Start here if you are new
Follow Physical Hardware, Memory Hierarchy, then Execution Model before jumping into tuning tools. That sequence makes the later pages much easier to interpret.
Main GPU Hub / Hardware + Execution + Performance + Tools
For AI Developers
GPU HUB
START HERE
This is your main GPU page. From here you can jump into architecture, execution internals, performance analysis, and practical GPU tools. The UI is intentionally clean white while keeping the technical style.
Learning Tracks
7 sections
Interactive Tools
15+ tools
Supported GPUs
50+ GPUs
GPU LEARNING PATH
01
Physical Hardware
Physical hardware foundations: SMs, cores, memory buses, and board design.
Explore02
Memory Hierarchy
Registers, shared memory, L2, VRAM, and how data moves across tiers.
Explore03
Execution Model
Warps, blocks, grids, scheduling behavior, and runtime execution flow.
Explore04
Compilation Pipeline
From CUDA source to PTX, SASS, and final kernel binaries.
Explore05
CUDA Programming
Kernel design, memory access patterns, synchronization, and tuning.
Explore06
Driver Stack
CUDA driver/runtime interactions, scheduling layers, and device control.
Explore07
Libraries & Frameworks
CUDA libraries and ML frameworks that map workloads onto GPU kernels.
ExplorePRECISION TOOLS
Tool
VRAM Calculator
Estimate memory footprint by parameters, precision, sequence length, and mode.
Open VRAM ToolTool
Kernel Occupancy Estimator
Model active warps and identify register/shared-memory pressure quickly.
Open Occupancy ToolTool
GPU Picker
Shortlist GPU options for training, fine-tune, and high-throughput inference.
Open GPU PickerTool
Warp Divergence Visualizer
See how branch conditions split a warp into active and waiting execution passes.
Open Divergence ToolTool
Roofline Model Analyzer
Plot kernels against memory and compute ceilings to see the real bottleneck fast.
Open Roofline ToolHARDWARE INDEX
| GPU | Arch | VRAM | FP16 TFLOP | Mem BW | Compute | Best For |
|---|---|---|---|---|---|---|
| NVIDIA H100 SXM5 | Hopper | 80GB HBM3 | 1,979 | 3.3 TB/s | 9.0 | Large-scale training |
| NVIDIA A100 SXM4 | Ampere | 80GB HBM2e | 312 | 2.0 TB/s | 8.0 | General DL workloads |
| RTX 4090 | Ada Lovelace | 24GB GDDR6X | 82.6 | 1.0 TB/s | 8.9 | Local inference |
| L40S | Ada Lovelace | 48GB GDDR6 | 183 | 0.8 TB/s | 8.9 | Multi-model serving |
NVIDIA H100 SXM5
Large-scale training
Arch
Hopper
VRAM
80GB HBM3
FP16
1,979
Mem BW
3.3 TB/s
Compute
9.0
Best For
Large-scale training
NVIDIA A100 SXM4
General DL workloads
Arch
Ampere
VRAM
80GB HBM2e
FP16
312
Mem BW
2.0 TB/s
Compute
8.0
Best For
General DL workloads
RTX 4090
Local inference
Arch
Ada Lovelace
VRAM
24GB GDDR6X
FP16
82.6
Mem BW
1.0 TB/s
Compute
8.9
Best For
Local inference
L40S
Multi-model serving
Arch
Ada Lovelace
VRAM
48GB GDDR6
FP16
183
Mem BW
0.8 TB/s
Compute
8.9
Best For
Multi-model serving
Start here if you already deploy models
Open the VRAM calculator and GPU picker first, then use the learning pages only where you need more explanation about bottlenecks or architecture tradeoffs.
Related decision guides
Pair this hub with the GPU and budget framework and low-VRAM model planning for end-to-end hardware decisions.