On Thu, 1 Apr 2021 11:18:10 +0200
Cédric Le Goater [off-list ref] wrote:
Hello,
On 4/1/21 10:04 AM, Greg Kurz wrote:
quoted
On Wed, 31 Mar 2021 16:45:05 +0200
Cédric Le Goater [off-list ref] wrote:
quoted
Hello,
ipistorm [*] can be used to benchmark the raw interrupt rate of an
interrupt controller by measuring the number of IPIs a system can
sustain. When applied to the XIVE interrupt controller of POWER9 and
POWER10 systems, a significant drop of the interrupt rate can be
observed when crossing the second node boundary.
This is due to the fact that a single IPI interrupt is used for all
CPUs of the system. The structure is shared and the cache line updates
impact greatly the traffic between nodes and the overall IPI
performance.
As a workaround, the impact can be reduced by deactivating the IRQ
lockup detector ("noirqdebug") which does a lot of accounting in the
Linux IRQ descriptor structure and is responsible for most of the
performance penalty.
As a fix, this proposal allocates an IPI interrupt per node, to be
shared by all CPUs of that node. It solves the scaling issue, the IRQ
lockup detector still has an impact but the XIVE interrupt rate scales
linearly. It also improves the "noirqdebug" case as showed in the
tables below.
As explained by David and others, NUMA nodes happen to match sockets
with current POWER CPUs but these are really different concepts. NUMA
is about CPU memory accesses latency,
This is exactly our problem. we have cache issues because hw threads
on different chips are trying to access the same structure in memory.
It happens on virtual platforms and baremetal platforms. This is not
restricted to pseries.
1 more quoted lines
Ok, I get it... the XIVE HW accesses structures in RAM, just like HW threads
do, so the closer, the better. This definitely looks NUMA related indeed. So
yes, the idea of having the XIVE HW to only access local in-RAM data when
handling IPIs between vCPUs in the same NUMA node makes sense.
What is less clear is the exact role of ibm,chip-id actually. This is
currently used on PowerNV only to pick up a default target on the same
"chip" as the source if possible. What is the detailed motivation behind
this ?
quoted
while in the case of XIVE you
really need to identify a XIVE chip localized in a given socket.
PAPR doesn't know about sockets, only cores. In other words, a PAPR
compliant guest sees all vCPUs like they all sit in a single socket.
There are also NUMA nodes on PAPR.
Yes but nothing prevents a NUMA node to span over multiple sockets
or having several NUMA nodes within the same socket, even if this
isn't the case in practice with current POWER hardware.
quoted
Same for the XIVE. Trying to introduce a concept of socket, either
by hijacking OPAL's ibm,chip-id or NUMA node ids, is a kind of
spec violation in this context. If the user cares for locality of
the vCPUs and XIVE on the same socket, then it should bind vCPU
threads to host CPUs from the same socket in the first place.
Yes. that's a must have of course. You need to reflect the real HW
topology in the guest or LPAR if you are after performance, or
restrict the virtual machine to be on a single socket/chip/node.
And this is not only a XIVE problem. XICS has the same problem with
a shared single IPI interrupt descriptor but XICS doesn't scale well
by design, so it doesn't show.
quoted
Isn't this enough to solve the performance issues this series
want to fix, without the need for virtual socket ids ?
what are virtual socket ids ? A new concept ?
For now, we have virtual CPUs identified by a virtual CPU id.
It thus seems natural to speak of a virtual socket id, but
anyway, the wording isn't really important here and you
don't answer the question ;-)
Thanks,
C.
quoted
quoted
* P9 DD2.2 - 2s * 64 threads
"noirqdebug"
Mint/s Mint/s
chips cpus IPI/sys IPI/chip IPI/chip IPI/sys
--------------------------------------------------------------
1 0-15 4.984023 4.875405 4.996536 5.048892
0-31 10.879164 10.544040 10.757632 11.037859
0-47 15.345301 14.688764 14.926520 15.310053
0-63 17.064907 17.066812 17.613416 17.874511
2 0-79 11.768764 21.650749 22.689120 22.566508
0-95 10.616812 26.878789 28.434703 28.320324
0-111 10.151693 31.397803 31.771773 32.388122
0-127 9.948502 33.139336 34.875716 35.224548
* P10 DD1 - 4s (not homogeneous) 352 threads
"noirqdebug"
Mint/s Mint/s
chips cpus IPI/sys IPI/chip IPI/chip IPI/sys
--------------------------------------------------------------
1 0-15 2.409402 2.364108 2.383303 2.395091
0-31 6.028325 6.046075 6.089999 6.073750
0-47 8.655178 8.644531 8.712830 8.724702
0-63 11.629652 11.735953 12.088203 12.055979
0-79 14.392321 14.729959 14.986701 14.973073
0-95 12.604158 13.004034 17.528748 17.568095
2 0-111 9.767753 13.719831 19.968606 20.024218
0-127 6.744566 16.418854 22.898066 22.995110
0-143 6.005699 19.174421 25.425622 25.417541
0-159 5.649719 21.938836 27.952662 28.059603
0-175 5.441410 24.109484 31.133915 31.127996
3 0-191 5.318341 24.405322 33.999221 33.775354
0-207 5.191382 26.449769 36.050161 35.867307
49 more quoted lines