How Can We Help?
| Floating-point formats |
|---|
| IEEE 754 |
|
| Other |
| Alternatives |
TensorFloat-32 or TF32 is a numeric floating point format designed for Tensor Core running on certain Nvidia GPUs.
Format
The binary format is:
- 1 sign bit
- 8 exponent bits
- 10 fraction bits (also called mantissa, or precision bits)
The total 19 bits fits within a double word (32 bits), and while it lacks precision compared with a normal 32 bit IEEE 754 floating point number, provides much faster computation, up to 8 times on a V100.[1]
See also
References
- ^ https://deeprec.readthedocs.io/en/latest/NVIDIA-TF32.html accessed 23 May 2024
External links
Recent Comments