This repository has been archived by the owner on Sep 5, 2023. It is now read-only.
-
Notifications
You must be signed in to change notification settings - Fork 56
Commit
This commit does not belong to any branch on this repository, and may belong to a fork outside of the repository.
Merge pull request #1226 from osalyk/cpuload_part
tools: add part_cpuload templates
- Loading branch information
Showing
3 changed files
with
202 additions
and
12 deletions.
There are no files selected for viewing
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
| Original file line number | Diff line number | Diff line change |
|---|---|---|
| @@ -0,0 +1,143 @@ | ||
| { | ||
| "common": { | ||
| "server_global": [ | ||
| "create_serialize=0", | ||
| "kb_base=1000", | ||
| "serverip=$serverip", | ||
| "port=7204", | ||
| "thread" | ||
| ], | ||
| "server_cpu": [ | ||
| "[cpuio]", | ||
| "ioengine=cpuio", | ||
| "cpuload={0, 25, 50, 75, 80, 85, 90, 95, 99, 100}", | ||
| "numjobs=$cores_per_socket" | ||
| ], | ||
| "client_global": [ | ||
| "create_serialize=0", | ||
| "serverip=$serverip", | ||
| "port=7204", | ||
| "thread", | ||
| "disable_clat=1", | ||
| "lat_percentiles=1", | ||
| "percentile_list=99.0:99.9:99.99:99.999" | ||
| ], | ||
| "client_common": [ | ||
| "readwrite={write, randwrite}", | ||
| "blocksize=$blocksize", | ||
| "ramp_time=15", | ||
| "time_based", | ||
| "runtime=60" | ||
| ] | ||
| }, | ||
| "config": { | ||
| "lat": { | ||
| "type": "kvtable", | ||
| "APM Server - FIO engine configuration": [ | ||
| "[global]", | ||
| "ioengine=librpma_apm_server", | ||
| "{server_global}", | ||
| "", | ||
| "[server]", | ||
| "direct_write_to_pmem=1", | ||
| "numjobs=1", | ||
| "size=100MiB", | ||
| "filename=/dev/dax/path", | ||
| "", | ||
| "{server_cpu}" | ||
| ], | ||
| "APM Client - FIO engine configuration": [ | ||
| "[global]", | ||
| "ioengine=librpma_apm_client", | ||
| "{client_global}", | ||
| "", | ||
| "[client]", | ||
| "sync=1", | ||
| "{client_common}" | ||
| ], | ||
| "GPSPM(-RT) Server - FIO engine configuration": [ | ||
| "[global]", | ||
| "ioengine=librpma_gpspm_server", | ||
| "{server_global}", | ||
| "", | ||
| "[server]", | ||
| "direct_write_to_pmem=0", | ||
| "numjobs=1", | ||
| "iodepth=1", | ||
| "size=100MiB", | ||
| "filename=/dev/dax/path", | ||
| "busy_wait_polling={{0, 1}} # 1 for GPSPM-RT", | ||
| "", | ||
| "time_based", | ||
| "runtime=365d", | ||
| "", | ||
| "{server_cpu}" | ||
| ], | ||
| "GPSPM(-RT) Client - FIO engine configuration": [ | ||
| "[global]", | ||
| "ioengine=librpma_gpspm_client", | ||
| "{client_global}", | ||
| "", | ||
| "[client]", | ||
| "sync=1", | ||
| "{client_common}" | ||
| ] | ||
| }, | ||
| "bw": { | ||
| "type": "kvtable", | ||
| "APM Server - FIO engine configuration": [ | ||
| "[global]", | ||
| "ioengine=librpma_apm_server", | ||
| "{server_global}", | ||
| "", | ||
| "[server]", | ||
| "direct_write_to_pmem=1", | ||
| "numjobs=$numjobs", | ||
| "size=100MiB", | ||
| "filename=/dev/dax/path", | ||
| "", | ||
| "{server_cpu}" | ||
| ], | ||
| "APM Client - FIO engine configuration": [ | ||
| "[global]", | ||
| "ioengine=librpma_apm_client", | ||
| "{client_global}", | ||
| "", | ||
| "[client]", | ||
| "numjobs=$numjobs", | ||
| "group_reporting=1", | ||
| "iodepth=2", | ||
| "{client_common}" | ||
| ], | ||
| "GPSPM(-RT) Server - FIO engine configuration": [ | ||
| "[global]", | ||
| "ioengine=librpma_gpspm_server", | ||
| "{server_global}", | ||
| "", | ||
| "[server]", | ||
| "direct_write_to_pmem=0", | ||
| "numjobs=$numjobs", | ||
| "iodepth=2", | ||
| "size=100MiB", | ||
| "filename=/dev/dax/path", | ||
| "busy_wait_polling={{0, 1}} # 1 for GPSPM-RT", | ||
| "", | ||
| "time_based", | ||
| "runtime=365d", | ||
| "", | ||
| "{server_cpu}" | ||
| ], | ||
| "GPSPM(-RT) Client - FIO engine configuration": [ | ||
| "[global]", | ||
| "ioengine=librpma_gpspm_client", | ||
| "{client_global}", | ||
| "", | ||
| "[client]", | ||
| "numjobs=$numjobs", | ||
| "group_reporting=1", | ||
| "iodepth=2", | ||
| "{client_common}" | ||
| ] | ||
| } | ||
| } | ||
| } |
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
| Original file line number | Diff line number | Diff line change |
|---|---|---|
| @@ -0,0 +1,47 @@ | ||
| {% if menu is true %} | ||
| <a class="pure-menu-heading" href="#cpuload">CPU load impact</a> | ||
| <ul class="pure-menu-list"> | ||
| <li class="pure-menu-item"><a href="#cpuload-lat" class="pure-menu-link">Latency</a></li> | ||
| <li class="pure-menu-item"><a href="#cpuload-bw" class="pure-menu-link">Bandwidth</a></li> | ||
| </ul> | ||
| {% else %} | ||
|
|
||
| <h2 id="cpuload">CPU load impact</h2> | ||
|
|
||
| Benchmarking an impact of the CPU load running on **the RPMA Target** on writing data persistently using both ways of achieving remote persistency: **APM** and **GPSPM(-RT)**. | ||
|
|
||
| As a starting point are chosen well-performing workload configurations in respect of either latency or bandwidth. The CPU load on **the RPMA Target** is expected to affect all persistency methods requiring active participation of the process running on **the RPMA Target** namely **GPSPM-RT** and **GPSPM**. Whereas **APM** since it does not make use of the CPU of **the RPMA Target** should not be significantly affected. | ||
|
|
||
| <h3 id="cpuload-lat">CPU load impact: Latency</h3> | ||
|
|
||
| Comparing the latency of **APM** to PMem on **the RPMA Target** vs the latency of **GPSPM(-RT)** to PMem on **the RPMA Target** (with *Direct Write to PMem* disabled) affected by various CPU loads running on **the RPMA Target**. | ||
|
|
||
| {{config.lat}} | ||
|
|
||
| {{figure.apm_vs_gpspm_cpu_00_99.lat_avg}} | ||
|
|
||
| {{figure.apm_vs_gpspm_cpu_00_99.lat_pctl_999}} | ||
|
|
||
| {{figure.apm_vs_gpspm_cpu_00_99.lat_pctl_9999}} | ||
|
|
||
| {{figure.apm_vs_gpspm_cpu_75_99.lat_avg}} | ||
|
|
||
| {{figure.apm_vs_gpspm_cpu_75_99.lat_pctl_999}} | ||
|
|
||
| {{figure.apm_vs_gpspm_cpu_75_99.lat_pctl_9999}} | ||
|
|
||
| <h3 id="cpuload-bw">CPU load impact: Bandwidth</h3> | ||
|
|
||
| Comparing the bandwidth of **APM** to PMem on **the RPMA Target** vs the bandwidth of **GPSPM(-RT)** to PMem on **the RPMA Target** affected by various CPU loads running on **the RPMA Target**. | ||
|
|
||
| {{config.bw}} | ||
|
|
||
| {{figure.apm_vs_gpspm_cpu_00_99.bw}} | ||
|
|
||
| {{figure.apm_vs_gpspm_cpu_00_99.bw_mt}} | ||
|
|
||
| {{figure.apm_vs_gpspm_cpu_75_99.bw}} | ||
|
|
||
| {{figure.apm_vs_gpspm_cpu_75_99.bw_mt}} | ||
|
|
||
| {% endif %} |