⬡ CacheSim
Silicon Memory Coherence Lab
GitHub Repo ↗

Lab Controller

Address Register Decomposition

AWAITING CYCLE
Tag
Idx
Off
----------------------------------------------------------------
Tag: 0x-- Index: -- Offset: -- Way: --

SRAM Memory Cell Matrix

Click RUN SIMULATION to map cache layout.

Lab Timeline Scrubber Debugger

0 / 0
Playback Speed:

Silicon Performance Analytics

Hit Rate 0.0%
Miss Rate 0.0%
Memory Accesses 0
Writebacks 0
Hit Rate Convergence Timeline
3C Miss Triage
Cold Conflict Capacity

Silicon Optimization Explorer

Memory bus transaction stream log

Cycle Address Op Result Triage Set Way

Silicon Lecture hall

📚 Address bits partition decomposition
Type Hex Address:
📊 Coherent AMAT Interactive Simulator
L1 local hit time: 4t
L1 miss rate: 10%
L2 local hit time: 12t
L2 miss rate: 20%
DRAM latency: 200t
🔍 replacement policies comparison

LRU evicts least-recently referenced lines. Optimal recency accuracy but high logic gates footprint.
FIFO evicts oldest loaded lines ignoring hits, avoiding age register updates but yielding Belady anomalies.
PLRU maps nodes in a binary direction bitmask. Flip decisions point away from accessed blocks, using only N-1 bits.

EDA Export Suite