Re: [PATCH 2/6] tools/perf/json: Add cache metrics for s390 z16

[Thomas Richter <tmricht@linux.ibm.com>]

On 3/22/23 21:59, Arnaldo Carvalho de Melo wrote:
> Em Tue, Mar 14, 2023 at 06:36:24PM -0300, Arnaldo Carvalho de Melo escreveu:
>> Em Tue, Mar 14, 2023 at 09:34:46AM -0700, Ian Rogers escreveu:
>>> On Tue, Mar 14, 2023 at 1:20 AM Thomas Richter <tmricht@linux.ibm.com> wrote:
>>>>
>>>> On 3/13/23 19:33, Arnaldo Carvalho de Melo wrote:
>>>>> Em Mon, Mar 13, 2023 at 08:22:44AM -0700, Ian Rogers escreveu:
>>>>>> On Mon, Mar 13, 2023 at 1:30 AM Thomas Richter <tmricht@linux.ibm.com> wrote:
>>>>>>>
>>>>>>> Add metrics for s390 z16
>>>>>>> - Percentage sourced from Level 2 cache
>>>>>>> - Percentage sourced from Level 3 on same chip cache
>>>>>>> - Percentage sourced from Level 4 Local cache on same book
>>>>>>> - Percentage sourced from Level 4 Remote cache on different book
>>>>>>> - Percentage sourced from memory
>>>>>>>
>>>>>>> For details about the formulas see this documentation:
>>>>>>> https://www.ibm.com/support/pages/system/files/inline-files/CPU%20MF%20Formulas%20including%20z16%20-%20May%202022_1.pdf
>>>>>>>
>>>>>>> Outpuf after:
>>>>>>>  # ./perf stat -M l4rp -- dd if=/dev/zero of=/dev/null bs=10M count=10K
>>>>>>>  .... dd output deleted
>>>>>>>
>>>>>>>  Performance counter stats for 'dd if=/dev/zero of=/dev/null bs=10M count=10K':
>>>>>>>
>>>>>>>                  0      IDCW_OFF_DRAWER_CHIP_HIT         #     0.00 l4rp
>>>>>>>            431,866      L1I_DIR_WRITES
>>>>>>>              2,395      IDCW_OFF_DRAWER_IV
>>>>>>>                  0      ICW_OFF_DRAWER
>>>>>>>                  0      IDCW_OFF_DRAWER_DRAWER_HIT
>>>>>>>              1,437      DCW_OFF_DRAWER
>>>>>>>        425,960,793      L1D_DIR_WRITES
>>>>>>>
>>>>>>>       12.165030699 seconds time elapsed
>>>>>>>
>>>>>>>        0.001037000 seconds user
>>>>>>>       12.162140000 seconds sys
>>>>>>>
>>>>>>>  #
>>>>>>>
>>>>>>> Signed-off-by: Thomas Richter <tmricht@linux.ibm.com>
>>>>>>> Acked-By: Sumanth Korikkar <sumanthk@linux.ibm.com>
>>>>>>
>>>>>> Acked-by: Ian Rogers <irogers@google.com>
>>>>>
>>>>> Thanks, applied the first two patches, please address the review
>>>>> suggestions for patches 3-6 and resubmit only those.
>>>>>
>>>>> The patches will be in the public perf-tools-next branch later today.
>>>>>
>>>>> - Arnaldo
>>>>>
>>>>
>>>> I would really prefer the current implementation without using "ScaleUnit": "100%"
>>>> The reason is that these formulars are given to me from the s390 Performance team.
>>>> They want to use the exact same formulars on all platforms running on s390
>>>> which includes z/OS and z/VM. This way they are sure to get the same numbers.
>>>>
>>>> Hope this background info helps.
>>>
>>> For the series:
>>> Acked-by: Ian Rogers <irogers@google.com>
>>
>> Thanks, applied.
>>
>> - Arnaldo
> 
> While trying to cross build to s390 on:
> 
> ubuntu:18.04
> 
> using python3
>  
> 
>    CC      /tmp/build/perf/tests/parse-events.o
> Exception processing pmu-events/arch/s390/cf_z16/extended.json
> Traceback (most recent call last):
>   File "pmu-events/jevents.py", line 997, in <module>
>     main()
>   File "pmu-events/jevents.py", line 979, in main
>     ftw(arch_path, [], preprocess_one_file)
>   File "pmu-events/jevents.py", line 935, in ftw
>     ftw(item.path, parents + [item.name], action)
>   File "pmu-events/jevents.py", line 933, in ftw
>     action(parents, item)
>   File "pmu-events/jevents.py", line 514, in preprocess_one_file
>     for event in read_json_events(item.path, topic):
>   File "pmu-events/jevents.py", line 388, in read_json_events
>     events = json.load(open(path), object_hook=JsonEvent)
>   File "/usr/lib/python3.6/json/__init__.py", line 296, in load
>     return loads(fp.read(),
>   File "/usr/lib/python3.6/encodings/ascii.py", line 26, in decode
>     return codecs.ascii_decode(input, self.errors)[0]
> UnicodeDecodeError: 'ascii' codec can't decode byte 0xe2 in position 4271: ordinal not in range(128)
> 
>

Hmmm, this is very strange. After reading this mail I installed Ubuntu 18.04
on my s390 system. The build works fine, no errors at all.


# pmu-events/jevents.py s390 all pmu-events/arch pmu-events/pmu-events.c
# ll pmu-events/pmu-events.c
-rw-r--r-- 1 root root 317284 Mar 23 10:46 pmu-events/pmu-events.c
#

The file has the correct contents and the build works fine too.
# make 
....
Auto-detecting system features:
...                                   dwarf: [ on  ]
...                      dwarf_getlocations: [ on  ]
...                                   glibc: [ on  ]
...                                  libbfd: [ on  ]
...                          libbfd-buildid: [ on  ]
...                                  libcap: [ OFF ]
...                                  libelf: [ on  ]
...                                 libnuma: [ on  ]
...                  numa_num_possible_cpus: [ on  ]
...                                 libperl: [ on  ]
...                               libpython: [ on  ]
...                               libcrypto: [ on  ]
...                               libunwind: [ OFF ]
...                      libdw-dwarf-unwind: [ on  ]
...                                    zlib: [ on  ]
...                                    lzma: [ on  ]
...                               get_cpuid: [ OFF ]
...                                     bpf: [ on  ]
...                                  libaio: [ on  ]
...                                 libzstd: [ OFF ]

  INSTALL libsubcmd_headers
  INSTALL libsymbol_headers
  INSTALL libperf_headers
  INSTALL libapi_headers
  INSTALL libbpf_headers
  CC      pmu-events/pmu-events.o
  LD      pmu-events/pmu-events-in.o
  LINK    perf
# ./perf list | grep -A 20 basic:
basic:
  CPU_CYCLES
       [Cycle Count. Unit: cpum_cf]
  INSTRUCTIONS
       [Instruction Count. Unit: cpum_cf]
  L1D_DIR_WRITES
       [Level-1 D-Cache Directory Write Count. Unit: cpum_cf]
  L1D_PENALTY_CYCLES
       [Level-1 D-Cache Penalty Cycle Count. Unit: cpum_cf]
  L1I_DIR_WRITES
       [Level-1 I-Cache Directory Write Count. Unit: cpum_cf]
  L1I_PENALTY_CYCLES
       [Level-1 I-Cache Penalty Cycle Count. Unit: cpum_cf]
  PROBLEM_STATE_CPU_CYCLES
       [Problem-State Cycle Count. Unit: cpum_cf]
  PROBLEM_STATE_INSTRUCTIONS
       [Problem-State Instruction Count. Unit: cpum_cf]


So everythings works as usual.
-- 
Thomas Richter, Dept 3303, IBM s390 Linux Development, Boeblingen, Germany
--
Vorsitzender des Aufsichtsrats: Gregor Pillen
Geschäftsführung: David Faller
Sitz der Gesellschaft: Böblingen / Registergericht: Amtsgericht Stuttgart, HRB 243294