Bangladesh Ordinance Factory Assistant Programmer Exam: 2022 (MIST)প্রশ্ন ও সমাধান

Direct answer: Running a cable beyond its rated maximum length causes signal attenuation to exceed the receiver's detection threshold, leading to data errors, slower effective speeds, or complete loss of connectivity.

Why this happens: every cable type has a maximum length (e.g. Cat5e/Cat6 Ethernet: 100 m) beyond which the signal weakens (attenuates) and picks up more noise/crosstalk than the receiver's electronics can reliably distinguish from a valid signal.

Consequence	Explanation
Signal attenuation	Signal strength drops below the receiver's minimum detectable level
Increased bit errors	Weak/noisy signal is misread, corrupting frames (retransmits triggered by CRC checks)
Reduced throughput	Retransmissions and error-correction overhead slow effective data rate
Intermittent link/no link	NIC or switch port may fail to establish or repeatedly drop the link
Standards non-compliance	Voids certification/warranty and can break interoperability with certified equipment

Fix: insert a repeater or switch partway along the run to regenerate the signal, use a higher-category cable rated for the distance, or switch to fiber optic cable, which supports far greater distances with minimal attenuation.

Example: A Cat6 Ethernet run of 150 m (beyond the 100 m limit) will typically auto-negotiate to a lower link speed or drop entirely; adding a switch at the 90 m mark to regenerate the signal restores a reliable full-speed link.

Exam takeaway: Cable length limits exist because of signal attenuation — beyond the rated distance, use a repeater/switch to regenerate the signal or switch to fiber for long runs.

Direct answer: Hyper-Threading (Intel's implementation of Simultaneous Multithreading, SMT) lets one physical CPU core present itself to the OS as two logical processors, so it can juggle two independent instruction streams (threads) and fill idle execution-unit cycles more efficiently.

How it works:

graph TD
    Core["Physical Core"] --> T1["Logical Processor 1"]
    Core --> T2["Logical Processor 2"]
    T1 --> EU["Shared Execution Units, Cache, ALU"]
    T2 --> EU

• The core duplicates only the lightly-used front-end resources (architectural registers, instruction pointer, pipeline state) for each thread, while execution units, cache, and the ALU stay shared.
• When one thread stalls (e.g. waiting on a cache miss), the core can immediately issue instructions from the second thread instead of leaving execution units idle.

Benefits:

• Better resource utilization — reduces wasted CPU cycles during memory-bound stalls.
• Higher throughput for multi-threaded workloads (video encoding, compilation, virtualization).
• Cheap way to gain roughly 15-30% extra performance without adding a full physical core.

Limitation: because execution units are shared, two hyper-threaded logical cores do not perform like two real physical cores — gains are workload-dependent and sometimes negligible.

Example: A video-editing workstation with a 4-core/8-thread CPU (Hyper-Threading enabled) renders a timeline noticeably faster than the same CPU with Hyper-Threading disabled, because encoding threads overlap well with memory-wait stalls.

Exam takeaway: Hyper-Threading = 1 physical core presented as 2 logical processors by duplicating only front-end state, improving throughput by filling pipeline stalls — it is not equivalent to doubling physical cores.