Thursday, 4 February 2016

Intel Platform Quality of Service and Cache Allocation Technology

One issue when running parallel processes is contention of shared resources such as the Last Level Cache (aka LLC or L3 Cache).  For example, a server may be running a set of Virtual Machines with processes that are memory and cache intensive hence producing a large amount of cache activity. This can impact on the other VMs and is known as the "Noisy Neighbour" problem.

Fortunately the next generation Intel processors allow one to monitor and also fine tune cache allocation using Intel Cache Monitoring Technology (CMT) and Cache Allocation Technology (CAT).

Intel kindly loaned me a 12 thread development machine with CMT and CAT support to experiment with this technology using the Intel pqos tool.   For my experiment, I installed Ubuntu Xenial Server on the machine. I then installed KVM and an VM instance of Ubuntu Xenial Server.   I then loaded the instance using stress-ng running a memory bandwidth stressor:

 stress-ng --stream 1 -v --stream-l3-size 16M  
..which allocates 16MB in 4 buffers and performs various read/compute and writes to these, hence causing a "noisy neighbour".

Using pqos,  one can monitor and see the cache/memory activity:
sudo apt-get install intel-cmt-cat
sudo modprobe msr  
sudo pqos -r  
TIME 2016-02-04 10:25:06
    CORE   IPC  MISSES    LLC[KB]  MBL[MB/s]  MBR[MB/s]
       0  0.59 168259k     9144.0    12195.0        0.0
       1  1.33    107k        0.0        3.3        0.0
       2  0.20      2k        0.0        0.0        0.0
       3  0.70    104k        0.0        2.0        0.0
       4  0.86     23k        0.0        0.7        0.0
       5  0.38     42k       24.0        1.5        0.0
       6  0.12      2k        0.0        0.0        0.0
       7  0.24     48k        0.0        3.0        0.0
       8  0.61     26k        0.0        1.6        0.0
       9  0.37     11k      144.0        0.9        0.0
      10  0.48      1k        0.0        0.0        0.0
      11  0.45      2k        0.0        0.0        0.0
Now to run a stress-ng stream stressor on the host and see the performance while the noisy neighbour is also running:
stress-ng --stream 4 --stream-l3-size 2M --perf --metrics-brief -t 60
stress-ng: info:  [2195] dispatching hogs: 4 stream
stress-ng: info:  [2196] stress-ng-stream: stressor loosely based on a variant of the STREAM benchmark code
stress-ng: info:  [2196] stress-ng-stream: do NOT submit any of these results to the STREAM benchmark results
stress-ng: info:  [2196] stress-ng-stream: Using L3 CPU cache size of 2048K
stress-ng: info:  [2196] stress-ng-stream: memory rate: 1842.22 MB/sec, 736.89 Mflop/sec (instance 0)
stress-ng: info:  [2198] stress-ng-stream: memory rate: 1847.88 MB/sec, 739.15 Mflop/sec (instance 2)
stress-ng: info:  [2199] stress-ng-stream: memory rate: 1833.89 MB/sec, 733.56 Mflop/sec (instance 3)
stress-ng: info:  [2197] stress-ng-stream: memory rate: 1847.16 MB/sec, 738.86 Mflop/sec (instance 1)
stress-ng: info:  [2195] successful run completed in 60.01s (1 min, 0.01 secs)
stress-ng: info:  [2195] stressor      bogo ops real time  usr time  sys time   bogo ops/s   bogo ops/s
stress-ng: info:  [2195]                          (secs)    (secs)    (secs)   (real time) (usr+sys time)
stress-ng: info:  [2195] stream           22101     60.01    239.93      0.04       368.31        92.10
stress-ng: info:  [2195] stream:
stress-ng: info:  [2195]            547,520,600,744 CPU Cycles                     9.12 B/sec
stress-ng: info:  [2195]             69,959,954,760 Instructions                   1.17 B/sec (0.128 instr. per cycle)
stress-ng: info:  [2195]             11,066,905,620 Cache References               0.18 B/sec
stress-ng: info:  [2195]             11,065,068,064 Cache Misses                   0.18 B/sec (99.98%)
stress-ng: info:  [2195]              8,759,154,716 Branch Instructions            0.15 B/sec
stress-ng: info:  [2195]                  2,205,904 Branch Misses                 36.76 K/sec ( 0.03%)
stress-ng: info:  [2195]             23,856,890,232 Bus Cycles                     0.40 B/sec
stress-ng: info:  [2195]            477,143,689,444 Total Cycles                   7.95 B/sec
stress-ng: info:  [2195]                         36 Page Faults Minor              0.60 sec
stress-ng: info:  [2195]                          0 Page Faults Major              0.00 sec
stress-ng: info:  [2195]                         96 Context Switches               1.60 sec
stress-ng: info:  [2195]                          0 CPU Migrations                 0.00 sec
stress-ng: info:  [2195]                          0 Alignment Faults               0.00 sec
.. so about 1842 MB/sec memory rate and 736 Mflop/sec per CPU across 4 CPUs.  And pqos shows the cache/memory actitivity as:
sudo pqos -r
TIME 2016-02-04 10:35:27
    CORE   IPC  MISSES    LLC[KB]  MBL[MB/s]  MBR[MB/s]
       0  0.14  43060k     1104.0     2487.9        0.0
       1  0.12 3981523k     2616.0     2893.8        0.0
       2  0.26    320k       48.0       18.0        0.0
       3  0.12 3980489k     1800.0     2572.2        0.0
       4  0.12 3979094k     1728.0     2870.3        0.0
       5  0.12 3970996k     2112.0     2734.5        0.0
       6  0.04     20k        0.0        0.3        0.0
       7  0.04     29k        0.0        1.9        0.0
       8  0.09    143k        0.0        5.9        0.0
       9  0.15      0k        0.0        0.0        0.0
      10  0.07      2k        0.0        0.0        0.0
      11  0.13      0k        0.0        0.0        0.0
Using pqos again, we can find out how much LLC cache the processor has:
sudo pqos -v
NOTE:  Mixed use of MSR and kernel interfaces to manage
       CAT or CMT & MBM may lead to unexpected behavior.
INFO: Monitoring capability detected
INFO: CPUID.0x7.0: CAT supported
INFO: CAT details: CDP support=0, CDP on=0, #COS=16, #ways=12, ways contention bit-mask 0xc00
INFO: LLC cache size 9437184 bytes, 12 ways
INFO: LLC cache way size 786432 bytes
INFO: L3CA capability detected
INFO: Detected PID API (perf) support for LLC Occupancy
INFO: Detected PID API (perf) support for Instructions/Cycle
INFO: Detected PID API (perf) support for LLC Misses
ERROR: IPC and/or LLC miss performance counters already in use!
Use -r option to start monitoring anyway.
Monitoring start error on core(s) 5, status 6
So this CPU has 12 cache "ways", each of 786432 bytes (768K).  One or more  "Class of Service" (COS)  types can be defined that can use one or more of these ways.  One uses a bitmap with each bit representing a way to indicate how the ways are to be used by a COS.  For example, to use all the 12 ways on my example machine, the bit map is 0xfff  (111111111111).   A way can be exclusively mapped to a COS or shared, or not used at all.   Note that the ways in the bitmap must be contiguously allocated, so a mask such as 0xf3f (111100111111) is invalid and cannot be used.

In my experiment, I want to create 2 COS types, the first COS will have just 1 cache way assigned to it and CPU 0 will be bound to this COS as well as pinning the VM instance to CPU 0  The second COS will have the other 11 cache ways assigned to it, and all the other CPUs can use this COS.

So, create COS #1 with just 1 way of cache, and bind CPU 0 to this COS, and pin the VM to CPU 0:
sudo pqos -e llc:1=0x0001
sudo pqos -a llc:1=0
sudo taskset  -apc 0 $(pidof qemu-system-x86_64)
And create COS #2, with 11 ways of cache and bind CPUs 1-11 to this COS:
sudo pqos -e "llc:2=0x0ffe"
sudo pqos -a "llc:2=1-11"
And let's see the new configuration:
sudo pqos  -s
NOTE:  Mixed use of MSR and kernel interfaces to manage
       CAT or CMT & MBM may lead to unexpected behavior.
L3CA COS definitions for Socket 0:
    L3CA COS0 => MASK 0xfff
    L3CA COS1 => MASK 0x1
    L3CA COS2 => MASK 0xffe
    L3CA COS3 => MASK 0xfff
    L3CA COS4 => MASK 0xfff
    L3CA COS5 => MASK 0xfff
    L3CA COS6 => MASK 0xfff
    L3CA COS7 => MASK 0xfff
    L3CA COS8 => MASK 0xfff
    L3CA COS9 => MASK 0xfff
    L3CA COS10 => MASK 0xfff
    L3CA COS11 => MASK 0xfff
    L3CA COS12 => MASK 0xfff
    L3CA COS13 => MASK 0xfff
    L3CA COS14 => MASK 0xfff
    L3CA COS15 => MASK 0xfff
Core information for socket 0:
    Core 0 => COS1, RMID0
    Core 1 => COS2, RMID0
    Core 2 => COS2, RMID0
    Core 3 => COS2, RMID0
    Core 4 => COS2, RMID0
    Core 5 => COS2, RMID0
    Core 6 => COS2, RMID0
    Core 7 => COS2, RMID0
    Core 8 => COS2, RMID0
    Core 9 => COS2, RMID0
    Core 10 => COS2, RMID0
    Core 11 => COS2, RMID0
..showing Core 0 bound to COS1, and Cores 1-11 bound to COS2, with COS1 with 1 cache way and COS2 with the remaining 11 cache ways.

Now re-run the stream stressor and see if the VM has less impact on the LL3 cache:
stress-ng --stream 4 --stream-l3-size 1M --perf --metrics-brief -t 60
stress-ng: info:  [2232] dispatching hogs: 4 stream
stress-ng: info:  [2233] stress-ng-stream: stressor loosely based on a variant of the STREAM benchmark code
stress-ng: info:  [2233] stress-ng-stream: do NOT submit any of these results to the STREAM benchmark results
stress-ng: info:  [2233] stress-ng-stream: Using L3 CPU cache size of 1024K
stress-ng: info:  [2235] stress-ng-stream: memory rate: 2616.90 MB/sec, 1046.76 Mflop/sec (instance 2)
stress-ng: info:  [2233] stress-ng-stream: memory rate: 2562.97 MB/sec, 1025.19 Mflop/sec (instance 0)
stress-ng: info:  [2234] stress-ng-stream: memory rate: 2541.10 MB/sec, 1016.44 Mflop/sec (instance 1)
stress-ng: info:  [2236] stress-ng-stream: memory rate: 2652.02 MB/sec, 1060.81 Mflop/sec (instance 3)
stress-ng: info:  [2232] successful run completed in 60.00s (1 min, 0.00 secs)
stress-ng: info:  [2232] stressor      bogo ops real time  usr time  sys time   bogo ops/s   bogo ops/s
stress-ng: info:  [2232]                          (secs)    (secs)    (secs)   (real time) (usr+sys time)
stress-ng: info:  [2232] stream           62223     60.00    239.97      0.00      1037.01       259.29
stress-ng: info:  [2232] stream:
stress-ng: info:  [2232]            547,364,185,528 CPU Cycles                     9.12 B/sec
stress-ng: info:  [2232]             97,037,047,444 Instructions                   1.62 B/sec (0.177 instr. per cycle)
stress-ng: info:  [2232]             14,396,274,512 Cache References               0.24 B/sec
stress-ng: info:  [2232]             14,390,808,440 Cache Misses                   0.24 B/sec (99.96%)
stress-ng: info:  [2232]             12,144,372,800 Branch Instructions            0.20 B/sec
stress-ng: info:  [2232]                  1,732,264 Branch Misses                 28.87 K/sec ( 0.01%)
stress-ng: info:  [2232]             23,856,388,872 Bus Cycles                     0.40 B/sec
stress-ng: info:  [2232]            477,136,188,248 Total Cycles                   7.95 B/sec
stress-ng: info:  [2232]                         44 Page Faults Minor              0.73 sec
stress-ng: info:  [2232]                          0 Page Faults Major              0.00 sec
stress-ng: info:  [2232]                         72 Context Switches               1.20 sec
stress-ng: info:  [2232]                          0 CPU Migrations                 0.00 sec
stress-ng: info:  [2232]                          0 Alignment Faults               0.00 sec
Now with the noisy neighbour VM constrained to use just 1 way of LL3 cache, the stream stressor on the host now can achieve about 2592 MB/sec and about 1030 Mflop/sec per CPU across 4 CPUs.

This is a relatively simple example.  With the ability to monitor cache and memory bandwidth activity with one can carefully tune a system to make best use of the limited LL3 cache resource and maximise throughput where needed.

There are many applications where Intel CMT/CAT can be useful, for example fine tuning containers or VM instances, or pinning user space networking buffers to cache ways in DPDK for improved throughput.

Sunday, 31 January 2016

Pagemon improvements

Over the past month I've been finding the odd moments [1] to add some small improvements and fix a few bugs to pagemon (a tool to monitor process memory).  The original code went from a sketchy proof of concept prototype to a somewhat more usable tool in a few weeks, so my main concern recently was to clean up the code and make it more efficient.

With the use of tools such as valgrind's cachegrind and perf I was able to work on some of the code hot-spots [2] and reduce it from ~50-60% CPU down to 5-9% CPU utilisation on my laptop, so it's definitely more machine friendly now.  In addition I've added the following small features:
  • Now one can specify the name of a process to monitor as well as the PID.  This also allows one to run pagemon on itself(!), which is a bit meta.
  • Perf events showing Page Faults and Kernel Page Allocates and Frees, toggled on/off with the 'p' key.
  • Improved and snappier clean up and exit when a monitored process exits.
  • Far more efficient page map reading and rendering.
  • Out of Memory (OOM) scores added to VM statistics window.
  • Process activity (busy, sleeping, etc) to VM statistics window.
  • Zoom mode min/max with '[' (min) and ']' (max) keys.
  • Close pop-up windows with key 'c'.
  • Improved handling of rapid map expansion and shrinking.
  • Jump to end of map using 'End' key.
  • Improve the man page.
I've tried to keep the tool small and focused and I don't want feature bloat to make it unwieldy and overly complexed.  "Do one job, and do it well" is the philosophy behind pagemon. At just 1500 lines of C, it is as complex as I want it to be for now.

Version 0.01.08 should be hitting the Ubuntu 16.04 Xenial Xerus archive in the next 24 hours or so.  I have also the lastest version in my PPA (ppa:colin-king/pagemon) built for Trusty, Vivid, Wily and Xenial.


Pagemon is useful for spotting unexpected memory activity and it is just interesting watching the behaviour memory hungry processes such as web-browsers and Virtual Machines.

Notes:
[1] Mainly very late at night when I can't sleep (but that's another story...).  The git log says it all.
[2] Reading in /proc/$PID/maps and efficiently reading per page data from /proc/$PID/pagemap

Thursday, 28 January 2016

Forcing out bugs with stress-ng

stress-ng logo
Over the past few months I've been adding several new stress tests and a lot more stressor options to stress-ng for Ubuntu 16.04 Xenial Xerus.  I try to track new system calls and features landing in the kernel and where appropriate add a stress test to try and force out bugs.

Stress-ng has found various kernel bugs, such as CVE-2015-1333 and LP:#1526811 as well as bugs in user space (for example, daemons crashing) when memory pressure is very high.  Simple abusive tricks, such as aggressively trying to allocate every free page in memory are useful in finding drivers that don't necessary check for memory allocation failures.  For example, today I was caught out when a USB ethernet dongle driver didn't check for a null pointer due to an allocation failure and stress-ng ended up triggering a kernel oops (fortunately, this bug was fixed in a recent kernel).

The underlying philosophy for stress-ng is "use and abuse standard Linux interfaces and see how far we can push them to destruction".  I'm pretty sure there are plenty of creative folk out there who can dream up dastardly ways to make stress-ng even more stressy, so contributions are always warmly accepted!  I have a mirrorred copy of the git repository on github to make it easy for developers to get their hands on the code.

We've been using stress-ng on ARM based SoC kernels to force out bugs and this has been useful in finding areas where non-swap based systems break. You really don't want your kernel oopsing or processes segfaulting when a IoT device has run low on memory.

My original intent for stress-ng was just to make a system run hot and force thermal overruns. However, I soon discovered it is useful to force kernel bugs out by attempting to (pathologically) thrash most of the system calls.  I've also added perf stats to stress-ng to track performance of standard stress scenarios over kernel versions to get an early warning of any potential performance regressions.  So stress-ng is a bit of a mixed bag of stress tests and performance measuring goodness.

When I get some free time I hope to run stress-ng against a GCOV instrumented kernel at see how much test coverage I get on a kernel. I suspect there are a lot of core kernel functionality still not being touched by stress-ng.

I've also tried to make stress-ng portable, so it can build fine on GNU/Hurd and Debian kFreeBSD (with Linux specific tests not built-in of course). It also contains some architecture specific features, such as handling the data and instruction cache as well as the x86 rdrand instruction and cache line locking. If there are any ARM specific features than can be stressed I'd like to know and perhaps implement stressors for them.

Anyhow, I believe stress-ng is almost feature complete for Ubuntu Xenial, however, I expect it to grow in features over time since there is always new functionality landing in the Linux kernel that needs to be thrashed tested.

Friday, 8 January 2016

FIXME and TODO comments in the Linux kernel source

While looking at some code in the Linux Kernel this morning I spotted a few FIXME comments and that got me wondering just how many there are in the source code.  After a quick grep I found nearly 4200 in v4.4.0-rc8 and that got me thinking about other similar comment tags such as TODO that are in the source and how this has been changing over time.


So the trends are certainly upwards, but then again, so is the size of the kernel source:

Note: Data gathered using sloccount on the lines of C in the kernel source.

Using the sloccount data I then calculated the number of FIXME and TODOs per 1000 lines of code to see what the underlying trend is:

So FIXMEs are actually dropping in relative terms to the size of the kernel where as TODOs are increasing.

Of course, these statistics are bogus because it is dependent on kernel developers adding and removing FIXMEs and TODOs in a consistent manner, however, it is interesting to see how many comments exist and hence how much work has been tagged in comments as work to be done later. I wonder how this compares to other large open source projects.