feat: evaluate DSpark alongside DFlash 2, add DSpark llama-server presets

client/src/components/settings/LocalLlmTab.jsx

@@ -22,17 +22,36 @@ const BACKENDS = [
 ];
 const labelFor = (id) => BACKENDS.find((b) => b.id === id)?.label || id;
 
-const DFLASH_PRESETS = [
+// llama.cpp exposes one `--spec-type` per drafter family. `draft-dspark` merged
+// upstream on 2026-07-28 (ggml-org/llama.cpp#25173) and works on a stock
+// `brew install llama.cpp`; `draft-dflash` covers DFlash v1, while the DFlash 2
+// modules are still an open PR (#27342) and need a from-source build. Hence the
+// DSpark presets lead — see docs/research/2026-08-19-dspark-vs-dflash2.md.
+const SPEC_DECODE_PRESETS = [
+  {
+    id: 'qwen3.8-27b-dspark',
+    label: 'Qwen 3.8 27B + DSpark Drafter (Recommended — stock llama.cpp)',
+    model: 'models/Qwen3.8-27B-Instruct-Q4_K_M.gguf',
+    draftModel: 'models/Qwen3.8-27B-DSpark-bf16.gguf',
+    specType: 'draft-dspark',
+  },
+  {
+    id: 'qwen3-8b-dspark',
+    label: 'Qwen 3 8B + DSpark Drafter (small target)',
+    model: 'models/Qwen3-8B-Instruct-Q4_K_M.gguf',
+    draftModel: 'models/dspark_qwen3_8b_block7-bf16.gguf',
+    specType: 'draft-dspark',
+  },
   {
     id: 'qwen3.8-27b',
-    label: 'Qwen 3.8 27B + DFlash 2 Drafter (Recommended)',
+    label: 'Qwen 3.8 27B + DFlash 2 Drafter (needs llama.cpp PR #27342 build)',
     model: 'models/Qwen3.8-27B-Instruct-Q4_K_M.gguf',
     draftModel: 'models/Qwen3.8-27B-DFlash2-Q4_K_M.gguf',
     specType: 'draft-dflash',
   },
   {
     id: 'muse-glimmer-30b',
-    label: 'Muse-Glimmer 30B + DFlash 2 Drafter',
+    label: 'Muse-Glimmer 30B + DFlash 2 Drafter (needs llama.cpp PR #27342 build)',
     model: 'models/Muse-Glimmer-30B-Instruct-Q4_K_M.gguf',
     draftModel: 'models/Muse-Glimmer-30B-DFlash2-Q4_K_M.gguf',
     specType: 'draft-dflash',
@@ -42,7 +61,7 @@ const DFLASH_PRESETS = [
     label: 'Custom GGUF / Manual Paths',
     model: '',
     draftModel: '',
-    specType: 'draft-dflash',
+    specType: 'draft-dspark',
   },
 ];
 
@@ -411,7 +430,7 @@ export function LocalLlmTab() {
   const [llamaForm, setLlamaForm] = useState({
     model: '',
     draftModel: '',
-    specType: 'draft-dflash',
+    specType: 'draft-dspark',
     port: 8080,
     host: '127.0.0.1',
     ctxSize: 32768,
@@ -771,7 +790,7 @@ export function LocalLlmTab() {
   };
 
   const handlePresetSelect = (presetId) => {
-    const preset = DFLASH_PRESETS.find((p) => p.id === presetId);
+    const preset = SPEC_DECODE_PRESETS.find((p) => p.id === presetId);
     if (!preset) return;
     if (preset.id !== 'custom') {
       setLlamaForm((prev) => ({
@@ -940,7 +959,7 @@ export function LocalLlmTab() {
         </div>
 
         <p className="text-xs text-gray-400 leading-relaxed">
-          DFlash 2 enables multi-token speculative drafting (yielding 2.5–3× speedups on local models like Qwen 2.5 / Qwen 3.8 and Muse-Glimmer). You can launch and manage a local <code className="text-gray-300">llama-server</code> with DFlash 2 speculative drafting directly from PortOS and connect using the <strong className="text-white">OpenCode llama TUI</strong> provider.
+          Speculative decoding pairs a small drafter with your target model for 2–3× faster generation at identical output. You can launch and manage a local <code className="text-gray-300">llama-server</code> from PortOS and connect using the <strong className="text-white">OpenCode llama TUI</strong> provider. <strong className="text-white">DSpark</strong> (<code className="text-gray-300">draft-dspark</code>) works on a stock <code className="text-gray-300">brew install llama.cpp</code>; the DFlash 2 presets need a from-source build of an unmerged llama.cpp branch.
         </p>
 
         {llamaStatus?.running ? (
@@ -952,7 +971,7 @@ export function LocalLlmTab() {
                   <p><span className="text-gray-500">Base Model:</span> <code className="text-gray-300">{llamaStatus.config.model}</code></p>
                 )}
                 {llamaStatus.config?.draftModel && (
-                  <p><span className="text-gray-500">DFlash Drafter:</span> <code className="text-port-accent">{llamaStatus.config.draftModel}</code> ({llamaStatus.config.specType || 'draft-dflash'})</p>
+                  <p><span className="text-gray-500">Drafter:</span> <code className="text-port-accent">{llamaStatus.config.draftModel}</code> ({llamaStatus.config.specType || 'draft-dflash'})</p>
                 )}
               </div>
               {llamaStatus.managed ? (
@@ -981,10 +1000,10 @@ export function LocalLlmTab() {
                   id="llama-preset-select"
                   aria-label="Preset"
                   onChange={(e) => handlePresetSelect(e.target.value)}
-                  defaultValue="qwen3.8-27b"
+                  defaultValue="qwen3.8-27b-dspark"
                   className="bg-port-card border border-port-border rounded px-2 py-1 text-xs text-port-accent focus:outline-none"
                 >
-                  {DFLASH_PRESETS.map((p) => (
+                  {SPEC_DECODE_PRESETS.map((p) => (
                     <option key={p.id} value={p.id}>
                       {p.label}
                     </option>

docs/README.md

@@ -34,7 +34,7 @@ Start with the [product surface map](./features/product-surfaces.md) for a compl
 
 App management: [app-wizard](./features/app-wizard.md) · [autofixer](./features/autofixer.md) · [browser](./features/browser.md) · [error-handling](./features/error-handling.md) · [jira-sprint-manager](./features/jira-sprint-manager.md)
 
-Chief of Staff: [chief-of-staff](./features/chief-of-staff.md) · [cos-agent-runner](./features/cos-agent-runner.md) · [cos-enhancement](./features/cos-enhancement.md) · [agent-context](./features/agent-context.md) · [agent-skills](./features/agent-skills.md) · [memory-system](./features/memory-system.md) · [claude-ollama](./features/claude-ollama.md) · [mtplx](./features/mtplx.md) · [dflash2](./features/dflash2.md) · [prompt-manager](./features/prompt-manager.md)
+Chief of Staff: [chief-of-staff](./features/chief-of-staff.md) · [cos-agent-runner](./features/cos-agent-runner.md) · [cos-enhancement](./features/cos-enhancement.md) · [agent-context](./features/agent-context.md) · [agent-skills](./features/agent-skills.md) · [memory-system](./features/memory-system.md) · [claude-ollama](./features/claude-ollama.md) · [mtplx](./features/mtplx.md) · [dflash2](./features/dflash2.md) ([DSpark vs DFlash 2](./research/2026-08-19-dspark-vs-dflash2.md)) · [prompt-manager](./features/prompt-manager.md)
 
 Identity & self: [digital-twin](./features/digital-twin.md) · [identity-system](./features/identity-system.md) · [soul-system](./features/soul-system.md) · [privacy-center](./features/privacy-center.md) · [post](./features/post.md) (insights design spike: [plans/2026-06-03](./plans/2026-06-03-cross-domain-insights-engine.md))
 

docs/features/dflash2.md

@@ -1,16 +1,20 @@
-# DFlash 2 Speculative Decoding — llama.cpp & OpenCode llama TUI
+# Speculative Decoding (DSpark / DFlash 2) — llama.cpp & OpenCode llama TUI
 
 [DFlash 2](https://huggingface.co/z-lab) provides deep, ultra-fast block-level speculative drafting for large language models (such as Qwen 2.5, Qwen 3.8, and Muse-Glimmer). By pairing a small speculative drafter model (typically 1.5–3 GB) with a target foundation model (e.g. 27B–30B), DFlash 2 achieves 2.5–3× end-to-end token generation speedups without sacrificing output quality.
 
-PortOS provides direct integration with DFlash 2 through the **OpenCode llama TUI** provider preset and local OpenAI-compatible inference servers like `llama-server`.
+[DSpark](https://arxiv.org/html/2607.05147v1) (DeepSeek) is the sibling drafter family — same lossless block-drafting idea, a rank-256 Markov head instead of DFlash 2's candidate selector, and a much wider set of published drafters.
+
+PortOS integrates both through the **OpenCode llama TUI** provider preset and the managed local `llama-server`.
+
+> **Which one to run:** `--spec-type draft-dspark` **merged into llama.cpp on 2026-07-28** ([#25173](https://github.com/ggml-org/llama.cpp/pull/25173)) and works on a stock `brew install llama.cpp`. DFlash 2's engine modules are **still an open PR** ([#27342](https://github.com/ggml-org/llama.cpp/pull/27342)) and need a from-source build of that branch. llama.cpp falls back **silently** on a spec-type/drafter mismatch rather than erroring, so a DFlash 2 drafter on a stock build degrades quietly. Start with DSpark unless you have built the DFlash 2 branch. Full comparison: [DSpark vs DFlash 2](../research/2026-08-19-dspark-vs-dflash2.md).
 
 ---
 
 ## What PortOS Adds
 
 1. **OpenCode llama TUI Provider**:
    - An attachable `tui` coding-agent provider preset (`opencode-llama-tui`) configured to connect to `http://127.0.0.1:8080/v1`.
-   - Seeded with default model aliases `["dflash", "qwen3.8-27b-dflash2", "Muse-Glimmer-30B-DFlash2"]` with default `dflash`.
+   - Seeded with default model aliases `["dflash", "qwen3.8-27b-dflash2", "Muse-Glimmer-30B-DFlash2"]` with default `dflash`. The launcher keeps `--alias dflash` for every drafter family so this alias resolves regardless of which one you run.
    - Fully enabled by default and equipped with OpenCode's agentic file-writing harness, tool calling, and session persistence.
 2. **Model Refresh**:
    - Support for dynamic model discovery via the **Refresh Models** button on AI Providers, querying the local `llama-server` `/v1/models` endpoint.
@@ -22,7 +26,21 @@ PortOS provides direct integration with DFlash 2 through the **OpenCode llama TU
 ## Setup & Running with llama-server
 
 ### 1. Download Base & Draft Models
-Download your base GGUF and corresponding DFlash 2 drafter GGUF from Hugging Face:
+Download your base GGUF and a matching drafter GGUF from Hugging Face. Drafter
+checkpoints are **not standalone models** — they only produce text once an engine
+pairs them with their specific target, which is why PortOS's model search filters
+them out of the install picker.
+
+DSpark pairs (stock llama.cpp):
+
+- **Qwen 3.8 27B**: base `Qwen/Qwen3.8-27B-Instruct-GGUF` + drafter `DimInfer/Qwen3.8-27B-Dspark-v1`
+- **Qwen 3 8B**: base `Qwen/Qwen3-8B-Instruct-GGUF` + drafter converted from `deepseek-ai/dspark_qwen3_8b_block7`
+
+DSpark drafters ship without tokenizers — converting one to GGUF requires passing
+`--target-model-dir` so it reuses the target's tokenizer. Keep the drafter at bf16;
+the target can be any quant.
+
+DFlash 2 pairs (require a source build of llama.cpp [#27342](https://github.com/ggml-org/llama.cpp/pull/27342)):
 
 - **Qwen 3.8 27B Draft Pair**:
   - Base: `Qwen/Qwen3.8-27B-Instruct-GGUF` (e.g. `Qwen3.8-27B-Instruct-Q4_K_M.gguf`)
@@ -32,13 +50,14 @@ Download your base GGUF and corresponding DFlash 2 drafter GGUF from Hugging Fac
   - Drafter: `z-lab/Muse-Glimmer-30B-DFlash2-GGUF`
 
 ### 2. Launch llama-server
-Start `llama-server` on loopback port `8080` with speculative decoding enabled:
+Start `llama-server` on loopback port `8080` with speculative decoding enabled
+(`--spec-type draft-dflash` for a DFlash drafter, `draft-dspark` for a DSpark one):
 
 ```bash
 llama-server \
   -m models/Qwen3.8-27B-Instruct-Q4_K_M.gguf \
-  --draft-model models/Qwen3.8-27B-DFlash2-Q4_K_M.gguf \
-  --spec-type draft-dflash \
+  --draft-model models/Qwen3.8-27B-DSpark-bf16.gguf \
+  --spec-type draft-dspark \
   --port 8080 \
   --host 127.0.0.1 \
   --alias dflash \
@@ -51,3 +70,17 @@ llama-server \
 2. Verify **OpenCode llama TUI** is enabled.
 3. Click **Refresh Models** to pull the live aliases from `llama-server`, or use the default `dflash` model.
 4. Select **OpenCode llama TUI** in the CoS task creator or terminal runner to execute coding and agent tasks with speculative acceleration.
+
+---
+
+## Ollama
+
+Ollama's vendored llama.cpp engine has carried DSpark since **0.32.6**
+([#17545](https://github.com/ollama/ollama/pull/17545)), but Ollama surfaces no
+control for it — its server hardcodes `--spec-type draft-mtp` whenever
+speculative decoding is on. Selecting DSpark requires overriding
+`LLAMA_ARG_SPEC_TYPE` / `LLAMA_ARG_SPEC_DRAFT_MODEL` on the Ollama process and
+pointing at a raw blob-store path; the one reported measurement through that
+path was ~13%, far below what the drafter delivers natively. Tracked upstream in
+[ollama#17016](https://github.com/ollama/ollama/issues/17016). **Use
+`llama-server` for speculative decoding in PortOS, not the Ollama backend.**