* [PATCH] ixgbe: Limit lowest interrupt rate for adaptive interrupt moderation to 12K
@ 2015-07-30 22:19 Alexander Duyck
2015-09-02 1:49 ` [Intel-wired-lan] " Alexander Duyck
0 siblings, 1 reply; 5+ messages in thread
From: Alexander Duyck @ 2015-07-30 22:19 UTC (permalink / raw)
To: netdev, intel-wired-lan; +Cc: jeffrey.t.kirsher
This patch updates the lowest limit for adaptive interrupt interrupt
moderation to roughly 12K interrupts per second.
The way I came about reaching 12K as the desired interrupt rate is by
testing with UDP flows. Specifically I had a simple test that ran a
netperf UDP_STREAM test at varying sizes. What I found was as the packet
sizes increased the performance fell steadily behind until we were only
able to receive at ~4Gb/s with a message size of 65507. A bit of digging
found that we were dropping packets for the socket in the network stack,
and looking at things further what I found was I could solve it by increasing
the interrupt rate, or increasing the rmem_default/rmem_max. What I found was
that when the interrupt coalescing resulted in more data being processed
per interrupt than could be stored in the socket buffer we started losing
packets and the performance dropped. So I reached 12K based on the
following math.
rmem_default = 212992
skb->truesize = 2994
212992 / 2994 = 71.14 packets to fill the buffer
packet rate at 1514 packet size is 812744pps
71.14 / 812744 = 87.9us to fill socket buffer
>From there it was just a matter of choosing the interrupt rate and
providing a bit of wiggle room which is why I decided to go with 12K
interrupts per second as that uses a value of 84us.
The data below is based on VM to VM over a direct assigned ixgbe interface.
The test run was:
netperf -H <ip> -t UDP_STREAM"
Socket Message Elapsed Messages CPU Service
Size Size Time Okay Errors Throughput Util Demand
bytes bytes secs # # 10^6bits/sec % SS us/KB
Before:
212992 65507 60.00 1100662 0 9613.4 10.89 0.557
212992 60.00 473474 4135.4 11.27 0.576
After:
212992 65507 60.00 1100413 0 9611.2 10.73 0.549
212992 60.00 974132 8508.3 11.69 0.598
Using bare metal the data is similar but not as dramatic as the throughput
increases from about 8.5Gb/s to 9.5Gb/s.
Signed-off-by: Alexander Duyck <alexander.h.duyck@redhat.com>
---
drivers/net/ethernet/intel/ixgbe/ixgbe.h | 3 +--
drivers/net/ethernet/intel/ixgbe/ixgbe_ethtool.c | 2 +-
drivers/net/ethernet/intel/ixgbe/ixgbe_lib.c | 2 +-
drivers/net/ethernet/intel/ixgbe/ixgbe_main.c | 4 ++--
4 files changed, 5 insertions(+), 6 deletions(-)
diff --git a/drivers/net/ethernet/intel/ixgbe/ixgbe.h b/drivers/net/ethernet/intel/ixgbe/ixgbe.h
index ac3ac2a20386..93ebb30e6cbd 100644
--- a/drivers/net/ethernet/intel/ixgbe/ixgbe.h
+++ b/drivers/net/ethernet/intel/ixgbe/ixgbe.h
@@ -539,8 +539,7 @@ struct hwmon_buff {
#define IXGBE_MIN_RSC_ITR 24
#define IXGBE_100K_ITR 40
#define IXGBE_20K_ITR 200
-#define IXGBE_10K_ITR 400
-#define IXGBE_8K_ITR 500
+#define IXGBE_12K_ITR 336
/* ixgbe_test_staterr - tests bits in Rx descriptor status and error fields */
static inline __le32 ixgbe_test_staterr(union ixgbe_adv_rx_desc *rx_desc,
diff --git a/drivers/net/ethernet/intel/ixgbe/ixgbe_ethtool.c b/drivers/net/ethernet/intel/ixgbe/ixgbe_ethtool.c
index f7aeb560a504..6a929cf2ce7c 100644
--- a/drivers/net/ethernet/intel/ixgbe/ixgbe_ethtool.c
+++ b/drivers/net/ethernet/intel/ixgbe/ixgbe_ethtool.c
@@ -2279,7 +2279,7 @@ static int ixgbe_set_coalesce(struct net_device *netdev,
adapter->tx_itr_setting = ec->tx_coalesce_usecs;
if (adapter->tx_itr_setting == 1)
- tx_itr_param = IXGBE_10K_ITR;
+ tx_itr_param = IXGBE_12K_ITR;
else
tx_itr_param = adapter->tx_itr_setting;
diff --git a/drivers/net/ethernet/intel/ixgbe/ixgbe_lib.c b/drivers/net/ethernet/intel/ixgbe/ixgbe_lib.c
index 68e1e757ecef..f3168bcc7d87 100644
--- a/drivers/net/ethernet/intel/ixgbe/ixgbe_lib.c
+++ b/drivers/net/ethernet/intel/ixgbe/ixgbe_lib.c
@@ -866,7 +866,7 @@ static int ixgbe_alloc_q_vector(struct ixgbe_adapter *adapter,
if (txr_count && !rxr_count) {
/* tx only vector */
if (adapter->tx_itr_setting == 1)
- q_vector->itr = IXGBE_10K_ITR;
+ q_vector->itr = IXGBE_12K_ITR;
else
q_vector->itr = adapter->tx_itr_setting;
} else {
diff --git a/drivers/net/ethernet/intel/ixgbe/ixgbe_main.c b/drivers/net/ethernet/intel/ixgbe/ixgbe_main.c
index 3e6a9319c718..d35dedbb9560 100644
--- a/drivers/net/ethernet/intel/ixgbe/ixgbe_main.c
+++ b/drivers/net/ethernet/intel/ixgbe/ixgbe_main.c
@@ -2252,7 +2252,7 @@ static void ixgbe_update_itr(struct ixgbe_q_vector *q_vector,
/* simple throttlerate management
* 0-10MB/s lowest (100000 ints/s)
* 10-20MB/s low (20000 ints/s)
- * 20-1249MB/s bulk (8000 ints/s)
+ * 20-1249MB/s bulk (12000 ints/s)
*/
/* what was last interrupt timeslice? */
timepassed_us = q_vector->itr >> 2;
@@ -2341,7 +2341,7 @@ static void ixgbe_set_itr(struct ixgbe_q_vector *q_vector)
new_itr = IXGBE_20K_ITR;
break;
case bulk_latency:
- new_itr = IXGBE_8K_ITR;
+ new_itr = IXGBE_12K_ITR;
break;
default:
break;
^ permalink raw reply related [flat|nested] 5+ messages in thread
* Re: [Intel-wired-lan] [PATCH] ixgbe: Limit lowest interrupt rate for adaptive interrupt moderation to 12K
2015-07-30 22:19 [PATCH] ixgbe: Limit lowest interrupt rate for adaptive interrupt moderation to 12K Alexander Duyck
@ 2015-09-02 1:49 ` Alexander Duyck
2015-09-02 5:44 ` Hisashi T Fujinaka
2015-09-02 10:07 ` Jeff Kirsher
0 siblings, 2 replies; 5+ messages in thread
From: Alexander Duyck @ 2015-09-02 1:49 UTC (permalink / raw)
To: Alexander Duyck, netdev, intel-wired-lan
On 07/30/2015 03:19 PM, Alexander Duyck wrote:
> This patch updates the lowest limit for adaptive interrupt interrupt
> moderation to roughly 12K interrupts per second.
>
> The way I came about reaching 12K as the desired interrupt rate is by
> testing with UDP flows. Specifically I had a simple test that ran a
> netperf UDP_STREAM test at varying sizes. What I found was as the packet
> sizes increased the performance fell steadily behind until we were only
> able to receive at ~4Gb/s with a message size of 65507. A bit of digging
> found that we were dropping packets for the socket in the network stack,
> and looking at things further what I found was I could solve it by increasing
> the interrupt rate, or increasing the rmem_default/rmem_max. What I found was
> that when the interrupt coalescing resulted in more data being processed
> per interrupt than could be stored in the socket buffer we started losing
> packets and the performance dropped. So I reached 12K based on the
> following math.
>
> rmem_default = 212992
> skb->truesize = 2994
> 212992 / 2994 = 71.14 packets to fill the buffer
>
> packet rate at 1514 packet size is 812744pps
> 71.14 / 812744 = 87.9us to fill socket buffer
>
> >From there it was just a matter of choosing the interrupt rate and
> providing a bit of wiggle room which is why I decided to go with 12K
> interrupts per second as that uses a value of 84us.
>
> The data below is based on VM to VM over a direct assigned ixgbe interface.
> The test run was:
> netperf -H <ip> -t UDP_STREAM"
>
> Socket Message Elapsed Messages CPU Service
> Size Size Time Okay Errors Throughput Util Demand
> bytes bytes secs # # 10^6bits/sec % SS us/KB
> Before:
> 212992 65507 60.00 1100662 0 9613.4 10.89 0.557
> 212992 60.00 473474 4135.4 11.27 0.576
>
> After:
> 212992 65507 60.00 1100413 0 9611.2 10.73 0.549
> 212992 60.00 974132 8508.3 11.69 0.598
>
> Using bare metal the data is similar but not as dramatic as the throughput
> increases from about 8.5Gb/s to 9.5Gb/s.
>
> Signed-off-by: Alexander Duyck <alexander.h.duyck@redhat.com>
Has there been any update on this patch? I submitted it just over a
month ago now and it hasn't received any feedback. I was hoping this
could be submitted before the merge window closes for net-next.
Thanks.
- Alex
^ permalink raw reply [flat|nested] 5+ messages in thread
* Re: [Intel-wired-lan] [PATCH] ixgbe: Limit lowest interrupt rate for adaptive interrupt moderation to 12K
2015-09-02 1:49 ` [Intel-wired-lan] " Alexander Duyck
@ 2015-09-02 5:44 ` Hisashi T Fujinaka
2015-09-02 10:07 ` Jeff Kirsher
1 sibling, 0 replies; 5+ messages in thread
From: Hisashi T Fujinaka @ 2015-09-02 5:44 UTC (permalink / raw)
To: Alexander Duyck; +Cc: Alexander Duyck, netdev, intel-wired-lan
On Tue, 1 Sep 2015, Alexander Duyck wrote:
> On 07/30/2015 03:19 PM, Alexander Duyck wrote:
>> This patch updates the lowest limit for adaptive interrupt interrupt
>> moderation to roughly 12K interrupts per second.
...
> Has there been any update on this patch? I submitted it just over a month
> ago now and it hasn't received any feedback. I was hoping this could be
> submitted before the merge window closes for net-next.
>
> Thanks.
I'm nobody, but it makes sense to me.
--
Hisashi T Fujinaka - htodd@twofifty.com (also todd.fujinaka@intel.com)
BSEE + BSChem + BAEnglish + MSCS + $2.50 = coffee
^ permalink raw reply [flat|nested] 5+ messages in thread
* Re: [Intel-wired-lan] [PATCH] ixgbe: Limit lowest interrupt rate for adaptive interrupt moderation to 12K
2015-09-02 1:49 ` [Intel-wired-lan] " Alexander Duyck
2015-09-02 5:44 ` Hisashi T Fujinaka
@ 2015-09-02 10:07 ` Jeff Kirsher
2015-09-10 19:49 ` Singh, Krishneil K
1 sibling, 1 reply; 5+ messages in thread
From: Jeff Kirsher @ 2015-09-02 10:07 UTC (permalink / raw)
To: Alexander Duyck; +Cc: Alexander Duyck, netdev, intel-wired-lan
[-- Attachment #1: Type: text/plain, Size: 2780 bytes --]
On Tue, 2015-09-01 at 18:49 -0700, Alexander Duyck wrote:
> On 07/30/2015 03:19 PM, Alexander Duyck wrote:
> > This patch updates the lowest limit for adaptive interrupt interrupt
> > moderation to roughly 12K interrupts per second.
> >
> > The way I came about reaching 12K as the desired interrupt rate is
> by
> > testing with UDP flows. Specifically I had a simple test that ran a
> > netperf UDP_STREAM test at varying sizes. What I found was as the
> packet
> > sizes increased the performance fell steadily behind until we were
> only
> > able to receive at ~4Gb/s with a message size of 65507. A bit of
> digging
> > found that we were dropping packets for the socket in the network
> stack,
> > and looking at things further what I found was I could solve it by
> increasing
> > the interrupt rate, or increasing the rmem_default/rmem_max. What I
> found was
> > that when the interrupt coalescing resulted in more data being
> processed
> > per interrupt than could be stored in the socket buffer we started
> losing
> > packets and the performance dropped. So I reached 12K based on the
> > following math.
> >
> > rmem_default = 212992
> > skb->truesize = 2994
> > 212992 / 2994 = 71.14 packets to fill the buffer
> >
> > packet rate at 1514 packet size is 812744pps
> > 71.14 / 812744 = 87.9us to fill socket buffer
> >
> > >From there it was just a matter of choosing the interrupt rate and
> > providing a bit of wiggle room which is why I decided to go with 12K
> > interrupts per second as that uses a value of 84us.
> >
> > The data below is based on VM to VM over a direct assigned ixgbe
> interface.
> > The test run was:
> > netperf -H <ip> -t UDP_STREAM"
> >
> > Socket Message Elapsed Messages CPU
> Service
> > Size Size Time Okay Errors Throughput Util
> Demand
> > bytes bytes secs # # 10^6bits/sec % SS
> us/KB
> > Before:
> > 212992 65507 60.00 1100662 0 9613.4 10.89
> 0.557
> > 212992 60.00 473474 4135.4 11.27
> 0.576
> >
> > After:
> > 212992 65507 60.00 1100413 0 9611.2 10.73
> 0.549
> > 212992 60.00 974132 8508.3 11.69
> 0.598
> >
> > Using bare metal the data is similar but not as dramatic as the
> throughput
> > increases from about 8.5Gb/s to 9.5Gb/s.
> >
> > Signed-off-by: Alexander Duyck <alexander.h.duyck@redhat.com>
>
> Has there been any update on this patch? I submitted it just over a
> month ago now and it hasn't received any feedback. I was hoping this
> could be submitted before the merge window closes for net-next.
It will be in the next series I push later today.
[-- Attachment #2: This is a digitally signed message part --]
[-- Type: application/pgp-signature, Size: 819 bytes --]
^ permalink raw reply [flat|nested] 5+ messages in thread
* RE: [Intel-wired-lan] [PATCH] ixgbe: Limit lowest interrupt rate for adaptive interrupt moderation to 12K
2015-09-02 10:07 ` Jeff Kirsher
@ 2015-09-10 19:49 ` Singh, Krishneil K
0 siblings, 0 replies; 5+ messages in thread
From: Singh, Krishneil K @ 2015-09-10 19:49 UTC (permalink / raw)
To: Kirsher, Jeffrey T, Alexander Duyck; +Cc: netdev, intel-wired-lan
-----Original Message-----
From: Intel-wired-lan [mailto:intel-wired-lan-bounces@lists.osuosl.org] On Behalf Of Jeff Kirsher
Sent: Wednesday, September 2, 2015 3:07 AM
To: Alexander Duyck <alexander.duyck@gmail.com>
Cc: netdev@vger.kernel.org; intel-wired-lan@lists.osuosl.org
Subject: Re: [Intel-wired-lan] [PATCH] ixgbe: Limit lowest interrupt rate for adaptive interrupt moderation to 12K
On Tue, 2015-09-01 at 18:49 -0700, Alexander Duyck wrote:
> On 07/30/2015 03:19 PM, Alexander Duyck wrote:
> > This patch updates the lowest limit for adaptive interrupt interrupt
> > moderation to roughly 12K interrupts per second.
> >
> > The way I came about reaching 12K as the desired interrupt rate is
> by
> > testing with UDP flows. Specifically I had a simple test that ran a
> > netperf UDP_STREAM test at varying sizes. What I found was as the
> packet
> > sizes increased the performance fell steadily behind until we were
> only
> > able to receive at ~4Gb/s with a message size of 65507. A bit of
> digging
> > found that we were dropping packets for the socket in the network
> stack,
> > and looking at things further what I found was I could solve it by
> increasing
> > the interrupt rate, or increasing the rmem_default/rmem_max. What I
> found was
> > that when the interrupt coalescing resulted in more data being
> processed
> > per interrupt than could be stored in the socket buffer we started
> losing
> > packets and the performance dropped. So I reached 12K based on the
> > following math.
> >
> > rmem_default = 212992
> > skb->truesize = 2994
> > 212992 / 2994 = 71.14 packets to fill the buffer
> >
> > packet rate at 1514 packet size is 812744pps
> > 71.14 / 812744 = 87.9us to fill socket buffer
> >
> > >From there it was just a matter of choosing the interrupt rate and
> > providing a bit of wiggle room which is why I decided to go with 12K
> > interrupts per second as that uses a value of 84us.
> >
> > The data below is based on VM to VM over a direct assigned ixgbe
> interface.
> > The test run was:
> > netperf -H <ip> -t UDP_STREAM"
> >
> > Socket Message Elapsed Messages CPU
> Service
> > Size Size Time Okay Errors Throughput Util
> Demand
> > bytes bytes secs # # 10^6bits/sec % SS
> us/KB
> > Before:
> > 212992 65507 60.00 1100662 0 9613.4 10.89
> 0.557
> > 212992 60.00 473474 4135.4 11.27
> 0.576
> >
> > After:
> > 212992 65507 60.00 1100413 0 9611.2 10.73
> 0.549
> > 212992 60.00 974132 8508.3 11.69
> 0.598
> >
> > Using bare metal the data is similar but not as dramatic as the
> throughput
> > increases from about 8.5Gb/s to 9.5Gb/s.
> >
> > Signed-off-by: Alexander Duyck <alexander.h.duyck@redhat.com>
>
> Has there been any update on this patch? I submitted it just over a
> month ago now and it hasn't received any feedback. I was hoping this
> could be submitted before the merge window closes for net-next.
It will be in the next series I push later today.
Tested-by: Krishneil Singh <krishneil.k.singh@intel.com>
^ permalink raw reply [flat|nested] 5+ messages in thread
end of thread, other threads:[~2015-09-10 19:49 UTC | newest]
Thread overview: 5+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2015-07-30 22:19 [PATCH] ixgbe: Limit lowest interrupt rate for adaptive interrupt moderation to 12K Alexander Duyck
2015-09-02 1:49 ` [Intel-wired-lan] " Alexander Duyck
2015-09-02 5:44 ` Hisashi T Fujinaka
2015-09-02 10:07 ` Jeff Kirsher
2015-09-10 19:49 ` Singh, Krishneil K
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).