Sorry, you need to enable JavaScript to visit this website.

Feedback

Your feedback is important to keep improving our website and offer you a more reliable experience.

Linux Kernel Performance

Linux development evolves rapidly. The performance and scalability of the OS kernel has been a key part of its success. However, discussions have appeared on LKML (Linux Kernel Mailing List) regarding large performance regression between kernel versions. These discussions underscore the need for a systematic and disciplined way to characterize, improve, and test Linux kernel performance. Our goal is to work with the Linux community to further enhance the Linux kernel with consistent performance increases (avoiding degradations) across releases. The information available on this site gives community members better information about what 0-Day and LKP (Linux Kernel Performance) are doing to preserve performance integrity of the kernel.

0-Day CI Linux Kernel Performance Report (v5.2)

BY Rong Chen ON Aug 12, 2019
  1. Introduction

0-Day CI is an automated Linux kernel test service that provides comprehensive test coverage of the Linux kernel. It covers kernel build, static analysis, boot, functional, performance and power tests. This report shows the recent observations of kernel performance status on IA platform based on the test results from 0-Day CI service. It is structured in the following manner:

  • Section 2, merged regressions and improvements in v5.2 release candidates

  • Section 3, test parameter description

  • Section 4, captured regressions and improvements by shift-left testing during developers’ and maintainers’ tree during v5.2 release cycle

  • Section 5, performance comparison among different kernel releases

  • Section 6, test machine list

 

  1. Test parameters

Here are the descriptions for each parameter/field used in the tests. 

 

Classification

Name

Description

General

runtime

Run the test case within a certain time period (seconds or minutes)

 

nr_task

If it is an integer, which means the number of processes/threads (to run the workload) of this job. Default is 1.

If it is a percentage, e.g. 200% means the number of processes/threads is double of cpu number

 

nr_threads

Alias of nr_task

 

iterations

Number to repeat this job

 

test_size

Test disk size or memory size

 

set_nic_irq_affinity

Set NIC interrupt affinity

 

disable_latency_stats

Latency_stats may introduce too much noise if there are too many context switches, allow to disable it

 

transparent_hugepage

Set transparent hugepage policy (/sys/kernel/mm/transparent_hugepage)

 

boot_params:bp1_memmap

Boot parameters of memmap

 

disk:nr_pmem

number of pmem partitions used by test

 

swap:priority

Priority means the  priority of the swap device. priority is a value between -1 and 32767, the default is -1 and higher priority with higher value. 

Test Machine

model

Name of Intel processor microarchitecture

 

brand

Brand name of cpu

 

cpu_number

Number of cpu

 

memory

Size of memory


 

  1. Linux Kernel v5.2 Release Test

The v5.2 release of the Linux kernel was on July 7, 2019. Some of the more significant changes in 5.2 are a new CLONE_PIDFD flag to clone() to obtain a pidfd for the new process, a significant BPF verifier performance improvement that allows the maximum size of a BPF program to be raised to 1 million instructions, a BPF hook to manage sysctl knobs, a new set of system calls for filesystem mounting, case-insensitive lookups for the ext4 filesystem, a process freezer for version-2 control groups, pressure-stall monitors, and, of course, a vast number of fixes. See the KernelNewbies 5.2 page for a lot more details.

 

0-Day CI monitored the release closely to trace down the performance status on IA platform. 0-Day observed 6 regressions and 5 improvements during feature development phase for v5.2. We will share more detailed information together with correlated patches that led to the results. Note that the assessment is limited by the test coverage 0-Day has now. The list is summarized in the observation summary section.

  1. Observation Summary

0-Day CI observed 6 regressions and 5 improvements during feature development phase for v5.2, which is in the time frame from v5.2-rc1 to v5.2 release. 

Test Indicator

Report

Test Scenario

Test Machine

Development Base

Status

aim7.jobs-per-min

[MD] 4bc034d353: -86.0% regression

disk: 4BRD_12G

md: RAID0

fs: xfs

test: sync_disk_rw

load: 300

cpufreq_governor: performance

lkp-ivb-ep01

v5.1-rc3


 

merged at v5.2-rc1,

fixing patch (merged in v5.2-rc1) provided to restore the performance. 

aim7.jobs-per-min

[btrfs] c8eaeac7b7: -11.7% regression

disk: 4BRD_12G

md: RAID0

fs: btrfs

test: disk_rr

load: 1500

cpufreq_governor: performance

lkp-ivb-ep01

v5.1-rc7

merged at v5.2-rc1, the author is working on it.

fio.write_bw_MBps

[btrfs] 8d47a0d8f7: -28.6% regression

disk: 2pmem

fs: btrfs

runtime: 200s

nr_task: 50%

time_based: tb

rw: randwrite

bs: 4k

ioengine: libaio

test_size: 100G

cpufreq_governor: performance

lkp-hsw-ep2

v5.1-rc7

merged at v5.2-rc1, no response from author yet

fxmark.hdd_btrfs_MWUM_1_bufferedio.works/sec

[loop] 56a85fd837: 224.0% improvement

disk: 1HDD

media: hdd

test: MWUM

fstype: btrfs

directio: bufferedio

cpufreq_governor: performance


 

ivb44

v5.1-rc3

merged at v5.2-rc1

netperf.Throughput_Mbps

[tcp] 8b27dae5a2: -25.7% regression

ip: ipv4

runtime: 900s

nr_threads: 200%

cluster: cs-localhost

test: TCP_STREAM

cpufreq_governor: performance

lkp-skl-fpga01

v5.0

merged at v5.2-rc1.

The author's fix patch merged in 5.2-rc6.

netperf.Throughput_Mbps

[tcp] ede61ca474: 36.5% improvement

ip: ipv4

runtime: 600s

nr_threads: 200%

cluster: cs-localhost

test: TCP_STREAM

cpufreq_governor: performance

lkp-skl-2sp7

v5.2-rc3

merged at v5.2-rc6

netperf.Throughput_Mbps

[tcp] 0b7d7f6b22: 9.7% improvement

ip: ipv4

runtime: 900s

nr_threads: 25%

cluster: cs-localhost

test: TCP_STREAM

cpufreq_governor: performance

lkp-skl-sp2

v5.2-rc3

merged at v5.2-rc6

will-it-scale.per_process_ops

[mm] 42a3003535: -25.9% regression

nr_task: 100%

mode: process

test: page_fault3

cpufreq_governor: performance

lkp-skl-4sp1

v5.1

merged at v5.2-rc1, the author sent a fixing patch and the performance is restored at v5.2-rc6

will-it-scale.per_process_ops

[mm] 815744d751: 43.3% improvement

nr_task: 100%

mode: process

test: page_fault3

cpufreq_governor: performance

lkp-csl-2ap1

v5.2-rc4

merged at v5.2-rc5

will-it-scale.per_process_ops

[x86/modules] f2c65fb322: -2.9% regression

nr_task: 100%

mode: process

test: poll1

cpufreq_governor: performance

lkp-csl-2ap1

v5.1-rc7

merged at v5.2-rc1,

the author is working on it.

will-it-scale.per_thread_ops

[crypto] 71052dcf4b: 59.5% improvement

nr_task: 100%

mode: thread

test: signal1

cpufreq_governor: performance

lkp-knm01

v5.1-rc1

merged at v5.2-rc1

  1. aim7.jobs-per-min

aim7 is a traditional UNIX system level benchmark suite which is used to test and measure the performance of a multiuser system. 

 

  1. scenario: disk_rr test on btrfs

 

Commit c8eaeac7b7 was reported to have -11.7% of aim7.jobs-per-min when comparing to v5.1-rc7." It was merged to mainline at v5.2-rc1.

 

Correlated commits

c8eaeac7b7

btrfs: reserve delalloc metadata differently

branch

linus/master

report

[btrfs] c8eaeac7b7: -11.7% regression

test scenario

disk: 4BRD_12G

md: RAID0

fs: btrfs

test: disk_rr

load: 1500

cpufreq_governor: performance

test machine

lkp-ivb-ep01

status

merged at v5.2-rc1,no response from author yet

 

  1. netperf.Throughput_Mbps

Netperf is a benchmark that can be used to measure the performance of many different types of networking. It provides tests for both unidirectional throughput, and end-to-end latency.

  1. scenario: TCP_STREAM test on localhost

 

 

Commit 8b27dae5a2 was reported to have -25.7% of netperf.Throughput_Mbps when comparing to v5.0." It was merged to mainline at v5.2-rc1.

 

Correlated commits

8b27dae5a2

tcp: add one skb cache for rx

branch

linux-next/master

report

[tcp] 8b27dae5a2: netperf.Throughput_Mbps -25.7% regression

test scenario

ip: ipv4

runtime: 900s

nr_threads: 200%

cluster: cs-localhost

test: TCP_STREAM

cpufreq_governor: performance

test machine

lkp-skl-fpga01

status

The author's fix patch restored the regression, and merged in 5.2-rc6

 

  1. will-it-scale.per_process_ops

Will-it-scale takes a test case and runs it from 1 through to n parallel copies to see if the test case will scale. It builds both process and threads based tests in order to see any differences between the two.

Scenario: process page_fault3

 

 

Commit 42a3003535 was reported to have -25.9% of will-it-scale.per_process_ops when comparing to v5.1." It was merged to mainline at v5.2-rc1.

 

Correlated commits

42a3003535

mm: memcontrol: fix recursive statistics correctness & scalability

branch

linus/master

report

[mm] 42a3003535: -25.9% regression

test scenario

nr_task: 100%

mode: process

test: page_fault3

cpufreq_governor: performance

test machine

lkp-skl-4sp1

status

merged at v5.2-rc1, the author sent a fixing patch and the performance is restored at v5.2-rc6

 

  1. Shift-Left Testing

Beyond testing trees in the upstream kernel, 0-Day CI also tests developers’ and maintainers’ trees, which can catch issues earlier and reduce wider impact. We call it “shift-left” testing. During the v5.2 release cycle, 0-Day CI had reported 12 major performance regressions and 5 major improvements by doing shift-left testing. We will share more detailed information together with possible code changes that led to this result for some of these, though the assessment is limited by the test coverage we have now. The whole list is summarized at report summary section.

  1. Report Summary

0-Day CI had reported 14 performance regressions and 5 improvements by doing shift-left testing on developer and maintainer repos.

 

Test Indicator

Mail

Test Scenario

Test Machine

Status

aim7.jobs-per-min

[sched/fair] 07cf623c04: 69.7% improvement

disk: 1BRD_48G

fs: xfs

test: disk_wrt

load: 3000

cpufreq_governor: performance

lkp-ivb-ep01

Currently not merged

aim7.jobs-per-min

[sched/fair] cca2ae2c00: -35.9% regression

disk: 1BRD_48G

fs: xfs

test: creat-clo

load: 1500

cpufreq_governor: performance

lkp-ivb-ep01

not merged, the regression has been fixed.

aim7.jobs-per-min

[xfs] 490d451fa5: -6.3% regression

disk: 4BRD_12G

md: RAID0

fs: xfs

test: disk_src

load: 3000

cpufreq_governor: performance

lkp-skl-2sp7

merged at v5.3-rc1, no response from author yet

aim7.jobs-per-min

[btrfs] c8eaeac7b7: -11.7% regression

disk: 4BRD_12G

md: RAID0

fs: btrfs

test: disk_rr

load: 1500

cpufreq_governor: performance

lkp-ivb-ep01

Currently not merged, no response from author yet

hackbench.throughput

[sched/fair] 07cf623c04: -70.5% regression

nr_threads: 1600%

mode: process

ipc: pipe

cpufreq_governor: performance

lkp-bdw-ex1

Currently not merged, no response from author yet

hackbench.throughput

94cc25049f: -19.4% regression

nr_threads: 50%

mode: process

ipc: socket

cpufreq_governor: performance

lkp-bdw-ep3

Currently not merged, no response from author yet

hackbench.throughput

[sched/fair] b3e5e9a1dd: 38.7% improvement

test_memory_size: 50%

nr_threads: 100%

mode: development

test: TCP

cpufreq_governor: performance

lkp-bdw-ep4

Currently not merged, no response from author yet

netperf.Throughput_Mbps

[tcp] 8b27dae5a2: -25.7% regression

nr_threads: 100%

mode: threads

ipc: pipe

cpufreq_governor: performance

lkp-cfl-e1

Currently not merged

reaim.jobs_per_min

[x86/topology] 212bf4fdb7: 36.4% improvement

runtime: 300s

nr_task: 100%

test: custom

cpufreq_governor: performance

lkp-bdw-ex1

merged at v5.3-rc1

reaim.jobs_per_min

[locking/rwsem] 4f23dbc1e6: -37.0% regression

runtime: 300s

nr_task: 7000t

test: shared_memory

cpufreq_governor: performance

lkp-ivb-2ep1

merged at v5.3-rc1, the author is working on it.

reaim.jobs_per_min

[sched/fair] d2a99d2651: -2.3% regression

runtime: 300s

nr_task: 100%

test: alltests

cpufreq_governor: performance

lkp-skl-sp2

Currently not merged, no response from author yet

reaim.jobs_per_min

[sched/fair] e1f4f2a13d: -20.1% regression

runtime: 300s

nr_task: 100%

test: fserver

cpufreq_governor: performance

lkp-skl-sp2

Currently not merged, no response from author yet

vm-scalability.median

1f7ab644d6: -11.8% regression

runtime: 300s

size: 512G

test: anon-cow-rand

cpufreq_governor: performance

lkp-bdw-ep6

Currently not merged, no response from author yet

vm-scalability.median

[locking/rwsem] 7d43f1ce9d: vm-scalability.median -16.5% regression

runtime: 300s

size: 8T

test: anon-w-seq-mt

cpufreq_governor: performance

lkp-csl-2ap1

merged at v5.3-rc1

vm-scalability.median

[mm, swap] a839bbfc83: -85.4% regression

runtime: 300

thp_enabled: always

thp_defrag: always

nr_task: 32

nr_ssd: 1

test: swap-w-seq

cpufreq_governor: performance

lkp-skl-2sp

Currently not merged, no response from author yet

will-it-scale.per_process_ops

[mm/page_alloc.c] 60f9751638: 9.1% improvement

nr_task: 100%

mode: process

test: page_fault2

cpufreq_governor: performance

lkp-skl-4sp1

Currently not merged, no response from author yet

will-it-scale.per_thread_ops

[locking/rwsem] cf69482d62: 320.6% improvement

nr_task: 16

mode: thread

test: malloc1

cpufreq_governor: performance

lkp-ivb-d02

merged at v5.3-rc1

  1. aim7.jobs-per-min

aim7 is a traditional UNIX system level benchmark suite which is used to test and measure the performance of a multiuser system. 

 

  1. scenario: creat-clo test on xfs

 

Commit cca2ae2c00 was reported to have -35.9% of aim7.jobs-per-min when comparing to v5.2-rc1.

 

Correlated commits

cca2ae2c00

sched/fair: Replace source_load() & target_load() w/ weighted_cpuload()

branch

linux-de/upstream/rem_rq_load

report

[sched/fair] cca2ae2c00: -35.9% regression

test scenario

disk: 1BRD_48G

fs: xfs

test: creat-clo

load: 1500

cpufreq_governor: performance

test machine

lkp-ivb-ep01

status

Not merged at v5.2, the regression has been fixed.

 

  1. hackbench.throughput

Hackbench is both a benchmark and a stress test for the Linux kernel scheduler. It's main job is to create a specified number of pairs of schedulable entities (either  threads or traditional processes) which communicate via either sockets or pipes and time how long it takes for each pair to send data back and forth.

  1. scenario: process pipe

 

Commit 07cf623c04 was reported to have -70.5% of hackbench.throughput when comparing to v5.1.

 

Correlated commits

07cf623c04

sched/fair: Replace source_load() & target_load() w/ weighted_cpuload()

branch

linux-de/upstream/rem_rq_load

report

[sched/fair] 07cf623c04: -70.5% regression

test scenario

nr_threads: 1600%

mode: process

ipc: pipe

cpufreq_governor: performance

test machine

lkp-bdw-ex1

status

Not merged at v5.2,  no response from author yet

 

  1. will-it-scale.per_process_ops

Will-it-scale takes a test case and runs it from 1 through to n parallel copies to see if the test case will scale. It builds both process and threads based tests in order to see any differences between the two.

Scenario: process page_fault2

 

 

Commit 60f9751638 was reported to have 9.1% of will-it-scale.per_process_ops when comparing to v5.2-rc1.

 

Correlated commits

60f9751638

mm/page_alloc.c: Set ppc->high fraction default to 512

branch

net-next-mlx4/net-next-mlx4

report

[mm/page_alloc.c] 60f9751638: 9.1% improvement

test scenario

nr_task: 100%

mode: process

test: page_fault2

cpufreq_governor: performance

test machine

lkp-skl-4sp1

status

merged at v5.2-rc2,  no response from author yet


 

  1. Latest Release Performance Comparing

 

This session gives some information about the performance difference among different kernel releases, especially between v5.2 and v5.1. There are 50+ performance benchmarks running in 0-Day CI, and we selected 9 benchmarks which historically showed the most regressions/improvements reported by 0-Day CI. Some typical configuration/parameters are used to run the test. For some of the regressions from the comparing, 0-Day did not successfully bisect it thus no related report sent out during the release development period, but it is still worth to check. The root cause to cause the regressions won’t be covered in this session. 

 

In the following figures, the value on the Y-axis is the relative performance number. We used the v5.1 data as the base (performance number is 100).

  1. test suite: vm-scalability

vm-scalability exercises functions and regions of the mm subsystem of the Linux kernel. Below 4 tests show the typical test results. 

 

vm-scalability Test 1

vm-scalability Test 2

 

Here are the test configuration and performance test summary for above tests:                                                                                                                                                                                                           

 

vm-scalability Test 1 

vm-scalability Test 2 

test machine

model: Skylake-SP

brand: Intel(R) Xeon(R) Gold 6139 CPU @ 2.30GHz

cpu_number: 72

memory: 128G

model: Haswell-EX

brand: Intel(R) Xeon(R) CPU E7-8890 v3 @ 2.50GHz

cpu_number: 144

memory: 512G

runtime

300s

300s

nr_task

32

No requirement

size

No requirement

1T

disk

nr_ssd: 1

No requirement

transparent_hugepage

thp_enabled: always

thp_defrag: always

No requirement

vm-scalability test parameter

test case: swap-w-seq

test case: lru-shm


 

performance summary

vm-scalability.throughput on kernel v5.2 has -70.4% regression when comparing to v5.1

(It’s caused by commit 1a5f439c7c “mm, swap: fix THP swap out” which is a functionality fix)

vm-scalability.throughput on kernel v5.2 has -11.7% regression when comparing to v5.1

 

  1. test suite: will-it-scale

Will-it-scale takes a test case and runs it from 1 through to n parallel copies to see if the test case will scale. It builds both process and threads based tests in order to see any differences between the two.

 

will-it-scale  Test 1

Will-it-scale Test 2

will-it-scale  Test 3

will-it-scale  Test 4

 

Here are the parameters and performance test summary for above tests:                                                                                                                                                                                                                          

 

will-it-scale Test 1 

will-it-scale Test 2 

will-it-scale Test 3 

will-it-scale Test 4 

test machine

model: Ivy Bridge

brand: Intel(R) Core(TM) i3-3220 CPU @ 3.30GHz

cpu_number: 4

memory: 4G

model: Broadwell-EP

brand: Intel(R) Xeon(R) CPU E5-2699 v4 @ 2.20GHz

cpu_number: 88

memory: 128G

model: Ivy Bridge

brand: Intel(R) Core(TM) i3-3220 CPU @ 3.30GHz

cpu_number: 4

memory: 4G

model: Skylake

brand: Intel(R) Xeon(R) Gold 6140 CPU @ 2.30GHz

cpu_number: 72

memory: 192G

nr_task

1

100%

1

100%

will-it-scale test parameter

test case: pread2

mode: process

test: signal1

test: brk1

mode: process
test: writeseek3

summary

will-it-scale.scalability on kernel v5.2 has -20.9% regression when comparing to v5.1

will-it-scale.per_process_ops on kernel v5.2 has -17.35% regression when comparing to v5.1

will-it-scale.per_thread_ops on kernel v5.2 has -11.95% regression when comparing to v5.1

will-it-scale.per_process_ops on kernel v5.2 has 15.05% improvement when comparing to v5.1

 

  1. test suite: unixbench

UnixBench is a system benchmark to provide a basic indicator of the performance of a Unix-like system.

 

Unixbench Test 1

 

 

Here are the test configuration and performance test summary for above tests:                                                                                                                                                                                                           

 

Unixbench Test 1 

test machine

model: Broadwell-DE

brand: Intel(R) Xeon(R) CPU D-1541 @ 2.10GHz

cpu_number: 16

memory: 8G

runtime

300s

nr_task

1

unixbench test parameter

test case: shell1

performance summary

unixbench.score on kernel v5.2 has -7.62% regression when comparing to v5.1


 

 

  1. test suite: reaim

reaim updates and improves the existing Open Source AIM 7 benchmark. aim7 is a traditional UNIX system level benchmark suite which is used to test and measure the performance of a multiuser system.

 

reaim Test 1                             




 

Here are the test configuration and performance test summary for above tests:                                                                                                                                                                                                                          

 

reaim Test 1 

test machine

model: Skylake-SP

cpu_number: 40

memory: 64G

runtime

300s

nr_task

100%

reaim test parameter

test case: dbase

performance  summary

reaim.jobs_per_min on kernel v5.2 has -2.78% regression when comparing to v5.1

 

  1. test suite: pigz

pigz, which stands for Parallel Implementation of GZip, is a fully functional replacement for gzip that exploits multiple processors and multiple cores to the hilt when compressing data.

pigz Test 1 


 

 

Here are the test configuration and performance test summary for above tests:          

 

 

pigz Test 1

test machine

model: Broadwell-EP

brand: Intel(R) Xeon(R) CPU E5-2699 v4 @ 2.20GHz

cpu_number: 88

memory: 128G

nr_threads

25%

pigz Test parameter

blocksize: 128K

performance  summary

pigz.throughput on kernel v5.2 almost the same with v5.1

            

  1. test suite: netperf

Netperf is a benchmark that can be used to measure the performance of many different types of networking. It provides tests for both unidirectional throughput, and end-to-end latency.

 

netperf Test 1

netperf Test 2



 

Here are the test configuration and performance test summary for above tests:                                                                                                                                                                                                                          

 

netperf Test 1 

netperf Test 2 

test machine

model: Haswell

brand: Intel(R) Core(TM) i7-4770 CPU @ 3.40GHz

cpu_number: 8

memory: 8G

model: Skylake-SP

cpu_number: 40

memory: 64G

disable_latency_stats

1

1

set_nic_irq_affinity

1

1

runtime

300s

300s

nr_threads

200%

25%

ip

ipv4

ipv4

netperf test parameter

send_size: 5K

test case: TCP_SENDFILE

test case: TCP_RR

performance  summary

netperf.Throughput_Mbps on kernel v5.2 has -7.62% regression when comparing to v5.1

netperf.Throughput_tps on kernel v5.2 has -6.84% regression when comparing to v5.1

 

  1. test suite: hackbench

Hackbench is both a benchmark and a stress test for the Linux kernel scheduler. It's  main job is to create a specified number of pairs of schedulable entities (either threads or traditional processes) which communicate via either sockets or pipes and time how long  it takes for each pair to send data back and forth.

hackbench Test 1


 

hackbench Test 2


 

Here are the test configuration and performance test summary for above tests:                                                                                                                                                                                                                          

 

hackbench Test 1 

hackbench Test 2 

test machine

model: Skylake-SP

cpu_number: 192

memory: 256G

model: Coffee Lake

brand: Intel(R) Xeon(R) E-2278G CPU @ 3.40GHz

cpu_number: 16

memory: 32G

disable_latency_stats

1

1

nr_threads

100%

100%

hackbench test parameter

mode: threads

ipc: pipe

mode: process

ipc: socket

performance  summary

hackbench.throughput on kernel v5.2 has -9.6% regression when comparing to v5.1

hackbench.throughput on kernel v5.2 almost the same with v5.1

 

  1. test suite: fio

Fio was originally written to save me the hassle of writing special test case programs when I wanted to test a specific workload, either for performance reasons or to find/reproduce a bug.

fio Test 1 

  

fio Test 2

Here are the test configuration and performance test summary for above tests:                                                                                                                                                                                                                          

 

fio Test 1 

fio Test 2

test machine

model: Broadwell-DE

brand: Intel(R) Xeon(R) CPU D-1541 @ 2.10GHz

cpu_number: 16

memory: 8G

model: Broadwell-DE

brand: Intel(R) Xeon(R) CPU D-1541 @ 2.10GHz

cpu_number: 16

memory: 8G

runtime

300s

300s

file system

btrfs

xfs

disk

1SSD

1SSD

nr_task

8

32

fio test parameter

fio-setup-basic:

  rw: randwrite

  bs: 4k

  ioengine: sync

  test_size: 400g

fio-setup-basic:

  rw: randwrite

  bs: 4k

  ioengine: sync

  test_size: 400g

performance  summary

fio.write_bw_MBps  on kernel v5.2 has -4.98% regression when comparing to v5.1

fio.write_iops on kernel v5.2 has 4.59% improvement when comparing to v5.1

 

  1. test suite: ebizzy

ebizzy is designed to generate a workload resembling common web application server workloads. It is highly threaded, has a large in-memory working set, and allocates and deallocates memory frequently.

ebizzy Test 1

 

 

Here are the test configuration and performance test summary for above test:                                                                                                                                                                                                                          

 

ebizzy Test 1

test machine

model: Skylake-SP

brand: Intel(R) Xeon(R) Gold 6139 CPU @ 2.30GHz

cpu_number: 72

memory: 128G

nr_threads

200%

iterations

100x

ebizzy test parameter

duration: 10s

performance  summary

ebizzy.throughput on kernel v5.2 almost the same with v5.1


 

  1. Test Machines

    1. IVB Desktop

model

Ivy Bridge

brand

Intel(R) Core(TM) i3-3220 CPU @ 3.30GHz

cpu number

8

memory

16G

 

model

Ivy Bridge

brand

Intel(R) Core(TM) i3-3220 CPU @ 3.30GHz

cpu number

4

memory

8G

 

  1. SKL SP

model

Skylake

brand

Intel(R) Xeon(R) CPU E5-2697 v2 @ 2.70GHz

cpu number

80

memory

64G

 

  1. BDW EP

model

Broadwell-EP

brand

Intel(R) Xeon(R) CPU E5-2699 v4 @ 2.20GHz

cpu number

88

memory

128G

 

  1. HSW EP

model

Haswell-EP

brand

Intel(R) Xeon(R) CPU E5-2699 v3 @ 2.30GHz

cpu number

72

memory

128G

 

  1. IVB EP

model

Ivy Bridge-EP

brand

Intel(R) Xeon(R) CPU E5-2690 v2 @ 3.00GHz

cpu number

40

memory

384G

 

model

Ivytown Ivy Bridge-EP

brand

Intel(R) Xeon(R) CPU E5-2697 v2 @ 2.70GHz

cpu number

48

memory

64G

 

  1. HSX EX

model

Brickland Haswell-EX

brand

Intel(R) Xeon(R) CPU E7-8890 v3 @ 2.50GHz

cpu number

144

memory

512G