[Skiboot] [PATCH] SLW: Increase stop4-5 residency by 10x

Akshay Adiga akshay.adiga at linux.vnet.ibm.com
Wed Mar 21 14:27:36 AEDT 2018


Using DGEMM benchmark we observed there was a drop of 5-9% throughput with
and without stop4/5. In this benchmark the GPU waits on the cpu to wakeup
and provide the subsequent data block to compute. The wakup latency
accumulates over the run and shows up as a performance drop.

Linux enters stop4/5 more aggressively for its wakeup latency. Increasing
the residency from 1ms to 10ms makes the performance drop <1%

Signed-off-by: Akshay Adiga <akshay.adiga at linux.vnet.ibm.com>
---
 hw/slw.c | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)

diff --git a/hw/slw.c b/hw/slw.c
index db238ec..515582b 100644
--- a/hw/slw.c
+++ b/hw/slw.c
@@ -598,7 +598,7 @@ static struct cpu_idle_states power9_cpu_idle_states[] = {
 	{
 		.name = "stop4",
 		.latency_ns = 100000,
-		.residency_ns = 1000000,
+		.residency_ns = 10000000,
 		.flags = 0*OPAL_PM_DEC_STOP \
 		       | 0*OPAL_PM_TIMEBASE_STOP  \
 		       | 1*OPAL_PM_LOSE_USER_CONTEXT \
@@ -614,7 +614,7 @@ static struct cpu_idle_states power9_cpu_idle_states[] = {
 	{
 		.name = "stop5",
 		.latency_ns = 200000,
-		.residency_ns = 2000000,
+		.residency_ns = 20000000,
 		.flags = 0*OPAL_PM_DEC_STOP \
 		       | 0*OPAL_PM_TIMEBASE_STOP  \
 		       | 1*OPAL_PM_LOSE_USER_CONTEXT \
-- 
2.5.5



More information about the Skiboot mailing list