Realtime Testing Best Practices - Revision history

Glenn: catspec Category:Linux

2008-02-10T20:52:16Z

2008-02-10T19:07:49Z

+cat

2007-11-13T03:25:02Z

‎OLS papers

2007-10-30T12:52:12Z

‎Presentations

2007-10-30T12:05:41Z

‎Presentations

2007-10-30T12:03:53Z

‎Test Table

2007-10-30T12:03:31Z

‎Test Table

2007-10-30T11:51:20Z

‎Presentations

2007-10-30T11:35:37Z

‎Presentations

2007-10-19T00:03:19Z

fix heading levels

← Older revision		Revision as of 19:07, 10 February 2008
Line 350:		Line 350:
	"deterministically" - so I guess RT benchmarks need a notion of usage-profile		"deterministically" - so I guess RT benchmarks need a notion of usage-profile
	to be of value.		to be of value.
		+
		+	[[Category:Categories]]

@@ Line 298: / Line 298: @@
 [FIXTHIS - need to scan for past papers]
 * OLS 2006 BOF - Steven Rostedt, RedHat and Klaas Van Gend, MontaVista - See [http://www.opentux.nl/artikelen/OLS2006_state_of_RT_and_common_mistakes.odp The State of RT and Common Mistakes (OLS 2006 BOF)]
-* OLS 2007 - Paper by Steven Rostedt - see http://www.linuxsymposium.org/2007/view_abstract.php?content_key=75
+* OLS 2007 - Paper by Steven Rostedt - see https://ols2006.108.redhat.com/2007/Reprints/rostedt-Reprint.pdf
 Darren Hart wrote:

@@ Line 231: / Line 231: @@
 === Presentations ===
 [Add links here, most recent at top]
 * [http://tree.celinuxforum.org/CelfPubWiki/ELC2007Presentations?action=AttachFile&do=get&target=CELF_ELC_Interrupt_Latency_2.4_vs_2.6.pdf  Analysis of Interrupt Entry Latency in Linux 2.4 vs 2.6] by !SangBae Lee of Samsung for ELC 2007

@@ Line 239: / Line 239: @@
 ** After reviewing code and finding that the interrupt code path was almost identical, a different, more lightweight tracer was used (Zoom-in tracer) showing latencies were almost the same between 2.4 kernel and 2.6 kernel
 *** This ZI instrumentation has low overhead, so it is suitable for interrupt reponse time measurement. It was written by SangBae Lee
-** Also measured on MIPS 264 MHZ (for real TV system)
+** Also measured on MIPS 264 MHZ (for real TV system), but following data was measured on OSK board
 *** Interrupt response time measured:
 **** with LTT instrumentation:

@@ Line 217: / Line 217: @@
 !Person           !!Company!!Hardware                 !!Kernel           !!test method   !!Measurement method!!Results
 |-
-|Sangbae Lee      ||Samsung||OSK - OMAP (ARM) 192 MHZ)||2.4.20 and 2.6.10||using two machine test||ZI instrumentation - measure interrupt reponse latency||2.4.20 - 30~35us, 2.6.20 - 30~35us max ||
+|Sangbae Lee      ||Samsung||OSK - OMAP (ARM) 192 MHZ)||2.4.20 and 2.6.10||using two machine test||ZI instrumentation - measure interrupt reponse latency||2.4.20 - 30~35us, 2.6.10 - 30~35us max ||
 |-
 |Sangbae Lee      ||Samsung||MIPS 264 MHZ             ||2.6.10 ??        ||??            ||??                ||??||

@@ Line 234: / Line 234: @@
 * [http://tree.celinuxforum.org/CelfPubWiki/ELC2007Presentations?action=AttachFile&do=get&target=CELF_ELC_Interrupt_Latency_2.4_vs_2.6.pdf  Analysis of Interrupt Entry Latency in Linux 2.4 vs 2.6] by !SangBae Lee of Samsung for ELC 2007
 ** Analyzed MV 3.1 (2.4.20) and MV 4.0 (2.6.10), using LTT, on OSK board (OMAP 5920 ARM 192 MHZ)
 ** Initial results were that linux.2.4.20 was 3X fast for best-case interrupt latency by using LTT
 ** After reviewing code and finding that the interrupt code path was almost identical, a different, more lightweight tracer was used (Zoom-in tracer) showing latencies were almost the same between 2.4 kernel and 2.6 kernel
 ** Also measured on MIPS 264 MHZ (for real TV system)
 *** Interrupt response time measured:
@@ Line 242: / Line 245: @@
 ***** 2.4.20 - min = 10 us, max = 30 us
 **** with ZI instrumentation:
-***** 2.6.10 - min = 3 us, max = ??
+***** 2.6.10 - min = 3 us, max = 30~35 us
-***** 2.4.20 - min = ??, max = ??
+***** 2.4.20 - min = 3 us, max = 30~35 us
 *** Basic result = Don't use LTT for measuring RT performance
 * [http://tree.celinuxforum.org/CelfPubWiki/ELC2007Presentations?action=AttachFile&do=get&target=preempt070418celfelc.pdf  Porting and Evaluating the Linux Realtime Preemption on Embedded Platform] by Katsuya Matsubara of Igel at ELC 2007

@@ Line 150: / Line 150: @@
 * How is performance measured?
-== Using the LATENCY_TRACE option ==
+=== Using the LATENCY_TRACE option ===
 Quote about latency-test from Ingo:
 <pre>
@@ Line 168: / Line 168: @@
 </pre>
-== Number of samples recommended ==
+=== Number of samples recommended ===
 Ingo wrote:
@@ Line 179: / Line 179: @@
   overhead.
-== Things to watch for in testing ==
+=== Things to watch for in testing ===
 Another note from Ingo - see [http://groups.google.com/group/linux.kernel/msg/8c7e61d0926dba80?hl=en& here]
 * Note the bit about IRQ 7 - what's up with that?
@@ Line 210: / Line 210: @@
 </pre>
-= Tests results taxonomy =
+== Tests results taxonomy ==
-== Test Table ==
+=== Test Table ===
 {| border="1" cellspacing="0" cellpadding="5"
@@ Line 306: / Line 306: @@
 * Low-latency HowTo (for audio) - http://lowlatency.linuxaudio.org/
-== Notes on ineffective tests ==
+=== Notes on ineffective tests ===
 Nicholas McGuire wrote: