see spin (and any other tokenised langage that may appear) will have a huge boost in PropII.
localroger: Yes, the instruction set is fantastic because there are no registers/accumulators as such, so the risc architecture is extremely fast. And yes, we all know the benefits of the prop with its fantastic soft control of the I/O. Today we can have many TV drivers, tomorrow lots of serial ports, all on the same chip, with just a few resistors. I see that as one of the major reasons we love the prop. It doesnt matter, within reason, what I/O we want, the same chip can do it. Therefore, we are pushing the ideas of how we can drive this beautiful chip to levels never intended.
I do not see us driving the chip to other levels as a disadvantage, just that we have to work within the design constraints imposed. Sometimes, it will be better to use another chip or another auxiliary chip. We just need to be aware that others view our shortcomings as a reason the prop is no good, and this is not the case.
Shortly I will be releasing a prop based pcb that is intended to be a PropPlug replacement. Many will ask why. I see it as a great way to replace the FTDI chip (albeit at more cost) and to get to understand the USB protocol. From that we may be able to experiment with USB further. More info soon. Why? Because I/we can!
I'm allmost done with a $50k project at work which contains 2 props controlled by a pc through ethernet.
The props controll 128 relays each which can power or read whats connected to it.
I my self did all the design, schematics, layout, ordering all the components(from the cabinet to resistors), asm programming on the prop. , 1 agilent DMM and a power supply, PC has teststand running labview modules and software from the customer, all components are controlled through ethernet with the PC.
It's alot of work and the propeller helped me alot, but i'm known with the Prop. so maybe that is why I choosed it.
45nm?! That's kinda surprising... It would be nice because the Hi-k Metal Gate process would chop the power consumption in half. (Meaning you can leave the 45nm HkMG Propeller II whirring along, underclocked at 100MHz, for a week on two Energizer Lithium AA batteries or longer - until the LDO / DC-DC inverter craps out due to battery draining to death.) Da*n, I want this thing... Although, they moved from 160 to 180MHz, which is a clear-cut evidence that they're now settled to 130nm SOI process. Like Leon said, 90nm is a possibility.
but... I WANT 45nm Propeller II. T____T I want it. (if propeller II is not to be made on 65 / 45nm HkMG process, the Propeller III will be for sure. We won't see Propeller III 'til 2013 - 2016 - maybe early if Parallax Semiconductors, Inc. is up and running by that time.)
Comments
localroger: Yes, the instruction set is fantastic because there are no registers/accumulators as such, so the risc architecture is extremely fast. And yes, we all know the benefits of the prop with its fantastic soft control of the I/O. Today we can have many TV drivers, tomorrow lots of serial ports, all on the same chip, with just a few resistors. I see that as one of the major reasons we love the prop. It doesnt matter, within reason, what I/O we want, the same chip can do it. Therefore, we are pushing the ideas of how we can drive this beautiful chip to levels never intended.
I do not see us driving the chip to other levels as a disadvantage, just that we have to work within the design constraints imposed. Sometimes, it will be better to use another chip or another auxiliary chip. We just need to be aware that others view our shortcomings as a reason the prop is no good, and this is not the case.
Shortly I will be releasing a prop based pcb that is intended to be a PropPlug replacement. Many will ask why. I see it as a great way to replace the FTDI chip (albeit at more cost) and to get to understand the USB protocol. From that we may be able to experiment with USB further. More info soon. Why? Because I/we can!
I'm allmost done with a $50k project at work which contains 2 props controlled by a pc through ethernet.
The props controll 128 relays each which can power or read whats connected to it.
I my self did all the design, schematics, layout, ordering all the components(from the cabinet to resistors), asm programming on the prop. , 1 agilent DMM and a power supply, PC has teststand running labview modules and software from the customer, all components are controlled through ethernet with the PC.
It's alot of work and the propeller helped me alot, but i'm known with the Prop. so maybe that is why I choosed it.
gr,
but... I WANT 45nm Propeller II. T____T I want it. (if propeller II is not to be made on 65 / 45nm HkMG process, the Propeller III will be for sure. We won't see Propeller III 'til 2013 - 2016 - maybe early if Parallax Semiconductors, Inc. is up and running by that time.)
We will find out very soon, though.