Nvidia GPU Architecture
来自Jack's Lab
目录 |
1 Overview
The GPU architecture is built around a scalable array of Streaming Multiprocessors (SM)
The key components of a Fermi SM:
- CUDA cores (ALU + FPU)
- Load/Store Units
- Special Function Units
- Register File
- Shared Memory/L1 Cache
- Warp Scheduler
2 Fermi Micro Architecture
The Fermi architecture was the first complete GPU computing architecture to deliver the features required for the most demanding HPC applications.
- 16 SM (32 CUDA cores/SM)
- 512 CUDA cores (1 ALU + 1 FPU per core)
- 1 SM: 32 CUDA cores + 16 Load/Store Unit + 4 SPU + 2 Warp Scheduler
- 1 SM: 2 Warps
- 1 Warp: 16 CUDA cores + 16 Load/Store Unit(shared) + 4 SPU(shared) + [32 threads context ?]
- Handle 48 warps per SM for a total of 1536 (48x32) threads resident in a single SM at a time [48 Warps context ?]
- Register file is 32KB
- 1345 GFLOPs
- 40 nm
- 3.2 billion transistors
- GTX480 250Watts
2.1 Fermi Cards
2.1.1 GeForce 400 Series
- Release date: April 12, 2010
- Codename: GF10x
- Architecture: Fermi
- Models:
- GeForce Series
- GeForce GT Series
- GeForce GTS Series
- GeForce GTX Series
- Fabrication process and transistors:
- 260M 40nm (GT218)
- 585M 40 nm (GF108)
- 1.170M 40 nm (GF106)
- 1.950M 40 nm (GF104)
- 1.950M 40 nm (GF114)
- 3.200M 40 nm (GF100)
- Cards:
- Entry-level GT420 GT430
- Mid-range GT440 GTS450 GTX460
- High-end GTX465 GTX470
- Enthusiast GTX480 (2010.3, 3.2 billion Transistors, 15 SMs, 1536MB, 1345 GFLOPS, 250W)
2.1.2 GeForce 500 Series
- Release date: 8 November 2010
- Codename: GF11x
- Architecture: Fermi
- Models:
- GeForce Series
- GeForce GT Series
- GeForce GTX Series
- Fabrication process and transistors:
- 292M 40nm (GF119)
- 585M 40 nm (GF108)
- 1.170M 40 nm (GF116)
- 1.950M 40 nm (GF114)
- 3.000M 40 nm (GF110)
- Cards:
- Entry-level 510 GT520 GT530
- Mid-range GT545 GTX550Ti GTX560 GTX560Ti
- High-end GTX570 GTX580 GTX590(2011.3, 2x3 billion transistors, 32 SMs, 2x1536MB, 2488GFLOPS, 365W)
3 Kepler Micro Architecture
Released in the fall of 2012
- 15 SM (192 CUDA cores per SM)
- 512 CUDA cores (1 ALU + 1 FPU per core)
- 1 SM: 192 CUDA cores + 64 DPU + 32 Load/Store Unit + 32 SPU + 4 Warp Scheduler (8 instruction dispatchers)
- 1 SM: 4 Warps
- 1 Warp: 16 CUDA cores + 16 Load/Store Unit(shared) + 4 SPU(shared) + [32 threads context ?]
- Handle 64 warps/SM for a total of 2048 (64x32) threads resident in a single SM at a time [64 Warps context ?]
- Register file size is 64K
- 2x2810 GFLOPs
- 28 nm
- 2x3.54 billion transistors
- GTX690 300Watts (2012.4)
4 Maxwell Micro Architecture
GM204:
- 16 SM
- 2048 CUDA cores
- 4612 GFLOPs
- 28 nm
- 5.2 billion transistors
- GTX980 165W
5 Reference
- https://en.wikipedia.org/wiki/Fermi_(microarchitecture)
- https://en.wikipedia.org/wiki/GeForce_400_series
- https://en.wikipedia.org/wiki/GeForce_500_series