Fermi (microarchitecture)

Fermi is the codename for a graphics processing unit (GPU) microarchitecture developed by Nvidia, released to retail in April 2010 as the successor to the Tesla microarchitecture.

Uses your system voice; voice and rate in settings.

Snapshot

≈ 5 min read

Names in other languages

No names in other languages are held for this entry yet.

Article bodies are English and are not machine-translated. A name in another language is not a translation of the article.

Fermi is the codename for a graphics processing unit (GPU) microarchitecture developed by Nvidia, released to retail in April 20101 as the successor to the Tesla microarchitecture. It was the primary microarchitecture used in the GeForce 400 series and 500 series. All desktop Fermi GPUs were manufactured in 40nm, mobile Fermi GPUs in 40nm and 28nm .

Fermi is the oldest microarchitecture from Nvidia that receives support for Microsoft's rendering API Direct3D 12 feature_level 11. Fermi was followed by Kepler, and used alongside Kepler in the GeForce 600 series, GeForce 700 series, and GeForce 800 series, in the latter two only in mobile GPUs.

In the workstation market, Fermi found use in the Quadro x000 series, Quadro NVS models, and in Nvidia Tesla computing modules. The architecture is named after Enrico Fermi, an Italian physicist.

Load/Store Units

Allow source and destination addresses to be calculated for 16 threads per clock. Load and store the data from/to cache or DRAM.

Special Functions Units (SFUs)

Execute transcendental instructions such as sin, cosine, reciprocal, and square root. Each SFU executes one instruction per thread, per clock; a warp executes over eight clocks. The SFU pipeline is decoupled from the dispatch unit, allowing the dispatch unit to issue to other execution units while the SFU is occupied.

Integer Arithmetic Logic Unit (ALU)

Supports full 32-bit precision for all instructions, consistent with standard programming language requirements. It is also optimized to efficiently support 64-bit in workstation and server models, but artificially crippled for consumer versions.

Floating Point Unit (FPU)

Implements the new IEEE 754-2008 floating-point standard, providing the fused multiply-add (FMA) instruction for both single and double precision arithmetic. Up to 16 double precision fused multiply-add operations can be performed per SM, per clock.

Dual Warp Scheduler

At the SM level, each warp scheduler distributes warps of 32 threads to its execution units. Each SM features two warp schedulers and two instruction dispatch units, allowing two warps to be issued and executed concurrently. The dual warp scheduler selects two warps, and issues one instruction from each warp to a group of 16 cores, 16 load/store units, or 4 SFUs.

Most instructions can be dual issued; two integer instructions, two floating instructions, or a mix of integer, floating point, load, store, and SFU instructions can be issued concurrently. Double precision instructions do not support dual dispatch with any other operation.

Registers

Each SM has 32K of 32-bit registers. Each thread has access to its own registers and not those of other threads. The maximum number of registers that can be used by a CUDA kernel is 63.

The number of available registers degrades gracefully from 63 to 21 as the workload (and hence resource requirements) increases by number of threads. Registers have a very high bandwidth: about 8,000 GB/s.

L1+Shared Memory

On-chip memory that can be used either to cache data for individual threads (register spilling/L1 cache) and/or to share data among several threads (shared memory). This 64 KB memory can be configured as either 48 KB of shared memory with 16 KB of L1 cache, or 16 KB of shared memory with 48 KB of L1 cache.

Shared memory enables threads within the same thread block to cooperate, facilitates extensive reuse of on-chip data, and greatly reduces off-chip traffic. Shared memory is accessible by the threads in the same thread block.

It provides low-latency access (10-20 cycles) and very high bandwidth (1,600 GB/s) to moderate amounts of data (such as intermediate results in a series of calculations, one row or column of data for matrix operations, a line of video, etc.). David Patterson says that this Shared Memory uses idea of local scratchpad

Local Memory

Local memory is meant as a memory location used to hold "spilled" registers. Register spilling occurs when a thread block requires more register storage than is available on an SM. Local memory is used only for some automatic variables (which are declared in the device code without any of the __device__, __shared__, or __constant__ qualifiers).

Generally, an automatic variable resides in a register except for the following: (1) Arrays that the compiler cannot determine are indexed with constant quantities; (2) Large structures or arrays that would consume too much register space; Any variable the compiler decides to spill to local memory when a kernel uses more registers than are available on the SM.

L2 Cache

768 KB unified L2 cache, shared among the 16 SMs, that services all load and store from/to global memory, including copies to/from CPU host, and also texture requests. The L2 cache subsystem also implements atomic operations, used for managing access to data that must be shared across thread blocks or even kernels.

Global memory

Global memory (VRAM) is accessible by all threads directly as well as the host system over the PCIe bus. It has a high latency of 400-800 cycles.

General

N. Brookwood, "NVIDIA Solves the GPU Computing Puzzle." P.N. Glaskowsky, "NVIDIA’s Fermi: The First Complete GPU Computing Architecture." N. Whitehead, A. Fit-Florea, "Precision & Performance: Floating Point and IEEE 754 Compliance for NVIDIA GPUs.", 2011.

R. Farber, "CUDA Application Design and Development," Morgan Kaufmann, 2011. NVIDIA Application Note "Tuning CUDA applications for Fermi". NVIDIA Fermi Architecture on Orange Owl Solutions

Text derived from https://en.wikipedia.org/w/index.php?oldid=1355823796 under CC BY-SA 4.0; modifications have been made.

References

  1. ↩Wikipedia “Fermi (microarchitecture)” revision 1355823796 · retrieved 2026-10-20 release date (×1) · confidence 0.95 revision URL