Document validated default RTX PRO runtime and v0.3.0 installer

README.md

@@ -52,7 +52,46 @@ It makes one serving-oriented layout change:
 Sidecar SHA-256:
 `b070f9644adf93794d8a1030584ab705809387e64396a9327a68fa3a3a6666b3`
 
-## Performance
+## v0.3.0 runtime update
+
+The v0.3.0 extension packages 17 runtime modules, including two required
+compatibility interfaces, in a hardware-selected,
+private SGLang runtime view, leaving the base SGLang installation untouched.
+**This is a runtime update, not a model-weight change.** On the RTX PRO 6000
+Blackwell (96 GB), the default enables adaptive four/eight-token MTP, corrected
+sparse-index packing, ReplaySSM, shared scratch/capture-stream resources, and
+early draft-head initialization. Strict acceptance thresholds remain 1, SSM
+remains FP32, and KV remains FP8. Default context remains 131,072 tokens;
+DGX Spark and portable-profile arguments are unchanged.
+
+Clean Ubuntu 24.04 installation and full-model checks passed on the RTX PRO
+6000: SSD GPU smoke, five GPU packing tests, and 12 functional checks covering
+tools, images, reasoning, and retrieval at 120,081 and 255,681 prompt tokens.
+Both x86_64 and aarch64 wheels passed 65 CPU tests each; aarch64 hardware
+acceptance remains pending. The 131K run recorded 30 window switches in each
+direction and CUDA-graph replay for all 275 logged decode batches.
+
+### Measured v0.3.0 default profile
+
+Same prepared model and RTX PRO 6000, 131,072-token context, strict acceptance,
+and 1,024-token completions. The control uses `--no-rtx-optimizations`:
+
+| Workload | Fixed-four tok/s | Default tok/s | Median change |
+| --- | ---: | ---: | ---: |
+| List | 148.40 | 151.71 | +2.23% |
+| Prose | 166.52 | 155.89 | -6.39% |
+| Code | 208.78 | 217.59 | +4.22% |
+| Reasoning | 134.02 | 136.59 | +1.92% |
+
+**Results are mixed, not a universal speedup: prose regressed in this sample.**
+Each arm/workload has six measured requests, excluding warmups. Outputs varied;
+these small synthetic samples are not a general quality evaluation. The final
+control ran in a separate supervised pause after a connection interruption.
+[Full measurements, validation results, and artifact identity](https://github.com/garnermccloud/sglang-ssd-stream/blob/main/benchmarks/results/rtx-pro-6000-v0.3.0.json)
+and the [benchmark client](https://github.com/garnermccloud/sglang-ssd-stream/blob/main/benchmarks/validate_rtx.py)
+are public. The older measurements below describe a different runtime.
+
+## Historical performance
 
 ### September 2026: qualified RTX PRO adaptive-MTP results
 
@@ -84,17 +123,19 @@ remain sensitive to output and run order. One candidate code request measured
 241.15 tok/s. The historical runtime used the original RadixArk checkpoint and
 legacy SSD integration, not a fresh installation of this prepared model page.
 
-**The normal installer below does not enable this bundle.** Its pinned SGLang
-revision and plugin interface differ from the qualified historical runtime.
-Porting and clean-install acceptance are still pending; the published overlay
-must not be copied into arbitrary SGLang installations. No complete legacy base
-image is included. See the [versioned publication](https://github.com/garnermccloud/sglang-ssd-stream/releases/tag/rtx-pro-6000-20260909)
+The historical production runtime also used FRSpec with a 65,536-entry
+vocabulary, draft-head NVFP4, and `lowm`; these are not in the v0.3.0 installer.
+v0.3.0 enables the ported runtime by default on the RTX PRO 6000, but does
+not reproduce the entire historical setup. **282.12 tok/s and +18.16% are
+historical bundle results, not claims for the new default.** The published
+historical overlay must not be copied into arbitrary SGLang installations.
+No complete legacy base image is included. See the [versioned publication](https://github.com/garnermccloud/sglang-ssd-stream/releases/tag/rtx-pro-6000-20260909)
 for applicability, verification commands, and test limitations.
 
 These results are a separate experiment and should not be compared directly
 with the earlier SSD-versus-RAM figures below.
 
-### RTX PRO 6000 Blackwell
+### Earlier RTX PRO 6000 Blackwell SSD-versus-RAM results
 
 Matched tests on one RTX PRO 6000 Blackwell used native MTP, CUDA graphs, and a
 1,024-token completion:
@@ -112,8 +153,9 @@ The measured result is the point of SSD Stream: removing the table from RAM did
 not make it the decode bottleneck. Content-heavy requests dominated by unseen,
 random lookup rows measured 126-137 tok/s.
 
-The RTX acceptance suite also passed structured tool calls, unrelated images,
-a 120,043-token retrieval request, native MTP, CUDA graph replay, restart reuse,
+The earlier RTX acceptance suite (not v0.3.0 acceptance) also passed structured
+tool calls, unrelated images, a 120,043-token retrieval request, native MTP,
+CUDA graph replay, restart reuse,
 and a sustained alternating workload without swap, OOM, restart, or host-memory
 growth.
 
@@ -162,7 +204,9 @@ acceptance and performance measurements remain pending.
 
 ## Run it
 
-Use Linux with one supported NVIDIA GPU and a local SSD:
+Use Linux with one supported NVIDIA GPU, a local SSD, and a standard C++
+compiler. The native SSD reader is prebuilt; no Rust toolchain is required.
+Install or update with:
 
 ```bash
 curl -LsSf https://raw.githubusercontent.com/garnermccloud/sglang-ssd-stream/main/install.sh | sh
@@ -182,9 +226,38 @@ curl -s http://127.0.0.1:30000/v1/chat/completions \
   }'
 ```
 
-The CLI automatically selects the validated RTX PRO 6000 profile, experimental
-DGX Spark profile, or experimental grouped CPU-offload profile for a supported
-24-32 GB x86_64 GPU.
+The CLI automatically selects the validated RTX PRO 6000 profile,
+experimental DGX Spark profile, or experimental grouped
+CPU-offload profile for a supported 24-32 GB x86_64 GPU.
+
+To opt out of the RTX ports and use the old fixed-four runtime for rollback or
+comparison:
+
+```bash
+sglang-ssd-stream serve --no-rtx-optimizations
+```
+
+The RTX default context remains 131,072 tokens. An explicit
+`sglang-ssd-stream serve --context 262144` override passed full-model retrieval
+with a 255,681-token prompt.
+
+### Updates
+
+Rerun the installer above for software updates. The v0.3.0 installer is
+version-bound to the 0.3.0 wheel. It builds and checks a new virtual environment
+at its permanent path, then atomically promotes the launcher. Previous
+environments are preserved, and installation failures leave the old launcher
+in place. The C++ compiler prerequisite remains.
+
+Software and model updates are separate. v0.3.0 does not require a model
+update. To explicitly download and validate a newer prepared snapshot:
+
+```bash
+sglang-ssd-stream update-model
+```
+
+The existing snapshot stays active until validation succeeds; old model files
+are not deleted automatically.
 
 ## Model capabilities