Nvidia Fermi
Logs
Memory ctrl.
GDDR3GDDR5DDR3
Features
- CUDA
- Direct3D 12
- OpenCL 1.2
- OpenGL 4.6
- Shader Model 5.1
Actions
| Core Execution Block (e.g., SM) | |
|---|---|
| Block Name | Streaming Multiprocessor (SM) |
| Cores per Block | 32 |
| Texture Units per Block | 4 |
| Dispatch Units per Block | 2 (Warp Schedulers) |
| Shared Memory L1 Cache per Block | 64 KB (Configurable as 48K Shared/16K L1 or 16K Shared/48K L1) |
| Feature and API Set | |
|---|---|
| Maximum OpenGL Support | 4.6 |
| CUDA Compute Capability | 2.0 |
| Parallel Computing Features | Concurrent Kernel Execution, ECC Support, C++ Support |
| Maximum DirectX Feature Level | 11.0 |
| Architecture Identity | |
|---|---|
| Architecture Name | Fermi 1.0 |
| Primary Market Segment | Desktop, Workstation, HPC |
| Successor Architecture | Fermi 2.0 |
| Predecessor Architecture | Tesla |
| Memory Architecture | |
|---|---|
| ECC Support | Yes (GF100-based models) |
| L2 Cache Architecture | Unified L2 Cache |
| Memory Controller Support | GDDR5, GDDR3, DDR3 |
| Video Processing Engine | |
|---|---|
| Video Decode Engine Version | VP4 |
| Video Encode Engine Version | N/A (Introduced in Kepler) |
No documents available
Notes
DirectX 11/OpenGL 4.6 class GPU architecture.
Used by Nvidia GF100, GF104, GF106, GF108, GF110, GF114, GF116, GF117 and GF119 chips.
All Nvidia Fermi GPUs are DirectX 12 (feature level 11_0).
Disclaimer
The info found in this page might not be entirely correct. Check out this guide to learn how you can improve it.