The model ships in FP8. Should you requantize to four bits to fit more of it on fewer GPUs?

Halving the weights halves the GPU count and moves the quality evaluation onto you. What the format actually buys in memory and in speed, why those are different questions, and the evaluation that has to run before it ships.

Updated Sep 2026 · Grounded in real AI infrastructure interview loops and written to a senior-engineer editorial bar, with every number worked and every diagram hand-built.

more free answers with an account · no card

The concepts behind this question

Ranked by how closely each one overlaps this question's topic, so the first card is the thing to read if the answer above moved too fast.

Foundational

🧮 Open Weights & Serving Engines

Weight Formats: FP8 Blocks, MXFP4 and AWQOpen-weights models now ship pre-quantized, and the format is part of the release rather than something you choose afterwards. Block-scaled FP8 gives one byte per parameter with a scale per tile. MXFP4 gives about 0.53 bytes by pairing four-bit values with a shared exponent every 32 elements. Integer schemes like AWQ reach similar sizes with a different error profile. What decides a deployment is not which is most accurate in the abstract but which one the model was released and evaluated in, and which one your engine and hardware can execute natively.

Foundational

🧮 Open Weights & Serving Engines

Serving Benchmarks That Do Not LieMost published serving numbers are not comparable to each other and not predictive of production, because they differ in the input distribution, the concurrency, whether the cache was warm, and which of several very different metrics is being reported. A benchmark that supports a decision has to fix all four, report a distribution rather than a mean, and be run against the traffic shape you actually serve. The single most useful discipline is to compute the bandwidth bound first, so you know what fraction of the possible you achieved.

Advanced

🚀 Inference & Serving🔒 Premium

Quantization for InferenceQuantization stores weights, and sometimes activations and the KV cache, in fewer bits, which cuts the bytes a decode step has to stream and the memory a model occupies. Weight-only int4 (GPTQ, AWQ) is a capacity and single-stream latency play; fp8 for weights and activations (W8A8) doubles tensor-core throughput and helps prefill and large batch; fp8 KV cache doubles context per GPU. Each has an accuracy cost you measure rather than assume, and knowing which one to reach for from the bottleneck is the interview question.

Core

🧩 GPU & Accelerator ArchitectureSign in

Numerics: FP32, BF16, FP8 and FP4Every number format is a trade between range (exponent bits), precision (mantissa bits) and throughput (fewer bits, more values per cycle through the tensor cores). bf16 won training because it keeps fp32's range; fp8 splits into E4M3 for precision and E5M2 for range and needs scaling factors; fp4 needs block scaling and careful outlier handling. Knowing which format goes where, and why accumulation stays fp32, is what the numerics question is really asking.

Explore the full curriculum →

UP NEXT ON YOUR JOURNEY

Next in this trackPrefix caching cut your benchmark's latency in half. Why might production see none of that?Next in this trackWould you split prefill and decode onto separate pools for a 753B model? Work the decision.Next in this trackA multi-node vLLM deployment starts, forms a cluster, and hangs on the first request. Debug it.

FEDITOR'S NOTE

Scored on separating the memory gain from the speed gain, on native versus emulated execution, and on owning the evaluation once you deviate from the released format.

DISCUSSION · 0

No comments yet — be the first to share your approach.