Nvidia - AshokBhat/notes GitHub Wiki
About
Products
- GeForce, consumer-oriented graphics processing products
- Quadro, computer-aided design and digital content creation workstation graphics processing products
- NVS, multi-display business graphics solution
- Tegra, a SoC series for mobile devices
- Tesla, GPGPU for high-end image generation applications in professional and scientific fields
- DGX - Nvidia produced servers and workstations - GPGPU to accelerate deep learning applications
Development boards
- Nvidia Jetson - For embedded segments
GPU Microarchitectures
| Micro-architecture | Released | Segment |
|---|---|---|
| Maxwell | 2014 | General purpose |
| Pascal | 2016 | For consumer |
| Volta | 2017 | For servers and datacenters |
| Turing | 2018 | General purpose |
| Ampere | 2020 | All |
| Hopper | 2022 | All |
Hopper
- GA: Q3, 2022
- Transformer Engines to accelerate ML - Expected to have 6x improvement in performance for transformers-based networks.
- TSMC 4nm
CPU Microarchitectures
| CPU | Architecture | Launched | Process |
|---|---|---|---|
| Denver | ARMv8.0-A | 2014 | |
| Carmel | ARMv8.2-A | 2018 | TSMC 12nm |
Server CPUs
- Grace CPU
Roadmap
NVIDIA software
- NGC - Nvidia GPU cloud - GPU-Optimized Software - DL, ML, and HPC Workflows
Other hardware companies
- [Intel]] ](/AshokBhat/notes/wiki/[AMD) | [Arm]]
- [Nvidia DGX]] ](/AshokBhat/notes/wiki/[[Nvidia-Jetson) | Nvidia Tesla