Keep the specifications close.
Start with the explanation here, then use the manuals to check the details of your target. These links open public documentation, papers, and vendor references.
How to read this reference
The articles explain general mechanisms. Vendor examples are identified as examples, and interactive models state what they leave out. Specifications evolve; check the exact architecture, generation, document revision, and supported features before writing device-specific code.
Reading links reviewed: October 2, 2026.AMD · Zen core architecture
Core architecture and processor families.
Linux · NUMA memory
Memory locality and placement.
TianoCore · UEFI driver architecture
Firmware interfaces and boot handoff.
AMD · FPGA lookup tables
An FPGA logic implementation.
Arm · Digital signal processing
Signal-processing hardware and instructions.
Linux · Network scaling
RSS, RPS, and packet placement.
Linux · Spectre side channels
Speculation and mitigation boundaries.
TSMC · Transistor structures
Transistor geometry and process research.
PCI-SIG · Public FAQ
Public interconnect definitions.
Linux · MSI driver guide
Message-signaled interrupts.
LLNL · Parallel computing
Parallel work and scaling laws.
LLVM · Language reference
IR, optimization contracts, and memory.
Oracle · Floating-point arithmetic
Representation and rounding behavior.
Apple · M4 chip architecture
An example of CPU, GPU, and Neural Engine integration.
Nand2Tetris · Boolean logic
Build gates from NAND.
Nand2Tetris · Memory
Latches, registers, and memory.
Intel · Optimization reference manuals
Microarchitecture and performance details.
RISC-V · ISA specifications
Instruction semantics and privilege.
NVIDIA · CUDA programming guide
GPU execution and memory model.
NVIDIA · Writing SIMT kernels
Threads, blocks, and warps.
AMD · RDNA performance guide
Wavefronts, occupancy, and GPU optimization.
Cornell · Caches
Locality, cache organization, and misses.
Micron · DDR5 SDRAM
Banks, refresh, and DDR generations.
Linux · Memory barriers
Ordering and synchronization.
Linux · DMA API guide
Device memory access and mapping.
NVM Express · Specifications
Storage command and transport specifications.
Arm · big.LITTLE
Heterogeneous CPU cores.
Qualcomm · Hexagon NPU
A vendor example of a neural accelerator.
Qualcomm · Oryon CPU
CPU branding versus microarchitecture.
Khronos · Vulkan synchronization
Execution and memory dependencies.
NVIDIA · Ray tracing architecture
BVH traversal and intersection hardware.
Linux · Page tables
Address translation and permissions.
Micron · Solid-state drives
Flash management and endurance.
Intel · False sharing
Diagnose cache-line contention.
Google · Tensor Processing Unit paper
An accelerator architecture study.
Linux · CPU performance scaling
Governors and operating states.
UC Berkeley · Roofline model
Compute and memory upper bounds.
Linux · Workload tracing
Counters and workload observation.
Linux · PCI drivers
Configuration, device resources, and interrupts.