GATE CSE 2024 | Set 1 | Question: 5

Question

​​​​Which one of the following statements is FALSE?

• A. In the cycle stealing mode of DMA, one word of data is transferred between an I/O device and main memory in a stolen cycle
• B. For bulk data transfer, the burst mode of DMA has a higher throughput than the cycle stealing mode
• C. Programmed I/O mechanism has a better CPU utilization than the interrupt driven I/O mechanism
• D. The CPU can start executing an interrupt service routine faster with vectored interrupts than with non-vectored interrupts 

Comments

@usher Burst Mode definitely degrades system performance as CPU is idle for long time whereas in Cycle stealing mode it does not degrades system performance so that is the case where Cycle stealing is better than burst mode where performance cannot be compromised.

---

@usher 
For Devices which are very very slow, in that case Cycle stealing can outperform Burst mode, even for Bulk data transfer. Because Burst mode will degrade System performance on higher scale. 

---

References : Morris Mano ,Carl Hamcher 

In a nonvectored interrupt, the branch address is assigned to a fixed location in memory. In a vectored interrupt, the source that interrupts supplies the branch information to the computer.

The vectored interrupt scheme is faster because it eliminates the time-consuming polling process needed to identify the interrupting device. In a non-vectored interrupt, the processor must execute a software polling routine to query devices sequentially, whereas in a vectored interrupt, the device directly supplies the vector address to the processor via the hardware bus."

Vectored Interrupt: The hardware device places its unique vector interupt onto the data bus alongside the interrupt signal. The CPU reads this vector and immediately loads the starting address of the Interrupt Service Routine into the Program Counter (PC). The time delay is $O(1)$.

Non-Vectored Interrupt: The CPU branches to a single, fixed memory address regardless of which device triggered the interrupt. The CPU must then run a software polling procedure that checks the status register of Device 1, then Device 2, then Device 3, and so on, until it identifies the source. The time delay is $O(N)$.

Answer 1

A.In the cycle stealing mode of DMA, one word of data is transferred between an I/O device and main memory in a stolen cycle 

ans.True (as defination)

B.For bulk data transfer, the burst mode of DMA has a higher throughput than the cycle stealing mode

ans.True(as define, through as compared to bulk data, for burst mode through put will be equal to 1 as bulk of data will be transferd at once but with cycle stealing mode transfer will take more time and throughput will be less)

C.Programmed I/O mechanism has a better CPU utilization than the interrupt driven I/O mechanism

ans.False (Programmed i/o is like polling(can remeber to recall the concept) so to check wether there is i/o who need's cpu ,cpu has to do alot of work, but in interrupt driven i/o ,i/o has given preveleged to tell cpu that i need you by sending the interrupt so cpu in is more utilized in interrupt driven i/o)

D.The CPU can start executing an interrupt service routine faster with vectored interrupts than with non-vectored interrupts 

ans.true (vectored interrupt mean's address of program which handles that interrupt is send with the interrupt itself where in non vectored interrupt it is not send so cpu has to run defualt service rountine to identify the address of that routine and then run that routing)

Comments

Where can I find theoretical content or any notes like that?

---

@Dev_Ops You can refer this : https://stackoverflow.com/questions/40541880/what-is-the-difference-between-programmed-polled-i-o-interrupt-driven-i-o-an

---

C is false as both cannot be compared directly without more context.

 

• for slow I/O devices or when the I/O operations are infrequent:
  CPU will waste lot of time polling in Programmed I/O, so Interrupt Based is most efficient.
• for fast I/O devices or when the I/O operations are frequent:
  delay caused by context switch for an interrupt can be more than polling, so Programmed I/O is better.

Answer 2

Correct answer - Option C

In a system using programmed I/O, the cpu polls the devices regarding the status update of the I/O event. till the event is completed, the cpu would have wasted cycles trying to poll devices for the completion of I/O Task. In an interrupt based mechanism the CPU keeps running uninturrupted till the device activates the interrupt line indicating to the cpu about the completion of the I/O event. Hence there will not be any wasted cycles. Hence Interrupt based I/O mechanisms have better CPU utilisation rate. 

Answer 3

Option C

Because the data transfer rate in programmed I/O is slower than that of an Interrupt Driven I/O.