Use verified FlashInfer profile cache
This commit is contained in:
@@ -12,6 +12,7 @@ OUTPUT_ROOT="${OUTPUT_ROOT:?OUTPUT_ROOT must be set}"
|
|||||||
RUN_DIR="$(pwd -P)"
|
RUN_DIR="$(pwd -P)"
|
||||||
PROFILE_DIR="${PROFILE_DIR:-${RUN_DIR%/runs/frontier-fidelity-envelope-v1}/runs/frontier-qwen30-vllm020-profile-v1}"
|
PROFILE_DIR="${PROFILE_DIR:-${RUN_DIR%/runs/frontier-fidelity-envelope-v1}/runs/frontier-qwen30-vllm020-profile-v1}"
|
||||||
VENV_ROOT="${VENV_ROOT:-/tmp/wjh/venvs/vllm-0.20.0-cu129-profiler-v1}"
|
VENV_ROOT="${VENV_ROOT:-/tmp/wjh/venvs/vllm-0.20.0-cu129-profiler-v1}"
|
||||||
|
PROFILE_HOME="${PROFILE_HOME:-/tmp/wjh}"
|
||||||
VLLM_SOURCE="${VLLM_SOURCE:-/home/admin/cpfs/wjh/agentic-kv/third_party/vllm_v20_build}"
|
VLLM_SOURCE="${VLLM_SOURCE:-/home/admin/cpfs/wjh/agentic-kv/third_party/vllm_v20_build}"
|
||||||
FRONTIER_SOURCE="${FRONTIER_SOURCE:-/home/admin/cpfs/wjh/aituner/frontier-t1-dash0-deadc4a}"
|
FRONTIER_SOURCE="${FRONTIER_SOURCE:-/home/admin/cpfs/wjh/aituner/frontier-t1-dash0-deadc4a}"
|
||||||
MODEL_ROOT="${MODEL_ROOT:-/home/admin/cpfs/wjh/models/Qwen/Qwen3-30B-A3B}"
|
MODEL_ROOT="${MODEL_ROOT:-/home/admin/cpfs/wjh/models/Qwen/Qwen3-30B-A3B}"
|
||||||
@@ -19,6 +20,9 @@ CAPTURE_BUCKETS="${CAPTURE_BUCKETS:-1 2 4 8 16 24 32 40 48 56 64 72 80 88 96 104
|
|||||||
KV_CONTEXTS="${KV_CONTEXTS:-128 1024 2048 4096 8192 16384 32768 40960}"
|
KV_CONTEXTS="${KV_CONTEXTS:-128 1024 2048 4096 8192 16384 32768 40960}"
|
||||||
WARMUP_ITERS="${WARMUP_ITERS:-3}"
|
WARMUP_ITERS="${WARMUP_ITERS:-3}"
|
||||||
REPEATS="${REPEATS:-5}"
|
REPEATS="${REPEATS:-5}"
|
||||||
|
export HOME="${PROFILE_HOME}"
|
||||||
|
export XDG_CACHE_HOME="${PROFILE_HOME}/.cache"
|
||||||
|
export VLLM_CACHE_ROOT="${XDG_CACHE_HOME}/vllm"
|
||||||
|
|
||||||
mkdir -p "${OUTPUT_ROOT}/logs" "${OUTPUT_ROOT}/provenance" "${OUTPUT_ROOT}/raw"
|
mkdir -p "${OUTPUT_ROOT}/logs" "${OUTPUT_ROOT}/provenance" "${OUTPUT_ROOT}/raw"
|
||||||
exec > >(tee -a "${OUTPUT_ROOT}/logs/attention-tp${TP}.log") 2>&1
|
exec > >(tee -a "${OUTPUT_ROOT}/logs/attention-tp${TP}.log") 2>&1
|
||||||
|
|||||||
@@ -6,12 +6,16 @@ OUTPUT_ROOT="${OUTPUT_ROOT:?OUTPUT_ROOT must be set}"
|
|||||||
RUN_DIR="$(pwd -P)"
|
RUN_DIR="$(pwd -P)"
|
||||||
PROFILE_DIR="${PROFILE_DIR:-${RUN_DIR%/runs/frontier-fidelity-envelope-v1}/runs/frontier-qwen30-vllm020-profile-v1}"
|
PROFILE_DIR="${PROFILE_DIR:-${RUN_DIR%/runs/frontier-fidelity-envelope-v1}/runs/frontier-qwen30-vllm020-profile-v1}"
|
||||||
VENV_ROOT="${VENV_ROOT:-/tmp/wjh/venvs/vllm-0.20.0-cu129-profiler-v1}"
|
VENV_ROOT="${VENV_ROOT:-/tmp/wjh/venvs/vllm-0.20.0-cu129-profiler-v1}"
|
||||||
|
PROFILE_HOME="${PROFILE_HOME:-/tmp/wjh}"
|
||||||
VLLM_SOURCE="${VLLM_SOURCE:-/home/admin/cpfs/wjh/agentic-kv/third_party/vllm_v20_build}"
|
VLLM_SOURCE="${VLLM_SOURCE:-/home/admin/cpfs/wjh/agentic-kv/third_party/vllm_v20_build}"
|
||||||
FRONTIER_SOURCE="${FRONTIER_SOURCE:-/home/admin/cpfs/wjh/aituner/frontier-t1-dash0-deadc4a}"
|
FRONTIER_SOURCE="${FRONTIER_SOURCE:-/home/admin/cpfs/wjh/aituner/frontier-t1-dash0-deadc4a}"
|
||||||
MODEL_ROOT="${MODEL_ROOT:-/home/admin/cpfs/wjh/models/Qwen/Qwen3-30B-A3B}"
|
MODEL_ROOT="${MODEL_ROOT:-/home/admin/cpfs/wjh/models/Qwen/Qwen3-30B-A3B}"
|
||||||
TOKENS_TEXT="${TOKENS:-1 2 4 8 16 24 32 40 48 56 64 72 80 88 96 104 112 120 128}"
|
TOKENS_TEXT="${TOKENS:-1 2 4 8 16 24 32 40 48 56 64 72 80 88 96 104 112 120 128}"
|
||||||
read -r -a TOKENS <<< "${TOKENS_TEXT}"
|
read -r -a TOKENS <<< "${TOKENS_TEXT}"
|
||||||
export MODEL_ROOT
|
export MODEL_ROOT
|
||||||
|
export HOME="${PROFILE_HOME}"
|
||||||
|
export XDG_CACHE_HOME="${PROFILE_HOME}/.cache"
|
||||||
|
export VLLM_CACHE_ROOT="${XDG_CACHE_HOME}/vllm"
|
||||||
|
|
||||||
mkdir -p "${OUTPUT_ROOT}/logs" "${OUTPUT_ROOT}/provenance" "${OUTPUT_ROOT}/profiles"
|
mkdir -p "${OUTPUT_ROOT}/logs" "${OUTPUT_ROOT}/provenance" "${OUTPUT_ROOT}/profiles"
|
||||||
exec > >(tee -a "${OUTPUT_ROOT}/logs/linear.log") 2>&1
|
exec > >(tee -a "${OUTPUT_ROOT}/logs/linear.log") 2>&1
|
||||||
|
|||||||
@@ -6,10 +6,14 @@ OUTPUT_ROOT="${OUTPUT_ROOT:?OUTPUT_ROOT must be set}"
|
|||||||
RUN_DIR="$(pwd -P)"
|
RUN_DIR="$(pwd -P)"
|
||||||
PROFILE_DIR="${PROFILE_DIR:-${RUN_DIR%/runs/frontier-fidelity-envelope-v1}/runs/frontier-qwen30-vllm020-profile-v1}"
|
PROFILE_DIR="${PROFILE_DIR:-${RUN_DIR%/runs/frontier-fidelity-envelope-v1}/runs/frontier-qwen30-vllm020-profile-v1}"
|
||||||
VENV_ROOT="${VENV_ROOT:-/tmp/wjh/venvs/vllm-0.20.0-cu129-profiler-v1}"
|
VENV_ROOT="${VENV_ROOT:-/tmp/wjh/venvs/vllm-0.20.0-cu129-profiler-v1}"
|
||||||
|
PROFILE_HOME="${PROFILE_HOME:-/tmp/wjh}"
|
||||||
VLLM_SOURCE="${VLLM_SOURCE:-/home/admin/cpfs/wjh/agentic-kv/third_party/vllm_v20_build}"
|
VLLM_SOURCE="${VLLM_SOURCE:-/home/admin/cpfs/wjh/agentic-kv/third_party/vllm_v20_build}"
|
||||||
FRONTIER_SOURCE="${FRONTIER_SOURCE:-/home/admin/cpfs/wjh/aituner/frontier-t1-dash0-deadc4a}"
|
FRONTIER_SOURCE="${FRONTIER_SOURCE:-/home/admin/cpfs/wjh/aituner/frontier-t1-dash0-deadc4a}"
|
||||||
MODEL_ROOT="${MODEL_ROOT:-/home/admin/cpfs/wjh/models/Qwen/Qwen3-30B-A3B}"
|
MODEL_ROOT="${MODEL_ROOT:-/home/admin/cpfs/wjh/models/Qwen/Qwen3-30B-A3B}"
|
||||||
TOKENS="${TOKENS:-1 2 4 8 16 24 32 40 48 56 64 72 80 88 96 104 112 120 128}"
|
TOKENS="${TOKENS:-1 2 4 8 16 24 32 40 48 56 64 72 80 88 96 104 112 120 128}"
|
||||||
|
export HOME="${PROFILE_HOME}"
|
||||||
|
export XDG_CACHE_HOME="${PROFILE_HOME}/.cache"
|
||||||
|
export VLLM_CACHE_ROOT="${XDG_CACHE_HOME}/vllm"
|
||||||
|
|
||||||
mkdir -p "${OUTPUT_ROOT}/logs" "${OUTPUT_ROOT}/provenance" "${OUTPUT_ROOT}/raw"
|
mkdir -p "${OUTPUT_ROOT}/logs" "${OUTPUT_ROOT}/provenance" "${OUTPUT_ROOT}/raw"
|
||||||
exec > >(tee -a "${OUTPUT_ROOT}/logs/moe-router.log") 2>&1
|
exec > >(tee -a "${OUTPUT_ROOT}/logs/moe-router.log") 2>&1
|
||||||
|
|||||||
Reference in New Issue
Block a user