In a typical processor instruction set, a load reads a value from a memory address into a register; a store writes a value from a register to a memory address. The terms describe opposite directions of data movement, but their exact source and destination depend on the context: JVM bytecode, for example, uses them for transfers between local variables and an operand stack.
How load and store work in a processor
Memory holds data at addresses. A processor uses registers—small, fast storage locations inside the processor—to work with values. In the common processor-level meaning, a load fetches data from an addressed memory location and places it in a register. A store takes a value, usually from a register, and writes it to an addressed memory location.
As an Amazon Associate I earn from qualifying purchases.
The LLVM Language Reference describes the same basic memory actions at the compiler intermediate-representation level: load reads from memory, while store writes to memory. LLVM IR is not itself a complete definition of every processor’s instruction set.
Do these 3 things before closing this tab:
1Scan for outdated or missing drivers - takes under a minute2Repair Windows errors before they cause bigger problems3Fix the driver behind crashes, sound loss and screen glitchesLoad vs. store
| Operation | Direction | Typical effect |
|---|---|---|
| Load | Memory → register | Reads the value at an address and makes it available to the processor. |
| Store | Register → memory | Writes a value to an address in memory. |
Why load and store are explicit in some architectures
In a load-store architecture, instructions for arithmetic and other computation operate on register values; separate load and store instructions move data between registers and memory. MIT OpenCourseWare’s Beta architecture example uses LD to read memory into a register and ST to write register data to memory. In that architecture, these are the only instructions that access data memory.
#1 Best Overall
The Beta example calculates an effective address by adding a register value to a sign-extended 16-bit constant encoded in the instruction. That is a Beta-specific addressing detail, not a rule that applies to every processor. See the MIT OpenCourseWare Computation Structures materials for the architecture example.
Load and store in JVM bytecode
In Java Virtual Machine bytecode, the terms refer to movement within a method’s execution frame, not directly to a memory/register transfer. A load instruction moves a value from a local variable onto the operand stack; a store instruction moves a value from the operand stack into a local variable. The JVM specification lists typed families such as iload, lload, fload, dload, and aload, with corresponding store instructions. Consult the Java Virtual Machine Specification, instruction set for these definitions.
Is a load the same as moving an immediate value?
No. In the usual processor terminology, a load reads from an address in memory. An immediate-move instruction uses a value encoded directly in the instruction, rather than fetching that value from a memory address. Exact instruction names and behavior vary across instruction sets.
The distinction also matters in JVM bytecode: the specification lists instructions for loading constants separately from instructions that load values from local variables. So the word “load” alone does not tell you what the source is; check the architecture or virtual-machine specification for the instruction in question.
Quick Recap
Best Value
How to interpret the terms in a specific context
- Processor instruction or assembly: identify the address being read or written and the register carrying the value.
- LLVM IR: a load reads through a pointer operand; a store writes a specified value through a destination pointer. LLVM also defines details for variants such as atomic and volatile operations.
- JVM bytecode: determine whether the instruction transfers between a local variable and the operand stack, or belongs to the separate constant-loading family.
Product prices and availability are accurate as of the date/time indicated and are subject to change. Any price and availability information displayed on Amazon at the time of purchase will apply.
