Silvaco’s Ultra High-Speed cache memory is an adaptable,
independent, non-coherent cache Intellectual Property
(IP) featuring an advanced cache architecture. This
architecture enhances system performance, scalability,
power efficiency, data locality, application responsiveness,
cost optimization, and market competitiveness, providing
a distinctive business value.
Ultra High-Speed Cache Memory Compiler
Overview
Key Features
- Up to 3.4 GHz operation in N3P process
- Cache size up to 16 Kb
- 4 – 64-bit word width
- Configurable way associativity
- Custom logic-rules bitcell
- Speed / Static Power tradeoff option
Benefits
- Cache Types
- Data array
- Instruction array
- Virtual tag array
- Physical tag array, PTAG
- Cache Applications
- AI Processors
- Data Center
- High Performance Computing
- Network Compute Units
- Neural Processor
Block Diagram

Technical Specifications
Foundry, Node
TSMC, 3nm
Maturity
Silicon Proven
Availability
Now
Related IPs
- Viterbi Compiler, High-Speed Parallel Decoder
- Compact High-Speed 32-bit CPU Core with Level-2 Cache
- Bulk 40ULP single port SRAM Compiler - ultra low power, low power retention mode
- ARC EM11D Enhanced 32-bit processor core, ARCv2DSP ISA with Cache and XY Memory
- Memory Compiler
- L2 cache/cluster shared memory option for multicore versions of ARC HS5x and HS6x processors