minke.yu/sglang
1
0
Fork 0
Code Issues Pull Requests Actions 12 Packages Projects Releases Wiki Activity

docs: add B200 NVFP4 recipes + benchmarks to GLM-5.2 cookbook (#29674)

Browse Source 
Co-authored-by: Claude Opus 4.8 <noreply@anthropic.com>
This commit is contained in:
zijiexia
2026-06-29 14:06:30 -07:00
committed by GitHub
co-authored by Claude Opus 4.8
parent a5c45a12bb
commit 74a197af9d
2 changed files with 97 additions and 4 deletions
Download Patch File Download Diff File Expand all files Collapse all files
docs_new/src/snippets/configs/zai-org/glm-5.2-benchmarks.jsx
+32
View File
@@ -102,4 +102,36 @@ export const benchmarks = [
{ match: { hw: "gb300", variant: "default", quant: "bf16", strategy: "low-latency", nodes: "multi-2" } },
{ match: { hw: "gb300", variant: "default", quant: "bf16", strategy: "balanced", nodes: "multi-2" } },
{ match: { hw: "gb300", variant: "default", quant: "bf16", strategy: "high-throughput", nodes: "multi-2" } },
// ---- B200 + NVFP4 ---- (8-GPU single node, TP8; nvidia/GLM-5.2-NVFP4 via --quantization modelopt_fp4,
// measured on the lmsysorg/sglang:dev-glm52-nvfp4 preview image, flush-cache every run.
// ttft_ms/tpot_ms are P50; tokens_per_sec_per_gpu = output tok/s/GPU.
// balanced & high-throughput add DP-Attention (dp8); low-latency uses MTP 5-1-6, balanced MTP 2-1-3.)
{
match: { hw: "b200", variant: "default", quant: "nvfp4", strategy: "low-latency", nodes: "single" },
sglang_version: "dev-glm52-nvfp4",
speed: [
{ workload: { dataset: "random", isl: 8192, osl: 1024, max_concurrency: 1 },
ttft_ms: 295, tpot_ms: 1.85, tokens_per_sec_per_gpu: 58.6 },
{ workload: { dataset: "random", isl: 8192, osl: 1024, max_concurrency: 16 },
ttft_ms: 2491, tpot_ms: 5.43, tokens_per_sec_per_gpu: 254.3 },
],
},
{
match: { hw: "b200", variant: "default", quant: "nvfp4", strategy: "balanced", nodes: "single" },
sglang_version: "dev-glm52-nvfp4",
speed: [
{ workload: { dataset: "random", isl: 8192, osl: 1024, max_concurrency: 64 },
ttft_ms: 5837, tpot_ms: 12.70, tokens_per_sec_per_gpu: 418.9 },
{ workload: { dataset: "random", isl: 8192, osl: 1024, max_concurrency: 256 },
ttft_ms: 16736, tpot_ms: 30.00, tokens_per_sec_per_gpu: 593.7 },
],
},
{
match: { hw: "b200", variant: "default", quant: "nvfp4", strategy: "high-throughput", nodes: "single" },
sglang_version: "dev-glm52-nvfp4",
speed: [
{ workload: { dataset: "random", isl: 8192, osl: 1024, max_concurrency: 1024 },
ttft_ms: 130174, tpot_ms: 67.12, tokens_per_sec_per_gpu: 589.4 },
],
},
];
docs_new/src/snippets/configs/zai-org/glm-5.2.jsx
+65 -4
View File
@@ -94,6 +94,7 @@ sgl-eval run aime25 \\
gb300: "lmsysorg/sglang:latest",
b300: "lmsysorg/sglang:latest",
// NVFP4 needs the dev image with modelopt_fp4 support (per-quant override).
"b200|nvfp4": "lmsysorg/sglang:dev-glm52-nvfp4",
"b300|nvfp4": "lmsysorg/sglang:dev-glm52-nvfp4",
"gb300|nvfp4": "lmsysorg/sglang:dev-glm52-nvfp4",
},
@@ -613,11 +614,71 @@ sgl-eval run aime25 \\
},
// ====================================================================
// NVFP4 (Blackwell Ultra) — nvidia/GLM-5.2-NVFP4 (Model Optimizer). TP4.
// B300: low-latency + balanced (the 4-GPU GB300 node fits the ~381 GB build).
// GB300: low-latency / balanced / high-throughput measured on a single 4xGB300
// node — balanced & high-throughput add DP-Attention (dp4); low-latency uses MTP 5-1-6.
// NVFP4 — nvidia/GLM-5.2-NVFP4 (Model Optimizer).
// B200: 8-GPU single node, TP8 (low-latency / balanced / high-throughput); balanced &
// high-throughput add DP-Attention (dp8). low-latency uses MTP 5-1-6, balanced MTP 2-1-3.
// B300/GB300: 4-GPU single node, TP4 (the node fits the ~381 GB build); GB300 adds dp4 on
// balanced & high-throughput. Blackwell NVFP4 measured on the dev-glm52-nvfp4 preview image.
// ====================================================================
{
match: { hw: "b200", variant: "default", quant: "nvfp4", strategy: "low-latency", nodes: "single" },
verified: true,
env: [],
flags: [
"--model-path {{MODEL_NAME}}",
"--tp 8",
"--quantization modelopt_fp4",
"--speculative-algorithm EAGLE",
"--speculative-num-steps 5",
"--speculative-eagle-topk 1",
"--speculative-num-draft-tokens 6",
"--chunked-prefill-size 8192",
"--mem-fraction-static 0.85",
"--host {{HOST_IP}}",
"--port {{PORT}}",
],
},
{
match: { hw: "b200", variant: "default", quant: "nvfp4", strategy: "balanced", nodes: "single" },
verified: true,
env: [],
flags: [
"--model-path {{MODEL_NAME}}",
"--tp 8",
"--quantization modelopt_fp4",
"--dp 8",
"--enable-dp-attention",
// Shorter draft (MTP 2-1-3) than low-latency's 5-1-6: at this concurrency the
// verify overhead of a long draft outweighs the accept-length gain.
"--speculative-algorithm EAGLE",
"--speculative-num-steps 2",
"--speculative-eagle-topk 1",
"--speculative-num-draft-tokens 3",
// Larger chunked-prefill (32768 → ~4096/rank under dp8) is the dominant balanced lever.
"--chunked-prefill-size 32768",
"--mem-fraction-static 0.92",
"--max-running-requests 256",
"--host {{HOST_IP}}",
"--port {{PORT}}",
],
},
{
match: { hw: "b200", variant: "default", quant: "nvfp4", strategy: "high-throughput", nodes: "single" },
verified: true,
env: [],
flags: [
"--model-path {{MODEL_NAME}}",
"--tp 8",
"--quantization modelopt_fp4",
"--dp 8",
"--enable-dp-attention",
"--chunked-prefill-size 32768",
"--mem-fraction-static 0.92",
"--max-running-requests 512",
"--host {{HOST_IP}}",
"--port {{PORT}}",
],
},
{
match: { hw: "b300", variant: "default", quant: "nvfp4", strategy: "low-latency", nodes: "single" },
verified: true,