[Klaud Cold] Deprecate GLM-5 / GLM-5.1 benchmark configs / 弃用 GLM-5 / GLM-5.1 基准测试配置

benchmarks/multi_node/glm5_fp8_mi355x_sglang-disagg.sh → benchmarks/multi_node/deprecated/glm5_fp8_mi355x_sglang-disagg.sh

@@ -1,6 +1,6 @@
 #!/usr/bin/env bash
 
-source "$(dirname "$0")/../benchmark_lib.sh"
+source "$(dirname "$0")/../../benchmark_lib.sh"
 
 check_env_vars \
     CONC_LIST \

benchmarks/single_node/fixed_seq_len/glm5.1_fp4_mi355x.sh → benchmarks/single_node/fixed_seq_len/deprecated/glm5.1_fp4_mi355x.sh

@@ -1,7 +1,7 @@
 #!/usr/bin/env bash
 set -x
 
-source "$(dirname "$0")/../../benchmark_lib.sh"
+source "$(dirname "$0")/../../../benchmark_lib.sh"
 
 check_env_vars \
     MODEL \

benchmarks/single_node/fixed_seq_len/glm5.1_fp4_mi355x_atom.sh → benchmarks/single_node/fixed_seq_len/deprecated/glm5.1_fp4_mi355x_atom.sh

@@ -1,6 +1,6 @@
 #!/usr/bin/env bash
 
-source "$(dirname "$0")/../../benchmark_lib.sh"
+source "$(dirname "$0")/../../../benchmark_lib.sh"
 
 check_env_vars \
     MODEL \

benchmarks/single_node/fixed_seq_len/glm5_fp4_b200.sh → benchmarks/single_node/fixed_seq_len/deprecated/glm5_fp4_b200.sh

@@ -1,6 +1,6 @@
 #!/usr/bin/env bash
 
-source "$(dirname "$0")/../../benchmark_lib.sh"
+source "$(dirname "$0")/../../../benchmark_lib.sh"
 
 check_env_vars \
     MODEL \

benchmarks/single_node/fixed_seq_len/glm5_fp4_b200_mtp.sh → benchmarks/single_node/fixed_seq_len/deprecated/glm5_fp4_b200_mtp.sh

@@ -1,6 +1,6 @@
 #!/usr/bin/env bash
 
-source "$(dirname "$0")/../../benchmark_lib.sh"
+source "$(dirname "$0")/../../../benchmark_lib.sh"
 
 check_env_vars \
     MODEL \

benchmarks/single_node/fixed_seq_len/glm5_fp4_b300.sh → benchmarks/single_node/fixed_seq_len/deprecated/glm5_fp4_b300.sh

@@ -4,7 +4,7 @@
 # does not have a B300-specific recipe, so this script reuses the existing
 # GLM-5 FP4 B200 SGLang recipe as-is until B300-specific tuning is available.
 
-source "$(dirname "$0")/../../benchmark_lib.sh"
+source "$(dirname "$0")/../../../benchmark_lib.sh"
 
 check_env_vars \
     MODEL \

benchmarks/single_node/fixed_seq_len/glm5_fp4_b300_mtp.sh → benchmarks/single_node/fixed_seq_len/deprecated/glm5_fp4_b300_mtp.sh

@@ -4,7 +4,7 @@
 # does not have a B300-specific recipe, so this script reuses the existing
 # GLM5 FP8 B200 SGLang recipe as-is until B300-specific tuning is available.
 
-source "$(dirname "$0")/../../benchmark_lib.sh"
+source "$(dirname "$0")/../../../benchmark_lib.sh"
 
 check_env_vars \
     MODEL \

benchmarks/single_node/fixed_seq_len/glm5_fp8_b200.sh → benchmarks/single_node/fixed_seq_len/deprecated/glm5_fp8_b200.sh

@@ -1,6 +1,6 @@
 #!/usr/bin/env bash
 
-source "$(dirname "$0")/../../benchmark_lib.sh"
+source "$(dirname "$0")/../../../benchmark_lib.sh"
 
 check_env_vars \
     MODEL \

benchmarks/single_node/fixed_seq_len/glm5_fp8_b200_mtp.sh → benchmarks/single_node/fixed_seq_len/deprecated/glm5_fp8_b200_mtp.sh

@@ -1,6 +1,6 @@
 #!/usr/bin/env bash
 
-source "$(dirname "$0")/../../benchmark_lib.sh"
+source "$(dirname "$0")/../../../benchmark_lib.sh"
 
 check_env_vars \
     MODEL \

benchmarks/single_node/fixed_seq_len/glm5_fp8_b300.sh → benchmarks/single_node/fixed_seq_len/deprecated/glm5_fp8_b300.sh

@@ -4,7 +4,7 @@
 # does not have a B300-specific recipe, so this script reuses the existing
 # GLM5 FP8 B200 SGLang recipe as-is until B300-specific tuning is available.
 
-source "$(dirname "$0")/../../benchmark_lib.sh"
+source "$(dirname "$0")/../../../benchmark_lib.sh"
 
 check_env_vars \
     MODEL \

benchmarks/single_node/fixed_seq_len/glm5_fp8_b300_mtp.sh → benchmarks/single_node/fixed_seq_len/deprecated/glm5_fp8_b300_mtp.sh

@@ -4,7 +4,7 @@
 # does not have a B300-specific recipe, so this script reuses the existing
 # GLM5 FP8 B200 SGLang recipe as-is until B300-specific tuning is available.
 
-source "$(dirname "$0")/../../benchmark_lib.sh"
+source "$(dirname "$0")/../../../benchmark_lib.sh"
 
 check_env_vars \
     MODEL \

benchmarks/single_node/fixed_seq_len/glm5_fp8_h200.sh → benchmarks/single_node/fixed_seq_len/deprecated/glm5_fp8_h200.sh

@@ -1,6 +1,6 @@
 #!/usr/bin/env bash
 
-source "$(dirname "$0")/../../benchmark_lib.sh"
+source "$(dirname "$0")/../../../benchmark_lib.sh"
 
 check_env_vars \
     MODEL \

benchmarks/single_node/fixed_seq_len/glm5_fp8_h200_mtp.sh → benchmarks/single_node/fixed_seq_len/deprecated/glm5_fp8_h200_mtp.sh

@@ -6,7 +6,7 @@
 # nsa/trtllm-mha) since those backends are Blackwell-specific and not
 # applicable to Hopper.
 
-source "$(dirname "$0")/../../benchmark_lib.sh"
+source "$(dirname "$0")/../../../benchmark_lib.sh"
 
 check_env_vars \
     MODEL \

benchmarks/single_node/fixed_seq_len/glm5_fp8_mi325x.sh → benchmarks/single_node/fixed_seq_len/deprecated/glm5_fp8_mi325x.sh

@@ -1,6 +1,6 @@
 #!/usr/bin/env bash
 
-source "$(dirname "$0")/../../benchmark_lib.sh"
+source "$(dirname "$0")/../../../benchmark_lib.sh"
 
 check_env_vars \
     MODEL \

benchmarks/single_node/fixed_seq_len/glm5_fp8_mi325x_mtp.sh → benchmarks/single_node/fixed_seq_len/deprecated/glm5_fp8_mi325x_mtp.sh

@@ -3,7 +3,7 @@
 # GLM-5 FP8 on MI325X with EAGLE / MTP speculative decoding.
 # Mirrors glm5_fp8_mi325x.sh and adds the speculative-* flags.
 
-source "$(dirname "$0")/../../benchmark_lib.sh"
+source "$(dirname "$0")/../../../benchmark_lib.sh"
 
 check_env_vars \
     MODEL \

benchmarks/single_node/fixed_seq_len/glm5_fp8_mi355x.sh → benchmarks/single_node/fixed_seq_len/deprecated/glm5_fp8_mi355x.sh

@@ -1,6 +1,6 @@
 #!/usr/bin/env bash
 
-source "$(dirname "$0")/../../benchmark_lib.sh"
+source "$(dirname "$0")/../../../benchmark_lib.sh"
 
 check_env_vars \
     MODEL \

benchmarks/single_node/fixed_seq_len/glm5_fp8_mi355x_atom.sh → benchmarks/single_node/fixed_seq_len/deprecated/glm5_fp8_mi355x_atom.sh

@@ -1,6 +1,6 @@
 #!/usr/bin/env bash
 
-source "$(dirname "$0")/../../benchmark_lib.sh"
+source "$(dirname "$0")/../../../benchmark_lib.sh"
 
 check_env_vars \
     MODEL \

benchmarks/single_node/fixed_seq_len/glm5_fp8_mi355x_mtp.sh → benchmarks/single_node/fixed_seq_len/deprecated/glm5_fp8_mi355x_mtp.sh

@@ -1,7 +1,7 @@
 #!/usr/bin/env bash
 set -x
 
-source "$(dirname "$0")/../../benchmark_lib.sh"
+source "$(dirname "$0")/../../../benchmark_lib.sh"
 
 check_env_vars \
     MODEL \

configs/amd-master.yaml

@@ -462,119 +462,6 @@ qwen3.5-fp8-mi300x-sglang:
       search-space:
       - { tp: 8, conc-start: 4, conc-end: 64 }
 
-glm5-fp8-mi355x-sglang:
-  image: lmsysorg/sglang-rocm:v0.5.12-rocm720-mi35x-20260517
-  model: zai-org/GLM-5-FP8
-  model-prefix: glm5
-  runner: mi355x
-  precision: fp8
-  framework: sglang
-  multinode: false
-  scenarios:
-    fixed-seq-len:
-    - isl: 8192
-      osl: 1024
-      search-space:
-      - { tp: 4, conc-start: 4, conc-end: 256 }
-
-glm5-fp8-mi355x-sglang-mtp:
-  image: lmsysorg/sglang-rocm:v0.5.12-rocm720-mi35x-20260517
-  model: zai-org/GLM-5-FP8
-  model-prefix: glm5
-  runner: mi355x
-  precision: fp8
-  framework: sglang
-  multinode: false
-  scenarios:
-    fixed-seq-len:
-    - isl: 8192
-      osl: 1024
-      search-space:
-      - { tp: 4, conc-start: 4, conc-end: 128, spec-decoding: mtp }
-      - { tp: 8, conc-start: 4, conc-end: 8, spec-decoding: mtp }
-
-glm5-fp8-mi355x-sglang-disagg:
-  image: lmsysorg/sglang-rocm:v0.5.12.post1-rocm720-mi35x-20260523
-  model: zai-org/GLM-5-FP8
-  model-prefix: glm5
-  runner: mi355x-disagg
-  precision: fp8
-  framework: sglang-disagg
-  router: { name: sglang-router, version: "0.3.2" }
-  kv-p2p-transfer: mori
-  multinode: true
-  disagg: true
-  scenarios:
-    fixed-seq-len:
-    - isl: 8192
-      osl: 1024
-      search-space:
-      # 1P+1D TP8/EP1 CI smoke sweep; dp-attn false (NSA / MoRI path)
-      - spec-decoding: "none"
-        conc-list: [ 8, 16, 32, 64, 128, 256, 512 ]
-        prefill:
-          num-worker: 1
-          tp: 8
-          ep: 1
-          dp-attn: false
-          additional-settings:
-          - "PREFILL_NODES=1"
-        decode:
-          num-worker: 1
-          tp: 8
-          ep: 1
-          dp-attn: false
-          additional-settings:
-          - "DECODE_NODES=1"
-          - "DECODE_MTP_SIZE=0"
-
-glm5-fp8-mi355x-atom:
-  image: rocm/atom:rocm7.2.2_ubuntu24.04_py3.12_pytorch_release_2.10.0_atom0.1.2.post
-  model: zai-org/GLM-5-FP8
-  model-prefix: glm5
-  runner: mi355x
-  precision: fp8
-  framework: atom
-  multinode: false
-  scenarios:
-    fixed-seq-len:
-    - isl: 8192
-      osl: 1024
-      search-space:
-      - { tp: 4, conc-start: 4, conc-end: 256 }
-      - { tp: 8, conc-start: 4, conc-end: 256 }
-
-glm5.1-fp4-mi355x-sglang:
-  image: lmsysorg/sglang-rocm:v0.5.12.post1-rocm720-mi35x-20260529
-  model: amd/GLM-5.1-MXFP4
-  model-prefix: glm5.1
-  runner: mi355x
-  precision: fp4
-  framework: sglang
-  multinode: false
-  scenarios:
-    fixed-seq-len:
-    - isl: 8192
-      osl: 1024
-      search-space:
-      - { tp: 2, conc-start: 4, conc-end: 256 }
-      - { tp: 4, conc-start: 4, conc-end: 16 }
-
-glm5.1-fp4-mi355x-atom:
-  image: rocm/atom:rocm7.2.2_ubuntu24.04_py3.12_pytorch_release_2.10.0_atom0.1.2.post
-  model: amd/GLM-5.1-MXFP4
-  model-prefix: glm5.1
-  runner: mi355x
-  precision: fp4
-  framework: atom
-  multinode: false
-  scenarios:
-    fixed-seq-len:
-    - isl: 8192
-      osl: 1024
-      search-space:
-      - { tp: 4, conc-start: 4, conc-end: 256 }
-
 kimik2.5-int4-mi355x-vllm:
   image: vllm/vllm-openai-rocm:v0.24.0
   model: moonshotai/Kimi-K2.5
@@ -1476,43 +1363,6 @@ qwen3.5-fp8-mi325x-sglang-mtp:
       search-space:
       - { tp: 8, ep: 1, conc-start: 4, conc-end: 64, spec-decoding: mtp }
 
-glm5-fp8-mi325x-sglang:
-  image: lmsysorg/sglang:v0.5.12-rocm720-mi30x
-  model: zai-org/GLM-5-FP8
-  model-prefix: glm5
-  runner: mi325x
-  precision: fp8
-  framework: sglang
-  multinode: false
-  scenarios:
-    fixed-seq-len:
-    - isl: 8192
-      osl: 1024
-      search-space:
-      - { tp: 8, conc-start: 4, conc-end: 64 }
-
-glm5-fp8-mi325x-sglang-mtp:
-  image: lmsysorg/sglang:v0.5.12-rocm720-mi30x
-  model: zai-org/GLM-5-FP8
-  model-prefix: glm5
-  runner: mi325x
-  precision: fp8
-  framework: sglang
-  multinode: false
-  scenarios:
-    fixed-seq-len:
-    - isl: 8192
-      osl: 1024
-      search-space:
-      - { tp: 8, ep: 1, conc-start: 4, conc-end: 64, spec-decoding: mtp }
-
-# ============================================================================
-# Net-new agentic recipes from chore/agentx-v0.3 (no overlap with main entries).
-# Recipes that ALREADY existed on main were intentionally left at main's version
-# to preserve main behavior; PR-branch modifications to those recipes are NOT
-# brought in here.
-# ============================================================================
-
 qwen3.5-fp8-mi355x-sglang-agentic-hicache:
   image: lmsysorg/sglang-rocm:v0.5.12-rocm720-mi35x-20260521
   model: Qwen/Qwen3.5-397B-A17B-FP8