I am currently porting a project from Haswell to Skylake (i5-6440EQ). When in 
desktop, I observe latency peaks when windows are moved around or zoomed. With 
Haswell CPUs, there is not impact of graphic workload. Unfortunately everything 
above 10us  is an issue in our real time system application.
Currently I am using Xubuntu 16.04 + Kernel 4.9.51 + Xenomai 3.0.6
I already disabled all power- and performance features like Hyper-Threading, 
SpeedStep, Turbo Mode, C-States in BIOS and also by kernel parameters. Also 
disabled i915 power savings like i915.enable_dc=0,  i915.enable_rc6=0, 
i915.enable_execlists=0... and disabled grafik turbo gt_max_freq_mhz = 350  
gt_boost_freq_mhz = 350

One difference we see between Haswell and Skylake: i915 raises a lot of MSI 
when moving windows around where there are almost zero at Haswell.

Any ideas for bringing the peaks down?

== Sampling period: 100 us
== Test mode: periodic user-mode task
== All results in microseconds
warming up...
RTT|  00:00:01  (periodic user-mode task, 100 us period, priority 99)
RTH|----lat min|----lat avg|----lat max|-overrun|---msw|---lat best|--lat worst
RTD|     -0.368|     -0.210|      0.259|       0|     0|     -0.368|      0.259
RTD|     -0.259|     -0.161|      1.601|       0|     0|     -0.368|      1.601
RTD|     -0.290|     -0.174|      1.367|       0|     0|     -0.368|      1.601
RTD|     -0.269|     -0.190|      1.004|       0|     0|     -0.368|      1.601
RTD|     -0.256|     -0.185|      1.000|       0|     0|     -0.368|      1.601
RTD|     -0.267|     -0.148|      1.370|       0|     0|     -0.368|      1.601
RTD|     -0.262|     -0.052|      9.010|       0|     0|     -0.368|      9.010
RTD|     -0.256|     -0.027|     12.111|       0|     0|     -0.368|     12.111 
  <<-- moved terminal window around
RTD|     -0.266|     -0.146|     13.730|       0|     0|     -0.368|     13.730
RTD|     -0.276|     -0.175|      5.509|       0|     0|     -0.368|     13.730
RTD|     -0.314|     -0.159|      6.377|       0|     0|     -0.368|     13.730
RTD|     -0.269|     -0.144|      4.462|       0|     0|     -0.368|     13.730
RTD|     -0.274|     -0.178|      1.749|       0|     0|     -0.368|     13.730
RTD|     -0.267|     -0.123|      2.436|       0|     0|     -0.368|     13.730
RTD|     -0.285|     -0.153|      1.387|       0|     0|     -0.368|     13.730
RTD|     -0.285|     -0.141|      3.113|       0|     0|     -0.368|     13.730
RTD|     -0.313|     -0.160|      1.021|       0|     0|     -0.368|     13.730
RTD|     -0.274|     -0.107|     20.256|       0|     0|     -0.368|     20.256 
 <<-- opened File Manager
RTD|     -0.273|     -0.165|      1.797|       0|     0|     -0.368|     20.256
RTD|     -0.270|     -0.152|      5.486|       0|     0|     -0.368|     20.256
RTD|     -0.328|     -0.189|      1.070|       0|     0|     -0.368|     20.256
RTT|  00:00:22  (periodic user-mode task, 100 us period, priority 99)
RTH|----lat min|----lat avg|----lat max|-overrun|---msw|---lat best|--lat worst
RTD|     -0.275|     -0.201|      2.641|       0|     0|     -0.368|     20.256
RTD|     -0.275|     -0.142|      3.372|       0|     0|     -0.368|     20.256
RTD|     -0.275|     -0.191|      1.020|       0|     0|     -0.368|     20.256
---|-----------|-----------|-----------|--------|------|-------------------------
RTS|     -0.368|     -0.153|     20.256|       0|     0|    00:00:24/00:00:24


Here's what we get with Haswell:
== Sampling period: 100 us
== Test mode: periodic user-mode task
== All results in microseconds
warming up...
RTT|  00:00:01  (periodic user-mode task, 100 us period, priority 99)
RTH|----lat min|----lat avg|----lat max|-overrun|---msw|---lat best|--lat worst
RTD|     -0.399|     -0.341|      1.515|       0|     0|     -0.399|      1.515
RTD|     -0.392|     -0.310|      0.835|       0|     0|     -0.399|      1.515
RTD|     -0.500|     -0.335|      0.634|       0|     0|     -0.500|      1.515
RTD|     -0.529|     -0.333|      0.223|       0|     0|     -0.529|      1.515
RTD|     -0.584|     -0.335|      1.217|       0|     0|     -0.584|      1.515
RTD|     -0.399|     -0.335|      0.566|       0|     0|     -0.584|      1.515
RTD|     -0.502|     -0.336|      0.328|       0|     0|     -0.584|      1.515
RTD|     -0.399|     -0.340|      0.523|       0|     0|     -0.584|      1.515
RTD|     -0.530|     -0.237|      1.819|       0|     0|     -0.584|      1.819
RTD|     -0.212|     -0.159|      2.192|       0|     0|     -0.584|      2.192
RTD|     -0.218|     -0.170|      2.125|       0|     0|     -0.584|      2.192
---|-----------|-----------|-----------|--------|------|-------------------------
RTS|     -0.584|     -0.293|      2.192|       0|     0|    00:00:12/00:00:12


Thanks a lot,
Martin


________________________________
Vertrauliche E-Mail von / Confidential e-mail from Data Modul AG
Vorstand / CEO: Dr. Florian Pesahl
Vorsitzende des Aufsichtsrates / Chairwoman of the Supervisory Board: Kristin 
D. Russell
Sitz der Gesellschaft / Registered Office: M?nchen
Registergericht / Registration Court: M?nchen - Handelsregister B 85 591

Datamodul
[cid:007f-0001@01d3babf.2b093302]
-------------- next part --------------
A non-text attachment was scrubbed...
Name: EW-SEE_2018_d.gif
Type: image/gif
Size: 10542 bytes
Desc: EW-SEE_2018_d.gif
URL: 
<http://xenomai.org/pipermail/xenomai/attachments/20180313/5dc32f75/attachment.gif>
_______________________________________________
Xenomai mailing list
Xenomai@xenomai.org
https://xenomai.org/mailman/listinfo/xenomai

Reply via email to