Update PERF_CLIENT_STABILIZE_WINDOW to 5000

triton-inference-server · Oct 14, 2024 · 776c4cb · 776c4cb
1 parent a280014
commit 776c4cb
Show file tree

Hide file tree

Showing 4 changed files with 4 additions and 4 deletions.
diff --git a/qa/L0_perf_nomodel/run_test.sh b/qa/L0_perf_nomodel/run_test.sh
@@ -36,7 +36,7 @@ CONCURRENCY=${CONCURRENCY:=1}
 TIMEOUT_PERIOD=600
 PERF_CLIENT_PROTOCOL=${PERF_CLIENT_PROTOCOL:=grpc}
 PERF_CLIENT_PERCENTILE=${PERF_CLIENT_PERCENTILE:=95}
-PERF_CLIENT_STABILIZE_WINDOW=${PERF_CLIENT_STABILIZE_WINDOW:=1000}
+PERF_CLIENT_STABILIZE_WINDOW=${PERF_CLIENT_STABILIZE_WINDOW:=5000}
 PERF_CLIENT_STABILIZE_THRESHOLD=${PERF_CLIENT_STABILIZE_THRESHOLD:=15}
 TENSOR_SIZE=${TENSOR_SIZE:=1}
 SHARED_MEMORY=${SHARED_MEMORY:="none"}

diff --git a/qa/L0_perf_nomodel/test.sh b/qa/L0_perf_nomodel/test.sh
@@ -55,7 +55,7 @@ PERF_CLIENT_SLOWDOWN_THRESHOLD=5.0
 
 # Length of window, in milliseconds, to use when stabilizing latency
 # and infer/sec results.
-PERF_CLIENT_STABILIZE_WINDOW=1000
+PERF_CLIENT_STABILIZE_WINDOW=5000
 
 # Threshold, as a percentage, to use when stabilizing latency and
 # infer/sec results. Values must vary by less than this percent over 3

diff --git a/qa/L0_perf_nomodel_new/run_test.sh b/qa/L0_perf_nomodel_new/run_test.sh
@@ -36,7 +36,7 @@ CONCURRENCY=${CONCURRENCY:=1}
 TIMEOUT_PERIOD=600
 PERF_CLIENT_PROTOCOL=${PERF_CLIENT_PROTOCOL:=grpc}
 PERF_CLIENT_PERCENTILE=${PERF_CLIENT_PERCENTILE:=95}
-PERF_CLIENT_STABILIZE_WINDOW=${PERF_CLIENT_STABILIZE_WINDOW:=1000}
+PERF_CLIENT_STABILIZE_WINDOW=${PERF_CLIENT_STABILIZE_WINDOW:=5000}
 PERF_CLIENT_STABILIZE_THRESHOLD=${PERF_CLIENT_STABILIZE_THRESHOLD:=15}
 TENSOR_SIZE=${TENSOR_SIZE:=1}
 SHARED_MEMORY=${SHARED_MEMORY:="none"}

diff --git a/qa/L0_perf_nomodel_new/test.sh b/qa/L0_perf_nomodel_new/test.sh
@@ -55,7 +55,7 @@ PERF_CLIENT_SLOWDOWN_THRESHOLD=5.0
 
 # Length of window, in milliseconds, to use when stabilizing latency
 # and infer/sec results.
-PERF_CLIENT_STABILIZE_WINDOW=1000
+PERF_CLIENT_STABILIZE_WINDOW=5000
 
 # Threshold, as a percentage, to use when stabilizing latency and
 # infer/sec results. Values must vary by less than this percent over 3