CISC vs RISC: How Complex and Simple Instruction Sets Compete
This article explains the fundamental differences between CISC and RISC architectures, covering their historical origins, design philosophies, characteristic trade‑offs, representative implementations, modern hybrid approaches, and the rising impact of ARM and RISC‑V in today’s computing landscape.
1. What is an instruction set?
An instruction is the basic operation a CPU performs, such as addition, jump, or memory access. The instruction set is the complete collection of instructions a CPU supports, essentially the processor’s "martial arts manual."
指令 = CPU执行的基本操作 比如:
- 加法指令:把两个数相加
- 跳转指令:跳到某个位置继续执行
- 读写内存指令:存取数据
指令集 = CPU支持的所有指令的集合 = CPU的"武功秘籍"2. CISC: Complex Instruction Set
Full name
Complex Instruction Set Computer (CISC)
Representative implementations
- x86/x64 (Intel, AMD)
- IA-64 (Itanium)
- Evolution: 8086 → 80286 → 80386 → Pentium → …Core idea
"用一条指令完成复杂任务"
Example: string copy
CISC: MOVS (single instruction copies a string)
RISC: needs a loop of LOAD/STORE instructionsCharacteristics
Instruction traits:
- Variable length (1–15 bytes)
- Variable execution time
- Many addressing modes
- Few registers (8–16 general‑purpose)
Example x86 instructions:
MUL BX ; AX = AX * BX
REP MOVSB ; copy string in one instructionWhy CISC has complex instructions?
Historical background (1970s): memory was expensive, so complex instructions aimed to reduce instruction count and memory usage, and simplify programming. In practice, decoding complex instructions is slow, consumes many transistors, and yields limited performance gains.3. RISC: Reduced Instruction Set
Full name
Reduced Instruction Set Computer (RISC)
Representative implementations
- ARM (dominant RISC)
- MIPS (academic example)
- RISC‑V (open‑source newcomer)
- PowerPC (IBM’s RISC)Core idea
"Simple instructions are fast; many of them can be combined to perform complex work"
Principles:
1. Fixed instruction length (4 bytes)
2. Fixed execution time (usually 1 cycle)
3. Simple addressing modes
4. Many registers (16–32 general‑purpose)Characteristics
Instruction traits:
- Fixed length
- Only LOAD/STORE can access memory
- Many registers
- Small instruction set (tens to a few hundred)
Example ARM instructions:
ADD R0, R1, R2 ; R0 = R1 + R2
LDR R0, [R1] ; load from memory
STR R0, [R1] ; store to memory4. Full comparison
Instruction count: CISC – many (hundreds); RISC – few (tens)
Instruction length: CISC – variable; RISC – fixed
Execution cycles: CISC – variable; RISC – fixed
Registers: CISC – few; RISC – many
Memory access: CISC – flexible; RISC – LOAD/STORE only
Decoding difficulty: CISC – complex; RISC – simple
Power consumption: CISC – higher; RISC – lower
Cost: CISC – higher; RISC – lower
Representative architecture: CISC – x86; RISC – ARM
5. Why ARM is so successful?
ARM = Advanced RISC Machine
Success factors:
1. Low power → preferred for mobile devices
2. Licensing model → used by Apple, Samsung, Qualcomm, etc.
3. Simple design → fewer transistors, low heat
4. Pipeline‑friendly → high‑frequency operation
Market share:
- >95 % of smartphone processors are ARM
- Dominates embedded market
- Apple M‑series ARM chips are challenging x86 in PCs6. Modern processors “fusion”
In reality the line between CISC and RISC is blurring.
Intel’s approach
x86 processors internally:
- External: compatible x86 instructions (CISC)
- Internal: translate CISC to micro‑operations (RISC‑like)
- Execution: RISC‑style pipeline
This “decode‑and‑execute” technique started with Pentium Pro (1995), the first CPU with a RISC core behind a CISC façade.ARM’s evolution
Early ARM: pure RISC
Modern ARM: adds complex instruction accelerators such as NEON SIMD, hardware floating‑point, and prefetch optimisations – moving toward “complex” capabilities.Apple M1 case
M1 = ARM architecture (RISC) but outperforms x86 because of:
1. Large + many cores
2. Unified memory architecture
3. Advanced micro‑architecture design
4. 5 nm process
Illustrates that the instruction set is only a starting point; micro‑architecture matters most.7. Pipeline: RISC’s accelerator
RISC’s speed comes from pipelining.
Five‑stage pipeline: IF → ID → EX → MEM → WB (fetch, decode, execute, memory, write‑back)
At any moment five instructions are in different stages, giving up to 5× throughput.
RISC suits pipelines because:
- Fixed instruction length → simple decode
- Fixed execution cycle → easy scheduling
- Many registers → fewer data dependencies8. RISC‑V: Open‑source new force
RISC‑V:
- Born 2010 at UC Berkeley
- Fully open and free, no licensing fees
- Simple ISA
China’s enthusiasm:
- Not subject to US export controls
- Enables independent design
- Fits IoT chips
Notable shipments: Alibaba’s Pingtouge >30 billion chips; Huawei, Zhaoyi also have RISC‑V product lines.9. Summary
CISC (x86):
- Strong compatibility, rich ecosystem
- Higher power, suited for desktop/server
- Intel/AMD dominate
RISC (ARM):
- Low power, ideal for mobile/embedded
- Open licensing, high flexibility
- Expanding into PC/server
Future trend:
- Fusion: RISC‑style cores with CISC compatibility, and CISC learning RISC design
- ARM challenging x86 in PCs/servers
- RISC‑V open‑source riseInstruction sets are just the starting point; the real battle is in micro‑architecture!
Key takeaways:
CISC instructions are complex and numerous; x86 is the flagship.
RISC instructions are simple and few; ARM is the flagship.
RISC fits pipelines, offers low power.
CISC offers better compatibility and a mature ecosystem.
Modern CPUs are converging toward a hybrid design.
Signed-in readers can open the original source through BestHub's protected redirect.
This article has been distilled and summarized from source material, then republished for learning and reference. If you believe it infringes your rights, please contactand we will review it promptly.
IT Learning Made Simple
Learn IT: using simple language and everyday examples to study.
How this landed with the community
Was this worth your time?
0 Comments
Thoughtful readers leave field notes, pushback, and hard-won operational detail here.
