Add Opt-In for GLM-5.3 Flash breakable prefill CUDA graphs (#38522)

docs/cookbook/autoregressive/GLM/GLM-5.3-Flash.mdx

@@ -27,7 +27,9 @@ Choose your hardware, then choose the operating point that matches your workload
 
 Every listed hardware platform exposes both strategies. A **Verified** badge means that exact hardware and command were tested. **Final Verification In Progress** means the recipe runs and is queued for measurement on the final weights. **Not Verified** means the command is a supported starting point that still needs workload validation. A choice is disabled only when the underlying runtime combination is known to be unsupported.
 
-The recommended selection is only a starting point. The same panel also lets you override the KV/DSA pairing, multimodal feature transport, and HiCache tiers. Changing an option that was not part of the measured command changes the badge to **Not Verified** without hiding the option.
+The recommended selection is only a starting point. The same panel also lets you override the KV/DSA pairing, multimodal feature transport, Breakable Cuda Graph, and HiCache tiers. Changing an option that was not part of the measured command changes the badge to **Not Verified** without hiding the option.
+
+**Breakable Cuda Graph** defaults to **Off**. Select **On** to add `--cuda-graph-backend-prefill breakable` to the generated command. This requires a build that includes [PR #38522](https://github.com/sgl-project/sglang/pull/38522).
 
 import { Deployment } from "/src/snippets/_deployment.jsx";
 import { config } from "/src/snippets/configs/zai-org/glm-5.3-flash.jsx";

docs/src/snippets/configs/zai-org/glm-5.3-flash.jsx

@@ -33,6 +33,7 @@ export const config = {
       s.kvDsaPair === pairing &&
       s.mmTransport === "auto" &&
       s.hicache === "off" &&
+      s.bcg === "off" &&
       s.dcp === "off"
     );
   },
@@ -68,6 +69,20 @@ export const config = {
         },
       ],
     },
+    {
+      id: "bcg",
+      title: "Breakable Cuda Graph",
+      default: "off",
+      options: [
+        { id: "off", label: "Off" },
+        {
+          id: "on",
+          label: "On",
+          flags: ["--cuda-graph-backend-prefill breakable"],
+          hints: ["Enables breakable prefill CUDA graphs; requires a build with PR #38522."],
+        },
+      ],
+    },
     {
       id: "mmTransport",
       title: "VLM Transport",
@@ -331,6 +346,7 @@ sgl-eval run gsm8k \\
       nnodes: 1,
       verified: true,
       verificationStatus: (s) =>
+        s.bcg !== "off" ? "unverified" :
         ["bf16-tilelang", "fp8-trtllm"].includes(s.kvDsaPair) &&
         s.mmTransport === "auto" &&
         s.hicache === "off" &&
@@ -362,6 +378,7 @@ sgl-eval run gsm8k \\
       nnodes: 1,
       verified: true,
       verificationStatus: (s) =>
+        s.bcg !== "off" ? "unverified" :
         ["bf16-tilelang", "fp8-trtllm"].includes(s.kvDsaPair) &&
         s.mmTransport === "auto" &&
         s.hicache === "off" &&
@@ -393,6 +410,7 @@ sgl-eval run gsm8k \\
       nnodes: 1,
       verified: true,
       verificationStatus: (s) =>
+        s.bcg !== "off" ? "unverified" :
         ["bf16-tilelang", "fp8-trtllm"].includes(s.kvDsaPair) &&
         s.mmTransport === "auto" &&
         s.hicache === "off" &&
@@ -426,6 +444,7 @@ sgl-eval run gsm8k \\
       nnodes: 1,
       verified: true,
       verificationStatus: (s) =>
+        s.bcg !== "off" ? "unverified" :
         ["bf16-tilelang", "fp8-trtllm"].includes(s.kvDsaPair) &&
         s.mmTransport === "auto" &&
         s.hicache === "off" &&
@@ -594,6 +613,7 @@ sgl-eval run gsm8k \\
       nnodes: 1,
       verified: true,
       verificationStatus: (s) =>
+        s.bcg !== "off" ? "unverified" :
         s.mmTransport === "auto" && s.hicache === "off"
           ? "verified"
           : "unverified",
@@ -623,6 +643,7 @@ sgl-eval run gsm8k \\
       nnodes: 1,
       verified: true,
       verificationStatus: (s) =>
+        s.bcg !== "off" ? "unverified" :
         ["off", "l2"].includes(s.hicache) ? "verified" : "unverified",
       env: [],
       flags: [
@@ -645,6 +666,7 @@ sgl-eval run gsm8k \\
       nnodes: 1,
       verified: true,
       verificationStatus: (s) =>
+        s.bcg !== "off" ? "unverified" :
         s.mmTransport === "auto" && s.hicache === "off"
           ? "verified"
           : "unverified",
@@ -674,6 +696,7 @@ sgl-eval run gsm8k \\
       nnodes: 1,
       verified: true,
       verificationStatus: (s) =>
+        s.bcg !== "off" ? "unverified" :
         ["off", "l2"].includes(s.hicache) ? "verified" : "unverified",
       env: [],
       flags: [
@@ -694,7 +717,8 @@ sgl-eval run gsm8k \\
       match: { hw: "b200", strategy: "low-latency", quant: "fp8" },
       nnodes: 1,
       verified: true,
-      verificationStatus: (s) => (s.hicache === "off" ? "verified" : "unverified"),
+      verificationStatus: (s) =>
+        s.bcg !== "off" ? "unverified" : (s.hicache === "off" ? "verified" : "unverified"),
       env: [],
       flags: [
         "--model-path {{MODEL_NAME}}",
@@ -720,6 +744,7 @@ sgl-eval run gsm8k \\
       nnodes: 1,
       verified: true,
       verificationStatus: (s) =>
+        s.bcg !== "off" ? "unverified" :
         ["off", "l2"].includes(s.hicache) ? "verified" : "unverified",
       env: [],
       flags: [
@@ -740,7 +765,8 @@ sgl-eval run gsm8k \\
       match: { hw: "b300", strategy: "low-latency", quant: "fp8" },
       nnodes: 1,
       verified: true,
-      verificationStatus: (s) => (s.hicache === "off" ? "verified" : "unverified"),
+      verificationStatus: (s) =>
+        s.bcg !== "off" ? "unverified" : (s.hicache === "off" ? "verified" : "unverified"),
       env: [],
       flags: [
         "--model-path {{MODEL_NAME}}",
@@ -766,6 +792,7 @@ sgl-eval run gsm8k \\
       nnodes: 1,
       verified: true,
       verificationStatus: (s) =>
+        s.bcg !== "off" ? "unverified" :
         ["off", "l2"].includes(s.hicache) ? "verified" : "unverified",
       env: [],
       flags: [