Skip to contentSkip to conversation

RFT: virtio_net: limit xmit polling

DISCUSSION Michael S. Tsirkin 3 participants 13 messages Jun 19, 2011 – Jul 17, 2011

Subsystems (5)
Networking Network drivers KVM s390 +1 more subsystems
lore
Share & tools
lore

Michael S. Tsirkin

author maintainer
#0 lore raw

OK, different people seem to test different trees. In the hope to get everyone on the same page, I created several variants of this patch so they can be compared. Whoever's interested, please check out the following, and tell me how these compare:

kernel:

git://git.kernel.org/pub/scm/linux/kernel/git/mst/vhost.git

virtio-net-limit-xmit-polling/base - this is net-next baseline to test against virtio-net-limit-xmit-polling/v0 - fixes checks on out of capacity virtio-net-limit-xmit-polling/v1 - previous revision of the patch
this does xmit,free,xmit,2*free,free
virtio-net-limit-xmit-polling/v2 - new revision of the patch
this does free,xmit,2*free,free

There's also this on top:
virtio-net-limit-xmit-polling/v3 -> don't delay avail index update I don't think it's important to test this one, yet

Userspace to use: event index work is not yet merged upstream so the revision to use is still this:
git://git.kernel.org/pub/scm/linux/kernel/git/mst/qemu-kvm.git
virtio-net-event-idx-v3

Signature · 1 line
MST
#1 lore raw
Quoted 4 lines · Michael S. Tsirkin

OK, different people seem to test different trees. In the hope to get everyone on the same page, I created several variants of this patch so they can be compared. Whoever's interested, please check out the following, and tell me how these compare:

I'm in the process of testing these patches. Base and v0 are complete and v1 is near complete with v2 to follow. I'm testing with a variety of TCP_RR and TCP_STREAM/TCP_MAERTS tests involving local guest-to-guest tests and remote host-to-guest tests. I'll post the results in the next day or two when the tests finish.

Thanks,
Tom

Quoted 20 lines

kernel:

git://git.kernel.org/pub/scm/linux/kernel/git/mst/vhost.git

virtio-net-limit-xmit-polling/base - this is net-next baseline to test against virtio-net-limit-xmit-polling/v0 - fixes checks on out of capacity virtio-net-limit-xmit-polling/v1 - previous revision of the patch
this does xmit,free,xmit,2*free,free
virtio-net-limit-xmit-polling/v2 - new revision of the patch
this does free,xmit,2*free,free

There's also this on top:
virtio-net-limit-xmit-polling/v3 -> don't delay avail index update I don't think it's important to test this one, yet

Userspace to use: event index work is not yet merged upstream so the revision to use is still this:
git://git.kernel.org/pub/scm/linux/kernel/git/mst/qemu-kvm.git
virtio-net-event-idx-v3

#4 lore raw
In reply to Michael S. Tsirkin
Quoted 4 lines · Michael S. Tsirkin

OK, different people seem to test different trees. In the hope to get everyone on the same page, I created several variants of this patch so they can be compared. Whoever's interested, please check out the following, and tell me how these compare:

I'm in the process of testing these patches. Base and v0 are complete and v1 is near complete with v2 to follow. I'm testing with a variety of TCP_RR and TCP_STREAM/TCP_MAERTS tests involving local guest-to-guest tests and remote host-to-guest tests. I'll post the results in the next day or two when the tests finish.

Thanks,
Tom

Quoted 20 lines

kernel:

git://git.kernel.org/pub/scm/linux/kernel/git/mst/vhost.git

virtio-net-limit-xmit-polling/base - this is net-next baseline to test against virtio-net-limit-xmit-polling/v0 - fixes checks on out of capacity virtio-net-limit-xmit-polling/v1 - previous revision of the patch
this does xmit,free,xmit,2*free,free
virtio-net-limit-xmit-polling/v2 - new revision of the patch
this does free,xmit,2*free,free

There's also this on top:
virtio-net-limit-xmit-polling/v3 -> don't delay avail index update I don't think it's important to test this one, yet

Userspace to use: event index work is not yet merged upstream so the revision to use is still this:
git://git.kernel.org/pub/scm/linux/kernel/git/mst/qemu-kvm.git
virtio-net-event-idx-v3

#2 lore raw
In reply to Tom Lendacky

Michael,

I am testing this too.
I have finished one round of testing. But am running it again just to
confirm.
This time I will see if I can collect some exit stats too. Will post results sometime this weekend.
I am just doing TCP_STREAM and TCP_MAERTS from guest to remote host.

Thanks,
Roopa

Quoted 6 lines · Tom Lendacky
Quoted 4 lines · Michael S. Tsirkin
Quoted 4 lines

OK, different people seem to test different trees. In the hope to get everyone on the same page, I created several variants of this patch so they can be compared. Whoever's interested, please check out the following, and tell me how these compare:

Quoted 30 lines

I'm in the process of testing these patches. Base and v0 are complete and v1 is near complete with v2 to follow. I'm testing with a variety of TCP_RR and TCP_STREAM/TCP_MAERTS tests involving local guest-to-guest tests and remote host-to-guest tests. I'll post the results in the next day or two when the tests finish.

Thanks,
Tom

Quoted 20 lines
Quoted 20 lines

kernel:

git://git.kernel.org/pub/scm/linux/kernel/git/mst/vhost.git

virtio-net-limit-xmit-polling/base - this is net-next baseline to test against virtio-net-limit-xmit-polling/v0 - fixes checks on out of capacity virtio-net-limit-xmit-polling/v1 - previous revision of the patch
this does xmit,free,xmit,2*free,free
virtio-net-limit-xmit-polling/v2 - new revision of the patch
this does free,xmit,2*free,free

There's also this on top:
virtio-net-limit-xmit-polling/v3 -> don't delay avail index update I don't think it's important to test this one, yet

Userspace to use: event index work is not yet merged upstream so the revision to use is still this:
git://git.kernel.org/pub/scm/linux/kernel/git/mst/qemu-kvm.git
virtio-net-event-idx-v3

_______________________________________________
Virtualization mailing list
[email protected]
https://lists.linux-foundation.org/mailman/listinfo/virtualization

#3 lore raw

Here are the results I am getting with a Cisco 10G VIC adapter. All tests are from the guest to an external host.

virtio-net-limit-xmit-polling/base:
TCP_STREAM: 8089Mbps
TCP_MAERTS: 9334Mbps

virtio-net-limit-xmit-polling/v0
TCP_STREAM: 8004Mbps
TCP_MAERTS: 9338Mbps

virtio-net-limit-xmit-polling/v1
TCP_STREAM: 8028Mbps
TCP_MAERTS: 9339Mbps

virtio-net-limit-xmit-polling/v2
TCP_STREAM: 8045Mbps
TCP_MAERTS: 9337Mbps

For the TCP_STREAM tests I don¹t get consistent results. Every run gives me slightly different results. But its always between 7900Mbps to 8100Mbps.
But I also see this with the base kernel so its not related to these patches.

Thanks,
Roopa

Quoted 53 lines · Roopa Prabhu

Michael,

I am testing this too.
I have finished one round of testing. But am running it again just to
confirm.
This time I will see if I can collect some exit stats too. Will post results sometime this weekend.
I am just doing TCP_STREAM and TCP_MAERTS from guest to remote host.

Thanks,
Roopa

Quoted 6 lines · Tom Lendacky
Quoted 4 lines · Michael S. Tsirkin
Quoted 4 lines

OK, different people seem to test different trees. In the hope to get everyone on the same page, I created several variants of this patch so they can be compared. Whoever's interested, please check out the following, and tell me how these compare:

Quoted 30 lines

I'm in the process of testing these patches. Base and v0 are complete and v1 is near complete with v2 to follow. I'm testing with a variety of TCP_RR and TCP_STREAM/TCP_MAERTS tests involving local guest-to-guest tests and remote host-to-guest tests. I'll post the results in the next day or two when the tests finish.

Thanks,
Tom

Quoted 20 lines
Quoted 20 lines

kernel:

git://git.kernel.org/pub/scm/linux/kernel/git/mst/vhost.git

virtio-net-limit-xmit-polling/base - this is net-next baseline to test against virtio-net-limit-xmit-polling/v0 - fixes checks on out of capacity virtio-net-limit-xmit-polling/v1 - previous revision of the patch
this does xmit,free,xmit,2*free,free
virtio-net-limit-xmit-polling/v2 - new revision of the patch
this does free,xmit,2*free,free

There's also this on top:
virtio-net-limit-xmit-polling/v3 -> don't delay avail index update I don't think it's important to test this one, yet

Userspace to use: event index work is not yet merged upstream so the revision to use is still this:
git://git.kernel.org/pub/scm/linux/kernel/git/mst/qemu-kvm.git
virtio-net-event-idx-v3

Quoted 4 lines

_______________________________________________
Virtualization mailing list
[email protected]
https://lists.linux-foundation.org/mailman/listinfo/virtualization

_______________________________________________
Virtualization mailing list
[email protected]
https://lists.linux-foundation.org/mailman/listinfo/virtualization

#5 lore raw
In reply to Michael S. Tsirkin
Quoted 7 lines · Michael S. Tsirkin

OK, different people seem to test different trees. In the hope to get everyone on the same page, I created several variants of this patch so they can be compared. Whoever's interested, please check out the following, and tell me how these compare:

kernel:

Quoted 9 lines

git://git.kernel.org/pub/scm/linux/kernel/git/mst/vhost.git

virtio-net-limit-xmit-polling/base - this is net-next baseline to test against virtio-net-limit-xmit-polling/v0 - fixes checks on out of capacity virtio-net-limit-xmit-polling/v1 - previous revision of the patch
this does xmit,free,xmit,2*free,free
virtio-net-limit-xmit-polling/v2 - new revision of the patch
this does free,xmit,2*free,free

Here's a summary of the results. I've also attached an ODS format spreadsheet (30 KB in size) that might be easier to analyze and also has some pinned VM results data. I broke the tests down into a local guest-to-guest scenario and a remote host-to-guest scenario.

Within the local guest-to-guest scenario I ran:
- TCP_RR tests using two different messsage sizes and four different instance counts among 1 pair of VMs and 2 pairs of VMs.
- TCP_STREAM tests using four different message sizes and two different instance counts among 1 pair of VMs and 2 pairs of VMs.

Within the remote host-to-guest scenario I ran:
- TCP_RR tests using two different messsage sizes and four different instance counts to 1 VM and 4 VMs.
- TCP_STREAM and TCP_MAERTS tests using four different message sizes and two different instance counts to 1 VM and 4 VMs.
over a 10GbE link.

*** Local Guest-to-Guest ***

Here's the local guest-to-guest summary for 1 VM pair doing TCP_RR with 256/256 request/response message size in transactions per second:

Instances Base V0 V1 V2
1 8,151.56 8,460.72 8,439.16 9,990.37
25 48,761.74 51,032.62 51,103.25 49,533.52
50 55,687.38 55,974.18 56,854.10 54,888.65
100 58,255.06 58,255.86 60,380.90 59,308.36

Here's the local guest-to-guest summary for 2 VM pairs doing TCP_RR with 256/256 request/response message size in transactions per second:

Instances Base V0 V1 V2
1 18,758.48 19,112.50 18,597.07 19,252.04
25 80,500.50 78,801.78 80,590.68 78,782.07
50 80,594.20 77,985.44 80,431.72 77,246.90
100 82,023.23 81,325.96 81,303.32 81,727.54

Here's the local guest-to-guest summary for 1 VM pair doing TCP_STREAM with 256, 1K, 4K and 16K message size in Mbps:

256:
Instances Base V0 V1 V2
1 961.78 1,115.92 794.02 740.37
4 2,498.33 2,541.82 2,441.60 2,308.26

1K:
1 3,476.61 3,522.02 2,170.86 1,395.57
4 6,344.30 7,056.57 7,275.16 7,174.09

4K:
1 9,213.57 10,647.44 9,883.42 9,007.29
4 11,070.66 11,300.37 11,001.02 12,103.72

16K:
1 12,065.94 9,437.78 11,710.60 6,989.93
4 12,755.28 13,050.78 12,518.06 13,227.33

Here's the local guest-to-guest summary for 2 VM pairs doing TCP_STREAM with 256, 1K, 4K and 16K message size in Mbps:

256:
Instances Base V0 V1 V2
1 2,434.98 2,403.23 2,308.69 2,261.35
4 5,973.82 5,729.48 5,956.76 5,831.86

1K:
1 5,305.99 5,148.72 4,960.67 5,067.76
4 10,628.38 10,649.49 10,098.90 10,380.09

4K:
1 11,577.03 10,710.33 11,700.53 10,304.09
4 14,580.66 14,881.38 14,551.17 15,053.02

16K:
1 16,801.46 16,072.50 15,773.78 15,835.66
4 17,194.00 17,294.02 17,319.78 17,121.09

*** Remote Host-to-Guest ***

Here's the remote host-to-guest summary for 1 VM doing TCP_RR with 256/256 request/response message size in transactions per second:

Instances Base V0 V1 V2
1 9,732.99 10,307.98 10,529.82 8,889.28
25 43,976.18 49,480.50 46,536.66 45,682.38
50 63,031.33 67,127.15 60,073.34 65,748.62
100 64,778.43 65,338.07 66,774.12 69,391.22

Here's the remote host-to-guest summary for 4 VMs doing TCP_RR with 256/256 request/response message size in transactions per second:

Instances Base V0 V1 V2
1 39,270.42 38,253.60 39,353.10 39,566.33
25 207,120.91 207,964.50 211,539.70 213,882.21
50 218,801.54 221,490.56 220,529.48 223,594.25
100 218,432.62 215,061.44 222,011.61 223,480.47

Here's the remote host-to-guest summary for 1 VM doing TCP_STREAM with 256, 1K, 4K and 16K message size in Mbps:

256:
Instances Base V0 V1 V2
1 2,274.74 2,220.38 2,245.26 2,212.30
4 5,689.66 5,953.86 5,984.80 5,827.94

1K:
1 7,804.38 7,236.29 6,716.58 7,485.09
4 7,722.42 8,070.38 7,700.45 7,856.76

4K:
1 8,976.14 9,026.77 9,147.32 9,095.58
4 7,532.25 7,410.80 7,683.81 7,524.94

16K:
1 8,991.61 9,045.10 9,124.58 9,238.34
4 7,406.10 7,626.81 7,711.62 7,345.37

Here's the remote host-to-guest summary for 1 VM doing TCP_MAERTS with 256, 1K, 4K and 16K message size in Mbps:

256:
Instances Base V0 V1 V2
1 1,165.69 1,181.92 1,152.20 1,104.68
4 2,580.46 2,545.22 2,436.30 2,601.74

1K:
1 2,393.34 2,457.22 2,128.86 2,258.92
4 7,152.57 7,606.60 8,004.64 7,576.85

4K:
1 9,258.93 8,505.06 9,309.78 9,215.05
4 9,374.20 9,363.48 9,372.53 9,352.00

16K:
1 9,244.70 9,287.72 9,298.60 9,322.28
4 9,380.02 9,347.50 9,377.46 9,372.98

Here's the remote host-to-guest summary for 4 VMs doing TCP_STREAM with 256, 1K, 4K and 16K message size in Mbps:

256:
Instances Base V0 V1 V2
1 9,392.37 9,390.74 9,395.58 9,392.46
4 9,394.24 9,394.46 9,395.42 9,394.05

1K:
1 9,396.34 9,397.46 9,396.64 9,443.26
4 9,397.14 9,402.25 9,398.67 9,391.09

4K:
1 9,397.16 9,398.07 9,397.30 9,396.33
4 9,395.64 9,400.25 9,397.54 9,397.75

16K:
1 9,396.58 9,397.01 9,397.58 9,397.70
4 9,399.15 9,400.02 9,399.66 9,400.16

Here's the remote host-to-guest summary for 4 VMs doing TCP_MAERTS with 256, 1K, 4K and 16K message size in Mbps:

256:
Instances Base V0 V1 V2
1 5,048.66 5,007.26 5,074.98 4,974.86
4 9,217.23 9,245.14 9,263.97 9,294.23

1K:
1 9,378.32 9,387.12 9,386.21 9,361.55
4 9,384.42 9,384.02 9,385.50 9,385.55

4K:
1 9,391.10 9,390.28 9,389.70 9,391.02
4 9,384.38 9,383.39 9,384.74 9,384.19

16K:
1 9,390.77 9,389.62 9,388.07 9,388.19
4 9,381.86 9,382.37 9,385.54 9,383.88

Tom

Quoted 8 lines

There's also this on top:
virtio-net-limit-xmit-polling/v3 -> don't delay avail index update I don't think it's important to test this one, yet

Userspace to use: event index work is not yet merged upstream so the revision to use is still this:
git://git.kernel.org/pub/scm/linux/kernel/git/mst/qemu-kvm.git
virtio-net-event-idx-v3

_______________________________________________
Virtualization mailing list
[email protected]
https://lists.linux-foundation.org/mailman/listinfo/virtualization

author maintainer
#6 lore raw
Quoted 29 lines · Tom Lendacky
Quoted 7 lines · Michael S. Tsirkin

OK, different people seem to test different trees. In the hope to get everyone on the same page, I created several variants of this patch so they can be compared. Whoever's interested, please check out the following, and tell me how these compare:

kernel:

Quoted 9 lines

git://git.kernel.org/pub/scm/linux/kernel/git/mst/vhost.git

virtio-net-limit-xmit-polling/base - this is net-next baseline to test against virtio-net-limit-xmit-polling/v0 - fixes checks on out of capacity virtio-net-limit-xmit-polling/v1 - previous revision of the patch
this does xmit,free,xmit,2*free,free
virtio-net-limit-xmit-polling/v2 - new revision of the patch
this does free,xmit,2*free,free

Here's a summary of the results. I've also attached an ODS format spreadsheet (30 KB in size) that might be easier to analyze and also has some pinned VM results data. I broke the tests down into a local guest-to-guest scenario and a remote host-to-guest scenario.

Within the local guest-to-guest scenario I ran:
- TCP_RR tests using two different messsage sizes and four different instance counts among 1 pair of VMs and 2 pairs of VMs.
- TCP_STREAM tests using four different message sizes and two different instance counts among 1 pair of VMs and 2 pairs of VMs.

Quoted 6 lines

Within the remote host-to-guest scenario I ran:
- TCP_RR tests using two different messsage sizes and four different instance counts to 1 VM and 4 VMs.
- TCP_STREAM and TCP_MAERTS tests using four different message sizes and two different instance counts to 1 VM and 4 VMs.
over a 10GbE link.

roprabhu, Tom,

Thanks very much for the testing. So on the first glance one seems to see a significant performance gain in V0 here, and a slightly less significant in V2, with V1 being worse than base. But I'm afraid that's not the whole story, and we'll need to work some more to know what really goes on, please see below.

Some comments on the results: I found out that V0 because of mistake on my part was actually almost identical to base.
I pushed out virtio-net-limit-xmit-polling/v1a instead that actually does what I intended to check. However,
the fact we get such a huge distribution in the results by Tom most likely means that the noise factor is very large.

Quoted 1 line

From my experience one way to get stable results is to

divide the throughput by the host CPU utilization (measured by something like mpstat).
Sometimes throughput doesn't increase (e.g. guest-host) by CPU utilization does decrease. So it's interesting.

Another issue is that we are trying to improve the latency of a busy queue here. However STREAM/MAERTS tests ignore the latency (more or less) while TCP_RR by default runs a single packet per queue. Without arguing about whether these are practically interesting workloads, these results are thus unlikely to be significantly affected by the optimization in question.

What we are interested in, thus, is either TCP_RR with a -b flag (configure with --enable-burst) or multiple concurrent TCP_RRs.

Quoted 173 lines

*** Local Guest-to-Guest ***

Here's the local guest-to-guest summary for 1 VM pair doing TCP_RR with 256/256 request/response message size in transactions per second:

Instances Base V0 V1 V2
1 8,151.56 8,460.72 8,439.16 9,990.37
25 48,761.74 51,032.62 51,103.25 49,533.52
50 55,687.38 55,974.18 56,854.10 54,888.65
100 58,255.06 58,255.86 60,380.90 59,308.36

Here's the local guest-to-guest summary for 2 VM pairs doing TCP_RR with 256/256 request/response message size in transactions per second:

Instances Base V0 V1 V2
1 18,758.48 19,112.50 18,597.07 19,252.04
25 80,500.50 78,801.78 80,590.68 78,782.07
50 80,594.20 77,985.44 80,431.72 77,246.90
100 82,023.23 81,325.96 81,303.32 81,727.54

Here's the local guest-to-guest summary for 1 VM pair doing TCP_STREAM with 256, 1K, 4K and 16K message size in Mbps:

256:
Instances Base V0 V1 V2
1 961.78 1,115.92 794.02 740.37
4 2,498.33 2,541.82 2,441.60 2,308.26

1K:
1 3,476.61 3,522.02 2,170.86 1,395.57
4 6,344.30 7,056.57 7,275.16 7,174.09

4K:
1 9,213.57 10,647.44 9,883.42 9,007.29
4 11,070.66 11,300.37 11,001.02 12,103.72

16K:
1 12,065.94 9,437.78 11,710.60 6,989.93
4 12,755.28 13,050.78 12,518.06 13,227.33

Here's the local guest-to-guest summary for 2 VM pairs doing TCP_STREAM with 256, 1K, 4K and 16K message size in Mbps:

256:
Instances Base V0 V1 V2
1 2,434.98 2,403.23 2,308.69 2,261.35
4 5,973.82 5,729.48 5,956.76 5,831.86

1K:
1 5,305.99 5,148.72 4,960.67 5,067.76
4 10,628.38 10,649.49 10,098.90 10,380.09

4K:
1 11,577.03 10,710.33 11,700.53 10,304.09
4 14,580.66 14,881.38 14,551.17 15,053.02

16K:
1 16,801.46 16,072.50 15,773.78 15,835.66
4 17,194.00 17,294.02 17,319.78 17,121.09

*** Remote Host-to-Guest ***

Here's the remote host-to-guest summary for 1 VM doing TCP_RR with 256/256 request/response message size in transactions per second:

Instances Base V0 V1 V2
1 9,732.99 10,307.98 10,529.82 8,889.28
25 43,976.18 49,480.50 46,536.66 45,682.38
50 63,031.33 67,127.15 60,073.34 65,748.62
100 64,778.43 65,338.07 66,774.12 69,391.22

Here's the remote host-to-guest summary for 4 VMs doing TCP_RR with 256/256 request/response message size in transactions per second:

Instances Base V0 V1 V2
1 39,270.42 38,253.60 39,353.10 39,566.33
25 207,120.91 207,964.50 211,539.70 213,882.21
50 218,801.54 221,490.56 220,529.48 223,594.25
100 218,432.62 215,061.44 222,011.61 223,480.47

Here's the remote host-to-guest summary for 1 VM doing TCP_STREAM with 256, 1K, 4K and 16K message size in Mbps:

256:
Instances Base V0 V1 V2
1 2,274.74 2,220.38 2,245.26 2,212.30
4 5,689.66 5,953.86 5,984.80 5,827.94

1K:
1 7,804.38 7,236.29 6,716.58 7,485.09
4 7,722.42 8,070.38 7,700.45 7,856.76

4K:
1 8,976.14 9,026.77 9,147.32 9,095.58
4 7,532.25 7,410.80 7,683.81 7,524.94

16K:
1 8,991.61 9,045.10 9,124.58 9,238.34
4 7,406.10 7,626.81 7,711.62 7,345.37

Here's the remote host-to-guest summary for 1 VM doing TCP_MAERTS with 256, 1K, 4K and 16K message size in Mbps:

256:
Instances Base V0 V1 V2
1 1,165.69 1,181.92 1,152.20 1,104.68
4 2,580.46 2,545.22 2,436.30 2,601.74

1K:
1 2,393.34 2,457.22 2,128.86 2,258.92
4 7,152.57 7,606.60 8,004.64 7,576.85

4K:
1 9,258.93 8,505.06 9,309.78 9,215.05
4 9,374.20 9,363.48 9,372.53 9,352.00

16K:
1 9,244.70 9,287.72 9,298.60 9,322.28
4 9,380.02 9,347.50 9,377.46 9,372.98

Here's the remote host-to-guest summary for 4 VMs doing TCP_STREAM with 256, 1K, 4K and 16K message size in Mbps:

256:
Instances Base V0 V1 V2
1 9,392.37 9,390.74 9,395.58 9,392.46
4 9,394.24 9,394.46 9,395.42 9,394.05

1K:
1 9,396.34 9,397.46 9,396.64 9,443.26
4 9,397.14 9,402.25 9,398.67 9,391.09

4K:
1 9,397.16 9,398.07 9,397.30 9,396.33
4 9,395.64 9,400.25 9,397.54 9,397.75

16K:
1 9,396.58 9,397.01 9,397.58 9,397.70
4 9,399.15 9,400.02 9,399.66 9,400.16

Here's the remote host-to-guest summary for 4 VMs doing TCP_MAERTS with 256, 1K, 4K and 16K message size in Mbps:

256:
Instances Base V0 V1 V2
1 5,048.66 5,007.26 5,074.98 4,974.86
4 9,217.23 9,245.14 9,263.97 9,294.23

1K:
1 9,378.32 9,387.12 9,386.21 9,361.55
4 9,384.42 9,384.02 9,385.50 9,385.55

4K:
1 9,391.10 9,390.28 9,389.70 9,391.02
4 9,384.38 9,383.39 9,384.74 9,384.19

16K:
1 9,390.77 9,389.62 9,388.07 9,388.19
4 9,381.86 9,382.37 9,385.54 9,383.88

Tom

Quoted 8 lines

There's also this on top:
virtio-net-limit-xmit-polling/v3 -> don't delay avail index update I don't think it's important to test this one, yet

Userspace to use: event index work is not yet merged upstream so the revision to use is still this:
git://git.kernel.org/pub/scm/linux/kernel/git/mst/qemu-kvm.git
virtio-net-event-idx-v3

author maintainer
#12 lore raw
In reply to Tom Lendacky
Quoted 29 lines · Tom Lendacky
Quoted 7 lines · Michael S. Tsirkin

OK, different people seem to test different trees. In the hope to get everyone on the same page, I created several variants of this patch so they can be compared. Whoever's interested, please check out the following, and tell me how these compare:

kernel:

Quoted 9 lines

git://git.kernel.org/pub/scm/linux/kernel/git/mst/vhost.git

virtio-net-limit-xmit-polling/base - this is net-next baseline to test against virtio-net-limit-xmit-polling/v0 - fixes checks on out of capacity virtio-net-limit-xmit-polling/v1 - previous revision of the patch
this does xmit,free,xmit,2*free,free
virtio-net-limit-xmit-polling/v2 - new revision of the patch
this does free,xmit,2*free,free

Here's a summary of the results. I've also attached an ODS format spreadsheet (30 KB in size) that might be easier to analyze and also has some pinned VM results data. I broke the tests down into a local guest-to-guest scenario and a remote host-to-guest scenario.

Within the local guest-to-guest scenario I ran:
- TCP_RR tests using two different messsage sizes and four different instance counts among 1 pair of VMs and 2 pairs of VMs.
- TCP_STREAM tests using four different message sizes and two different instance counts among 1 pair of VMs and 2 pairs of VMs.

Quoted 6 lines

Within the remote host-to-guest scenario I ran:
- TCP_RR tests using two different messsage sizes and four different instance counts to 1 VM and 4 VMs.
- TCP_STREAM and TCP_MAERTS tests using four different message sizes and two different instance counts to 1 VM and 4 VMs.
over a 10GbE link.

roprabhu, Tom,

Thanks very much for the testing. So on the first glance one seems to see a significant performance gain in V0 here, and a slightly less significant in V2, with V1 being worse than base. But I'm afraid that's not the whole story, and we'll need to work some more to know what really goes on, please see below.

Some comments on the results: I found out that V0 because of mistake on my part was actually almost identical to base.
I pushed out virtio-net-limit-xmit-polling/v1a instead that actually does what I intended to check. However,
the fact we get such a huge distribution in the results by Tom most likely means that the noise factor is very large.

From my experience one way to get stable results is to divide the throughput by the host CPU utilization (measured by something like mpstat).
Sometimes throughput doesn't increase (e.g. guest-host) by CPU utilization does decrease. So it's interesting.

Another issue is that we are trying to improve the latency of a busy queue here. However STREAM/MAERTS tests ignore the latency (more or less) while TCP_RR by default runs a single packet per queue. Without arguing about whether these are practically interesting workloads, these results are thus unlikely to be significantly affected by the optimization in question.

What we are interested in, thus, is either TCP_RR with a -b flag (configure with --enable-burst) or multiple concurrent TCP_RRs.

Quoted 173 lines

*** Local Guest-to-Guest ***

Here's the local guest-to-guest summary for 1 VM pair doing TCP_RR with 256/256 request/response message size in transactions per second:

Instances Base V0 V1 V2
1 8,151.56 8,460.72 8,439.16 9,990.37
25 48,761.74 51,032.62 51,103.25 49,533.52
50 55,687.38 55,974.18 56,854.10 54,888.65
100 58,255.06 58,255.86 60,380.90 59,308.36

Here's the local guest-to-guest summary for 2 VM pairs doing TCP_RR with 256/256 request/response message size in transactions per second:

Instances Base V0 V1 V2
1 18,758.48 19,112.50 18,597.07 19,252.04
25 80,500.50 78,801.78 80,590.68 78,782.07
50 80,594.20 77,985.44 80,431.72 77,246.90
100 82,023.23 81,325.96 81,303.32 81,727.54

Here's the local guest-to-guest summary for 1 VM pair doing TCP_STREAM with 256, 1K, 4K and 16K message size in Mbps:

256:
Instances Base V0 V1 V2
1 961.78 1,115.92 794.02 740.37
4 2,498.33 2,541.82 2,441.60 2,308.26

1K:
1 3,476.61 3,522.02 2,170.86 1,395.57
4 6,344.30 7,056.57 7,275.16 7,174.09

4K:
1 9,213.57 10,647.44 9,883.42 9,007.29
4 11,070.66 11,300.37 11,001.02 12,103.72

16K:
1 12,065.94 9,437.78 11,710.60 6,989.93
4 12,755.28 13,050.78 12,518.06 13,227.33

Here's the local guest-to-guest summary for 2 VM pairs doing TCP_STREAM with 256, 1K, 4K and 16K message size in Mbps:

256:
Instances Base V0 V1 V2
1 2,434.98 2,403.23 2,308.69 2,261.35
4 5,973.82 5,729.48 5,956.76 5,831.86

1K:
1 5,305.99 5,148.72 4,960.67 5,067.76
4 10,628.38 10,649.49 10,098.90 10,380.09

4K:
1 11,577.03 10,710.33 11,700.53 10,304.09
4 14,580.66 14,881.38 14,551.17 15,053.02

16K:
1 16,801.46 16,072.50 15,773.78 15,835.66
4 17,194.00 17,294.02 17,319.78 17,121.09

*** Remote Host-to-Guest ***

Here's the remote host-to-guest summary for 1 VM doing TCP_RR with 256/256 request/response message size in transactions per second:

Instances Base V0 V1 V2
1 9,732.99 10,307.98 10,529.82 8,889.28
25 43,976.18 49,480.50 46,536.66 45,682.38
50 63,031.33 67,127.15 60,073.34 65,748.62
100 64,778.43 65,338.07 66,774.12 69,391.22

Here's the remote host-to-guest summary for 4 VMs doing TCP_RR with 256/256 request/response message size in transactions per second:

Instances Base V0 V1 V2
1 39,270.42 38,253.60 39,353.10 39,566.33
25 207,120.91 207,964.50 211,539.70 213,882.21
50 218,801.54 221,490.56 220,529.48 223,594.25
100 218,432.62 215,061.44 222,011.61 223,480.47

Here's the remote host-to-guest summary for 1 VM doing TCP_STREAM with 256, 1K, 4K and 16K message size in Mbps:

256:
Instances Base V0 V1 V2
1 2,274.74 2,220.38 2,245.26 2,212.30
4 5,689.66 5,953.86 5,984.80 5,827.94

1K:
1 7,804.38 7,236.29 6,716.58 7,485.09
4 7,722.42 8,070.38 7,700.45 7,856.76

4K:
1 8,976.14 9,026.77 9,147.32 9,095.58
4 7,532.25 7,410.80 7,683.81 7,524.94

16K:
1 8,991.61 9,045.10 9,124.58 9,238.34
4 7,406.10 7,626.81 7,711.62 7,345.37

Here's the remote host-to-guest summary for 1 VM doing TCP_MAERTS with 256, 1K, 4K and 16K message size in Mbps:

256:
Instances Base V0 V1 V2
1 1,165.69 1,181.92 1,152.20 1,104.68
4 2,580.46 2,545.22 2,436.30 2,601.74

1K:
1 2,393.34 2,457.22 2,128.86 2,258.92
4 7,152.57 7,606.60 8,004.64 7,576.85

4K:
1 9,258.93 8,505.06 9,309.78 9,215.05
4 9,374.20 9,363.48 9,372.53 9,352.00

16K:
1 9,244.70 9,287.72 9,298.60 9,322.28
4 9,380.02 9,347.50 9,377.46 9,372.98

Here's the remote host-to-guest summary for 4 VMs doing TCP_STREAM with 256, 1K, 4K and 16K message size in Mbps:

256:
Instances Base V0 V1 V2
1 9,392.37 9,390.74 9,395.58 9,392.46
4 9,394.24 9,394.46 9,395.42 9,394.05

1K:
1 9,396.34 9,397.46 9,396.64 9,443.26
4 9,397.14 9,402.25 9,398.67 9,391.09

4K:
1 9,397.16 9,398.07 9,397.30 9,396.33
4 9,395.64 9,400.25 9,397.54 9,397.75

16K:
1 9,396.58 9,397.01 9,397.58 9,397.70
4 9,399.15 9,400.02 9,399.66 9,400.16

Here's the remote host-to-guest summary for 4 VMs doing TCP_MAERTS with 256, 1K, 4K and 16K message size in Mbps:

256:
Instances Base V0 V1 V2
1 5,048.66 5,007.26 5,074.98 4,974.86
4 9,217.23 9,245.14 9,263.97 9,294.23

1K:
1 9,378.32 9,387.12 9,386.21 9,361.55
4 9,384.42 9,384.02 9,385.50 9,385.55

4K:
1 9,391.10 9,390.28 9,389.70 9,391.02
4 9,384.38 9,383.39 9,384.74 9,384.19

16K:
1 9,390.77 9,389.62 9,388.07 9,388.19
4 9,381.86 9,382.37 9,385.54 9,383.88

Tom

Quoted 8 lines

There's also this on top:
virtio-net-limit-xmit-polling/v3 -> don't delay avail index update I don't think it's important to test this one, yet

Userspace to use: event index work is not yet merged upstream so the revision to use is still this:
git://git.kernel.org/pub/scm/linux/kernel/git/mst/qemu-kvm.git
virtio-net-event-idx-v3

#7 lore raw
In reply to Michael S. Tsirkin
Quoted 33 lines · Michael S. Tsirkin
Quoted 33 lines
Quoted 33 lines

roprabhu, Tom,

Thanks very much for the testing. So on the first glance one seems to see a significant performance gain in V0 here, and a slightly less significant in V2, with V1 being worse than base. But I'm afraid that's not the whole story, and we'll need to work some more to know what really goes on, please see below.

Some comments on the results: I found out that V0 because of mistake on my part was actually almost identical to base.
I pushed out virtio-net-limit-xmit-polling/v1a instead that actually does what I intended to check. However,
the fact we get such a huge distribution in the results by Tom most likely means that the noise factor is very large.

From my experience one way to get stable results is to divide the throughput by the host CPU utilization (measured by something like mpstat).
Sometimes throughput doesn't increase (e.g. guest-host) by CPU utilization does decrease. So it's interesting.

Another issue is that we are trying to improve the latency of a busy queue here. However STREAM/MAERTS tests ignore the latency (more or less) while TCP_RR by default runs a single packet per queue. Without arguing about whether these are practically interesting workloads, these results are thus unlikely to be significantly affected by the optimization in question.

Quoted 7 lines
Quoted 3 lines
Quoted 3 lines

What we are interested in, thus, is either TCP_RR with a -b flag (configure with --enable-burst) or multiple concurrent TCP_RRs.

ok sounds good. I am testing your v1a patch. Will try to get some results out end of this week. Thanks.

_______________________________________________
Virtualization mailing list
[email protected]
https://lists.linux-foundation.org/mailman/listinfo/virtualization

#8 lore raw
In reply to Michael S. Tsirkin
Quoted 70 lines · Michael S. Tsirkin
Quoted 30 lines · Tom Lendacky
Quoted 7 lines · Michael S. Tsirkin

OK, different people seem to test different trees. In the hope to get everyone on the same page, I created several variants of this patch so they can be compared. Whoever's interested, please check out the following, and tell me how these compare:

kernel:

Quoted 9 lines

git://git.kernel.org/pub/scm/linux/kernel/git/mst/vhost.git

virtio-net-limit-xmit-polling/base - this is net-next baseline to test against virtio-net-limit-xmit-polling/v0 - fixes checks on out of capacity virtio-net-limit-xmit-polling/v1 - previous revision of the patch
this does xmit,free,xmit,2*free,free
virtio-net-limit-xmit-polling/v2 - new revision of the patch
this does free,xmit,2*free,free

Here's a summary of the results. I've also attached an ODS format spreadsheet
(30 KB in size) that might be easier to analyze and also has some pinned VM results data. I broke the tests down into a local guest-to-guest scenario and a remote host-to-guest scenario.

Within the local guest-to-guest scenario I ran:
- TCP_RR tests using two different messsage sizes and four different instance counts among 1 pair of VMs and 2 pairs of VMs.
- TCP_STREAM tests using four different message sizes and two different instance counts among 1 pair of VMs and 2 pairs of VMs.

Quoted 6 lines

Within the remote host-to-guest scenario I ran:
- TCP_RR tests using two different messsage sizes and four different instance counts to 1 VM and 4 VMs.
- TCP_STREAM and TCP_MAERTS tests using four different message sizes and two different instance counts to 1 VM and 4 VMs.
over a 10GbE link.

roprabhu, Tom,

Thanks very much for the testing. So on the first glance one seems to see a significant performance gain in V0 here, and a slightly less significant in V2, with V1 being worse than base. But I'm afraid that's not the whole story, and we'll need to work some more to know what really goes on, please see below.

Some comments on the results: I found out that V0 because of mistake on my part was actually almost identical to base.
I pushed out virtio-net-limit-xmit-polling/v1a instead that actually does what I intended to check. However,
the fact we get such a huge distribution in the results by Tom most likely means that the noise factor is very large.

From my experience one way to get stable results is to divide the throughput by the host CPU utilization (measured by something like mpstat).
Sometimes throughput doesn't increase (e.g. guest-host) by CPU utilization does decrease. So it's interesting.

Another issue is that we are trying to improve the latency of a busy queue here. However STREAM/MAERTS tests ignore the latency (more or less) while TCP_RR by default runs a single packet per queue. Without arguing about whether these are practically interesting workloads, these results are thus unlikely to be significantly affected by the optimization in question.

Quoted 6 lines

What we are interested in, thus, is either TCP_RR with a -b flag (configure with --enable-burst) or multiple concurrent TCP_RRs.

Michael, below are some numbers I got from one round of runs. Thanks,
Roopa

256byte req/response.
Vcpus and irqs were pinned to 4 cores and the cpu utilization is Avg across 4 cores.

base:
Numof concurrent TCP_RRs Num of transactions/sec host cpu-util(%)

1                            7982.93                        15.72
25                           67873                         28.84
50                           112534                        52.25
100                          192057                       86.54

v1

Numof concurrent TCP_RRs    Num of transactions/sec    host cpu-util(%)
1                           7970.94                       10.8
25                          65496.8                       28
50                          109858                        53.22
100                         190155                        87.5

v1a

Numof concurrent TCP_RRs    Num of transactions/sec   host cpu-util (%)
1                           7979.81                       9.5
25                          66786.1                       28
50                          109552                        51
100                         190876                        88

v2
Numof concurrent TCP_RRs Num of transactions/sec host cpu-util (%)

1                            7969.87                     16.5
25                           67780.1                     28.44
50                           114966                      54.29
100                          177982                      79.9
#11 lore raw
In reply to Michael S. Tsirkin
Quoted 70 lines · Michael S. Tsirkin
Quoted 30 lines · Tom Lendacky
Quoted 7 lines · Michael S. Tsirkin

OK, different people seem to test different trees. In the hope to get everyone on the same page, I created several variants of this patch so they can be compared. Whoever's interested, please check out the following, and tell me how these compare:

kernel:

Quoted 9 lines

git://git.kernel.org/pub/scm/linux/kernel/git/mst/vhost.git

virtio-net-limit-xmit-polling/base - this is net-next baseline to test against virtio-net-limit-xmit-polling/v0 - fixes checks on out of capacity virtio-net-limit-xmit-polling/v1 - previous revision of the patch
this does xmit,free,xmit,2*free,free
virtio-net-limit-xmit-polling/v2 - new revision of the patch
this does free,xmit,2*free,free

Here's a summary of the results. I've also attached an ODS format spreadsheet
(30 KB in size) that might be easier to analyze and also has some pinned VM results data. I broke the tests down into a local guest-to-guest scenario and a remote host-to-guest scenario.

Within the local guest-to-guest scenario I ran:
- TCP_RR tests using two different messsage sizes and four different instance counts among 1 pair of VMs and 2 pairs of VMs.
- TCP_STREAM tests using four different message sizes and two different instance counts among 1 pair of VMs and 2 pairs of VMs.

Quoted 6 lines

Within the remote host-to-guest scenario I ran:
- TCP_RR tests using two different messsage sizes and four different instance counts to 1 VM and 4 VMs.
- TCP_STREAM and TCP_MAERTS tests using four different message sizes and two different instance counts to 1 VM and 4 VMs.
over a 10GbE link.

roprabhu, Tom,

Thanks very much for the testing. So on the first glance one seems to see a significant performance gain in V0 here, and a slightly less significant in V2, with V1 being worse than base. But I'm afraid that's not the whole story, and we'll need to work some more to know what really goes on, please see below.

Some comments on the results: I found out that V0 because of mistake on my part was actually almost identical to base.
I pushed out virtio-net-limit-xmit-polling/v1a instead that actually does what I intended to check. However,
the fact we get such a huge distribution in the results by Tom most likely means that the noise factor is very large.

From my experience one way to get stable results is to divide the throughput by the host CPU utilization (measured by something like mpstat).
Sometimes throughput doesn't increase (e.g. guest-host) by CPU utilization does decrease. So it's interesting.

Another issue is that we are trying to improve the latency of a busy queue here. However STREAM/MAERTS tests ignore the latency (more or less) while TCP_RR by default runs a single packet per queue. Without arguing about whether these are practically interesting workloads, these results are thus unlikely to be significantly affected by the optimization in question.

Quoted 6 lines

What we are interested in, thus, is either TCP_RR with a -b flag (configure with --enable-burst) or multiple concurrent TCP_RRs.

Michael, below are some numbers I got from one round of runs. Thanks,
Roopa

256byte req/response.
Vcpus and irqs were pinned to 4 cores and the cpu utilization is Avg across 4 cores.

base:
Numof concurrent TCP_RRs Num of transactions/sec host cpu-util(%)

1                            7982.93                        15.72
25                           67873                         28.84
50                           112534                        52.25
100                          192057                       86.54

v1

Numof concurrent TCP_RRs    Num of transactions/sec    host cpu-util(%)
1                           7970.94                       10.8
25                          65496.8                       28
50                          109858                        53.22
100                         190155                        87.5

v1a

Numof concurrent TCP_RRs    Num of transactions/sec   host cpu-util (%)
1                           7979.81                       9.5
25                          66786.1                       28
50                          109552                        51
100                         190876                        88

v2
Numof concurrent TCP_RRs Num of transactions/sec host cpu-util (%)

1                            7969.87                     16.5
25                           67780.1                     28.44
50                           114966                      54.29
100                          177982                      79.9
author maintainer
#9 lore raw
In reply to Roopa Prabhu
Quoted 3 lines · Roopa Prabhu

Michael, below are some numbers I got from one round of runs. Thanks,
Roopa

Thanks!
At this point it does not appear like there's any measureable impact from moving the polling around.

Signature · 1 line
MST
author maintainer
#10 lore raw
In reply to Roopa Prabhu
Quoted 3 lines · Roopa Prabhu

Michael, below are some numbers I got from one round of runs. Thanks,
Roopa

Thanks!
At this point it does not appear like there's any measureable impact from moving the polling around.

Signature · 1 line
MST