Answer.md

@@ -0,0 +1,73 @@
+## Question 1
+
+
+1 - What are atomic variables?
+
+Atomic variables are variables that provide thread-safe operations without using explicit locks. Operations such as incrementing or updating a value are performed atomically, meaning they cannot be interrupted by other threads.
+
+Ordinary variables do not provide this guarantee. For example, the expression `counter++` consists of multiple steps, and different threads may interfere with each other, causing race conditions.
+
+Atomic variables are mainly used to safely share data between multiple threads.
+
+## Question 2
+
+2 - Name at least four classes from the java.util.concurrent.atomic package that provide atomic operations for different data types.
+
+Some classes from the `java.util.concurrent.atomic` package are:
+
+- AtomicInteger
+- AtomicLong
+- AtomicBoolean
+- AtomicReference
+
+A common use case for `AtomicInteger` is a shared counter. Multiple threads can safely increment the counter without using locks.
+
+## Question 3
+
+3 - Compare locks with atomic variables.
+
+Atomic variables are usually faster and simpler for operations on a single variable. They have lower overhead and do not require explicit locking.
+
+Locks are better when multiple operations or multiple shared variables must be protected together. They provide more flexibility but usually have higher overhead.
+
+Use atomic variables for simple updates such as counters. Use locks for complex critical sections that involve several operations.
+
+## Question 4
+
+4 - A program is completely free of race conditions but still performs poorly under high contention.
+
+A program can be free of race conditions and still perform poorly because correctness does not guarantee scalability.
+
+Some factors that limit performance are:
+
+1. Lock contention, where many threads wait for the same lock.
+2. Thread blocking, which reduces parallel execution.
+3. Synchronization overhead, which adds extra work for coordinating threads.
+
+As a result, the program remains correct but may not scale well under heavy load.
+
+## Question 5
+
+5 - Many concurrent systems experience performance degradation as the number of threads increases.
+
+Adding more threads does not always improve performance.
+
+- Context switching takes CPU time when the operating system switches between threads.
+- Contention happens when many threads compete for the same resources.
+- Cache coherence creates extra work for processors to keep shared data consistent.
+- Synchronization overhead increases because locks and coordination mechanisms require additional processing.
+
+Because of these costs, too many threads can actually reduce performance.
+
+## Question 6
+
+6 - Deadlocks often only appear in production, not during testing.
+
+Deadlocks often depend on specific thread schedules. During testing, the required scheduling order may never occur. In production, higher load and different timing make deadlocks more likely.
+
+Two ways to expose deadlocks during testing are:
+
+1. Stress testing with many threads and repeated executions.
+2. Adding random delays to create different thread interleavings.
+
+These techniques increase the chance of reproducing deadlock situations before deployment.