speculative: --spec-draft-window and --spec-draft-n-max-tail (MTP drafting at every depth)

[professorpalmer]

One of the five pieces of #285, split per bri-prism's request on #317. Single commit on current prism (eaecb50c7), applies without conflicts.

What it does

--spec-draft-window N: the draft (MTP) context keeps only the last N rows. The server drops older rows before each batch, and the draft context is sized for the window (common_speculative_init), so its cells are reused and a draft pass costs the same at any depth. This retires the --spec-draft-depth-max 24576 cutoff from #221 (the flag stays, default 0). --spec-draft-n-max-tail K sets the draft size from the tiered-KV line (--kv-vram-cells) on; the MTP draft loop now honours the per-draft cap. Without the tiered-KV PR the tail flag is inert; the window works on its own.

Receipts (RTX 4070, Bonsai 2 27B PTQ1_0 with the MTP head, q8_0 K/V)

• Drafting at every depth instead of stopping at 24k: 32k 54.5 -> 103.6 tok/s, 64k 48.0 -> 90.1, acceptance unchanged. These numbers were taken together with cuda+server: q8_0 K/V at the full 262k window on 12 GB (tiered VMM KV cache), MTP drafting at every depth #285's FA GQA-decode commit (now going into cuda: use the MMA flash attention kernel for GQA above 4 with quantized K/V on Ada #307 with the ggml_cuda_fattn_mma_kv_native_supported guard); the window alone removes the depth cutoff, the speed at depth depends on the attention kernel the decode step uses.
• Tail draft 4 vs 2 past the VRAM line: +26% on code, +15% on prose at 180k (a PCIe-bound step makes extra verify columns nearly free).
• Draft size is a quality-neutral knob: greedy continuations stay valid greedy decodes; the identity notes on drafting from cuda+server: q8_0 K/V at the full 262k window on 12 GB (tiered VMM KV cache), MTP drafting at every depth #285 apply (draft-on is not always byte-identical to draft-off at the rounding level, with the stock kernels too, because the FA KV split follows the kernel instance and the padded KV length).

Serving recipe and receipts: https://github.com/professorpalmer/bonsai-ada-surgery/blob/main/docs/Q8_FULL_CONTEXT.md