Dear Delconi,
Please refer to my comment below.
Hi Sir,
I’m one of your students from CZ1006, apparently there’s a few simple question which I might have miss out during your lecture and hope to understand some concept better.
In the slide regarding Polled I/O Advantages, it was mention that “Programmer complete control over the entire process”, does it mean that the Programmer know what is happening at each stage? Like if they code it such that after maybe task 1 & 2, it will wait for I/O (without any interrupt), that’s why it’s call “complete control”?
[Jason] For Polled I/O, the CPU checks the I/O status periodically. CPU initiates a check for available data at the I/O. For Interrupt-triggered I/O, the external device triggers the CPU when data is available at the I/O.
[Jason] For Polled I/O, the CPU checks the I/O status periodically. CPU initiates a check for available data at the I/O. For Interrupt-triggered I/O, the external device triggers the CPU when data is available at the I/O.
And in both Polled and interrupt, it mention something about “hardware interface circuitry”. Does it means circuitry with external devices (I was assuming hardware = external device)? And why is more “hardware interface circuitry” bad? I’m assuming cause there’s a ‘transferring’ of instruction in and out and error might occur during those transaction, do hope to check if it’s the case.
[Jason] Hardware interface circuitry does not mean external device. Having more complex hardware interface circuit is not necessarily ‘bad’. More circuits translate to higher cost. Some applications do require periodic checking of available data at the I/O. For instance, data logger that logs the temperature of a building every 1 second. In that case, polling is sufficient.
Hope my questions sounded correct as I’m bad in phrasing my questions.
Thanks
With Regards,
Delconi Quek