Tag: design-hardware
All the articles with the tag "design-hardware".
-
System Design - How Floating Point Arithmetic Works (IEEE 754)
How IEEE 754 floating point works: the sign-exponent-mantissa bit layout, normalization, special values (infinity, NaN, denormals), rounding modes, and why 0.1 + 0.2 != 0.3 — with binary walkthroughs and real source code.
-
System Design - How CPU SIMD and Vector Instructions Work
How CPU SIMD (Single Instruction, Multiple Data) works: the evolution from MMX to AVX-512, register layouts, data-parallel execution, auto-vectorization, intrinsics, and real-world source code examples from databases and machine learning libraries.
-
System Design - How CPU Out-of-Order Execution Works
How modern CPUs execute instructions out of order: the fetch-decode-rename-dispatch-execute-retire pipeline, register renaming, reservation stations, reorder buffer, and how it all maintains the illusion of sequential execution.
-
System Design - How SSDs Work Internally (NAND Flash, FTL, and Write Amplification)
How Solid-State Drives work internally: NAND flash physics, pages and blocks, the Flash Translation Layer (FTL), garbage collection, wear leveling, write amplification, TRIM, and why understanding SSDs matters for database engineers.
-
System Design - How CPU Branch Prediction Works
How modern CPUs predict branch outcomes: pipeline stalls, static vs dynamic prediction, branch history tables, two-level adaptive predictors, branch target buffers, and real-world performance implications with code examples.