GATE CS 2021 Set 2 Q27 — Memory Hierarchy: Cache, Main Memory and Secondary Storage

Let us analyze the two statements individually:

Statement $S_1$: True

• In a write-through $L1$ cache, every write operation updates both the $L1$ cache and the next level ($L2$ cache) simultaneously.
• Because of this, the data in $L1$ is always consistent with $L2$, meaning $L1$ never contains any "dirty" lines (lines modified in $L1$ but not updated in $L2$).
• When a read miss occurs in $L1$, a line may need to be evicted to make room for the new line. Since there are no dirty lines in $L1$, no writeback to $L2$ is ever required.

Statement $S_2$: False

• The pairing of write policies is typically as follows:
  • Write-back caches are commonly paired with write-allocate to ensure that subsequent writes to the newly allocated block happen quickly in the cache.
  • Write-through caches are commonly paired with no-write-allocate (write-around) because writing directly to the lower level anyway makes bringing the block into $L1$ on a write miss unnecessary.
• The statement asserts the exact opposite and claims it "must" be used, which is incorrect.

Thus, $S_1$ is true and $S_2$ is false.