While performance of the QoS block drops sharply if the dequeue size is
greater than or equal to the enqueue size, increasing the dequeue size
to just under the enqueue one gives improved performance when the
scheduler is not keeping up with the line rate.
Signed-off-by: Bruce Richardson <bruce.richard...@intel.com>
---
doc/guides/sample_app_ug/qos_scheduler.rst | 2 +-
examples/qos_sched/main.h | 2 +-
2 files changed, 2 insertions(+), 2 deletions(-)
diff --git a/doc/guides/sample_app_ug/qos_scheduler.rst
b/doc/guides/sample_app_ug/qos_scheduler.rst
index f376554dd9..9936b99172 100644
--- a/doc/guides/sample_app_ug/qos_scheduler.rst
+++ b/doc/guides/sample_app_ug/qos_scheduler.rst
@@ -91,7 +91,7 @@ Optional application parameters include:
* B = I/O RX lcore write burst size to the output software rings,
worker lcore read burst size from input software rings,QoS enqueue size
(the default value is 64)
-* C = QoS dequeue size (the default value is 32)
+* C = QoS dequeue size (the default value is 63)
* D = Worker lcore write burst size to the NIC TX (the default value is 64)
diff --git a/examples/qos_sched/main.h b/examples/qos_sched/main.h
index b9c301483a..d8f3e32c83 100644
--- a/examples/qos_sched/main.h
+++ b/examples/qos_sched/main.h
@@ -26,7 +26,7 @@ extern "C" {
#define MAX_PKT_RX_BURST 64
#define PKT_ENQUEUE 64
-#define PKT_DEQUEUE 32
+#define PKT_DEQUEUE 63
#define MAX_PKT_TX_BURST 64
#define RX_PTHRESH 8 /**< Default values of RX prefetch threshold reg. */
--
2.37.2
