liveBook · Manning

8-bit signed integer

As explained in Quantization and Fast Inference

INT8 is an 8-bit signed or unsigned integer representation used for low-precision numerical computation and storage.

INT8 represents values with uniformly spaced levels. In symmetric INT8 quantization, the scale sets the maximum representable magnitude; the spacing therefore depends on that scale. section 8.2.1

Choose between INT8 and other low-precision formats by considering the value distribution and scaling granularity:

  • For values concentrated near zero, FP8 E4M3 has 49 representable values in [0, 0.5], compared with INT8 symmetric's 32. For a uniform distribution, INT8 is identified as the better choice. section 8.2.1
  • For weight reconstruction, per-channel INT8 scaling achieved lower MSE than per-tensor FP8 in the OPT-6.7B fc1 comparison; reconstruction MSE alone does not determine end-to-end perplexity. section 8.2.3
Figure 8.2 Representable values in [0, 8]. E4M3 and E5M2 cluster their values exponentially—dense near zero, sparse at larger magnitudes. INT8 symmetric spaces all 128 positive values uniformly, so its coverage ends abruptly at the scale-dependent maximum (here ≈2, assuming scale = 0.0157). Beyond that boundary, INT8 cannot represent any value at all.

Compare how INT8's uniform spacing and scale-dependent range differ from FP8's denser values near zero. section 8.2.1

INT8 in a PTQ decision

A large CNN targeting INT8 with a 2% accuracy tolerance is described as a PTQ-trivial case; architecture, target bit-width, and tolerated degradation inform whether to escalate to QAT. section 4.1

How Quantization and Fast Inference covers it

How deep it goes
Explains representable-value density and compares weight reconstruction MSE, while cautioning that MSE alone does not establish end-to-end perplexity.
Examples
Compares near-zero value counts and OPT-6.7B fc1 weight reconstruction across INT8 and FP8 formats.
What you can do after
Evaluate INT8 against FP8 using value distributions, scaling granularity, and the intended quantization workflow.
How it connects
Connects INT8 to FP8, per-channel scaling, post-training quantization, and quantization-aware training.

Also explained in

Compare how 2 books explain 8-bit signed integer →

Related concepts

This explainer was generated by Manning's knowledge graph from the text of Quantization and Fast Inference. The section links point to the parts of the book it draws on.