<!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 4.0 Transitional//EN">
<HTML xmlns="http://www.w3.org/TR/REC-html40" xmlns:v =
"urn:schemas-microsoft-com:vml" xmlns:o =
"urn:schemas-microsoft-com:office:office" xmlns:w =
"urn:schemas-microsoft-com:office:word" xmlns:m =
"http://schemas.microsoft.com/office/2004/12/omml"><HEAD>
<META content="text/html; charset=us-ascii" http-equiv=Content-Type>
<META name=GENERATOR content="MSHTML 8.00.6001.18812">
<STYLE>@font-face {
font-family: Calibri;
}
@page Section1 {size: 8.5in 11.0in; margin: 1.0in 1.0in 1.0in 1.0in; }
P.MsoNormal {
MARGIN: 0in 0in 0pt; FONT-FAMILY: "Calibri","sans-serif"; FONT-SIZE: 11pt
}
LI.MsoNormal {
MARGIN: 0in 0in 0pt; FONT-FAMILY: "Calibri","sans-serif"; FONT-SIZE: 11pt
}
DIV.MsoNormal {
MARGIN: 0in 0in 0pt; FONT-FAMILY: "Calibri","sans-serif"; FONT-SIZE: 11pt
}
A:link {
COLOR: blue; TEXT-DECORATION: underline; mso-style-priority: 99
}
SPAN.MsoHyperlink {
COLOR: blue; TEXT-DECORATION: underline; mso-style-priority: 99
}
A:visited {
COLOR: purple; TEXT-DECORATION: underline; mso-style-priority: 99
}
SPAN.MsoHyperlinkFollowed {
COLOR: purple; TEXT-DECORATION: underline; mso-style-priority: 99
}
P.MsoListParagraph {
MARGIN: 0in 0in 0pt 0.5in; FONT-FAMILY: "Calibri","sans-serif"; FONT-SIZE: 11pt; mso-style-priority: 34
}
LI.MsoListParagraph {
MARGIN: 0in 0in 0pt 0.5in; FONT-FAMILY: "Calibri","sans-serif"; FONT-SIZE: 11pt; mso-style-priority: 34
}
DIV.MsoListParagraph {
MARGIN: 0in 0in 0pt 0.5in; FONT-FAMILY: "Calibri","sans-serif"; FONT-SIZE: 11pt; mso-style-priority: 34
}
SPAN.EmailStyle17 {
FONT-FAMILY: "Calibri","sans-serif"; COLOR: windowtext; mso-style-type: personal-compose
}
.MsoChpDefault {
mso-style-type: export-only
}
DIV.Section1 {
page: Section1
}
OL {
MARGIN-BOTTOM: 0in
}
UL {
MARGIN-BOTTOM: 0in
}
</STYLE>
<!--[if gte mso 9]><xml>
<o:shapedefaults v:ext="edit" spidmax="1026" />
</xml><![endif]--><!--[if gte mso 9]><xml>
<o:shapelayout v:ext="edit">
<o:idmap v:ext="edit" data="1" />
</o:shapelayout></xml><![endif]--></HEAD>
<BODY lang=EN-US link=blue vLink=purple>
<DIV dir=ltr align=left><SPAN class=562433617-12082009><FONT color=#0000ff
size=2 face=Arial>Mike,</FONT></SPAN></DIV>
<DIV dir=ltr align=left><SPAN class=562433617-12082009><FONT color=#0000ff
size=2 face=Arial> I have
forwarded your note to the mvapich-discuss list, which is the right list for
MVAPICH2 related issues.</FONT></SPAN></DIV>
<DIV dir=ltr align=left><SPAN class=562433617-12082009><FONT color=#0000ff
size=2 face=Arial></FONT></SPAN> </DIV>
<DIV dir=ltr align=left><SPAN class=562433617-12082009><FONT color=#0000ff
size=2 face=Arial>Rajeev</FONT></SPAN></DIV>
<DIV dir=ltr align=left><SPAN class=562433617-12082009></SPAN> </DIV><BR>
<BLOCKQUOTE
style="BORDER-LEFT: #0000ff 2px solid; PADDING-LEFT: 5px; MARGIN-LEFT: 5px; MARGIN-RIGHT: 0px"
dir=ltr>
<DIV dir=ltr lang=en-us class=OutlookMessageHeader align=left>
<HR tabIndex=-1>
<FONT size=2 face=Tahoma><B>From:</B> mpich2-dev-bounces@mcs.anl.gov
[mailto:mpich2-dev-bounces@mcs.anl.gov] <B>On Behalf Of </B>Mike
Heinz<BR><B>Sent:</B> Wednesday, August 12, 2009 10:40 AM<BR><B>To:</B>
mpich2-dev@mcs.anl.gov<BR><B>Cc:</B> Todd Rimmer<BR><B>Subject:</B>
[mpich2-dev] MVAPICH2 does not work with specified PKEYs.<BR></FONT><BR></DIV>
<DIV></DIV>
<DIV class=Section1>
<P class=MsoNormal>My testers are reporting further problems with mvapich2. On
a fabric where the use of pkeys is required, mvapich2 is
failing.<o:p></o:p></P>
<P class=MsoNormal><o:p> </o:p></P>
<P style="TEXT-INDENT: -0.25in; MARGIN-LEFT: 0.25in; mso-list: l0 level1 lfo1"
class=MsoListParagraph><![if !supportLists]><SPAN
style="mso-list: Ignore">1)<SPAN
style="FONT: 7pt 'Times New Roman'">
</SPAN></SPAN><![endif]>The MV2_DEFAULT_PKEY parameter does not appear to be
supported when using mpirun_rsh.<o:p></o:p></P>
<P style="TEXT-INDENT: -0.25in; MARGIN-LEFT: 0.25in; mso-list: l0 level1 lfo1"
class=MsoListParagraph><![if !supportLists]><SPAN
style="mso-list: Ignore">2)<SPAN
style="FONT: 7pt 'Times New Roman'">
</SPAN></SPAN><![endif]>When using mpd and mpiexec, the MV2_DEFAULT_PKEY
parameter gets passed, but then fails. For example:<o:p></o:p></P>
<P class=MsoNormal><o:p> </o:p></P>
<P class=MsoNormal><SPAN style="FONT-FAMILY: 'Courier New'">[root@homer
mpi_apps]# export MV2_DEFAULT_PKEY=0xffff<o:p></o:p></SPAN></P>
<P class=MsoNormal><SPAN style="FONT-FAMILY: 'Courier New'">[root@homer
mpi_apps]# /usr/mpi/gcc/mvapich2-1.2p1/bin/mpiexec -machinefile
/opt/iba/src/mpi_apps/mpi_hosts -n 2 osu2/osu_bw<o:p></o:p></SPAN></P>
<P class=MsoNormal><SPAN style="FONT-FAMILY: 'Courier New'"> [0] Abort:
Can't find PKEY INDEX according to given PKEY<o:p></o:p></SPAN></P>
<P class=MsoNormal><SPAN style="FONT-FAMILY: 'Courier New'"> at line 1190
in file rdma_iba_priv.c<o:p></o:p></SPAN></P>
<P class=MsoNormal><SPAN style="FONT-FAMILY: 'Courier New'">rank 0 in job
6 homer.dev.silverstorm.com_33133 caused collective abort of
all ranks<o:p></o:p></SPAN></P>
<P class=MsoNormal><SPAN style="FONT-FAMILY: 'Courier New'"> exit status
of rank 0: killed by signal 9<o:p></o:p></SPAN></P>
<P class=MsoNormal><o:p> </o:p></P>
<P class=MsoNormal>(Note that 0xffff is actually the default
PKEY).<o:p></o:p></P>
<P class=MsoNormal><o:p> </o:p></P>
<P class=MsoNormal>A quick saquery reveals that the pkey is, in fact in the
table:<o:p></o:p></P>
<P class=MsoNormal><o:p> </o:p></P>
<P class=MsoNormal><SPAN style="FONT-FAMILY: 'Courier New'">[root@homer
mpi_apps]# iba_saquery -o pkey -l 1<o:p></o:p></SPAN></P>
<P class=MsoNormal><SPAN style="FONT-FAMILY: 'Courier New'">LID: 0x0001
PortNum: 1 BlockNum: 0<o:p></o:p></SPAN></P>
<P class=MsoNormal><SPAN
style="FONT-FAMILY: 'Courier New'">
0- 7: 0x9001 0xffff 0x9002 0x0000
0x0000 0x0000 0x0000 0x0000<o:p></o:p></SPAN></P>
<P class=MsoNormal><SPAN
style="FONT-FAMILY: 'Courier New'"> 8-
15: 0x0000 0x0000 0x0000 0x0000 0x0000
0x0000 0x0000 0x0000<o:p></o:p></SPAN></P>
<P class=MsoNormal><SPAN
style="FONT-FAMILY: 'Courier New'"> 16-
23: 0x0000 0x0000 0x0000 0x0000 0x0000
0x0000 0x0000 0x0000<o:p></o:p></SPAN></P>
<P class=MsoNormal><SPAN
style="FONT-FAMILY: 'Courier New'"> 24-
31: 0x0000 0x0000 0x0000 0x0000 0x0000
0x0000 0x0000 0x0000<o:p></o:p></SPAN></P>
<P class=MsoNormal><o:p> </o:p></P>
<P class=MsoNormal>When I examine ibv_param.c to see what was going on, here
is what I found:<o:p></o:p></P>
<P class=MsoNormal><o:p> </o:p></P>
<P class=MsoNormal><SPAN style="FONT-FAMILY: 'Courier New'">
if ((value = getenv("MV2_DEFAULT_PKEY")) != NULL) {<o:p></o:p></SPAN></P>
<P class=MsoNormal><SPAN
style="FONT-FAMILY: 'Courier New'">
rdma_default_pkey = (uint16_t)strtol(value, (char **) NULL,0) &
PKEY_MASK;<o:p></o:p></SPAN></P>
<P class=MsoNormal><SPAN style="FONT-FAMILY: 'Courier New'">
}<o:p></o:p></SPAN></P>
<P class=MsoNormal><SPAN
style="FONT-FAMILY: 'Courier New'">And…<o:p></o:p></SPAN></P>
<P class=MsoNormal><SPAN
style="FONT-FAMILY: 'Courier New'"><o:p> </o:p></SPAN></P>
<P class=MsoNormal><SPAN style="FONT-FAMILY: 'Courier New'">
#define PKEY_MASK 0x7fff /* the last bit is reserved */<o:p></o:p></SPAN></P>
<P class=MsoNormal><SPAN
style="FONT-FAMILY: 'Courier New'"><o:p> </o:p></SPAN></P>
<P class=MsoNormal>This makes it clear that mpiexec is doing bad things to the
pkey – if nothing else, the high bit <B>must</B> be set in order for the
connection to have full membership in an Infiniband partition. Without
setting this bit, a node will only have “limited membership”, and limited
nodes are not permitted to talk to each other.<o:p></o:p></P>
<P class=MsoNormal><o:p> </o:p></P>
<P class=MsoNormal>I’m going to try and see if I can quickly put together a
patch for you that fixes the problems with mpiexec – but I’m not sure what the
correct fix is for mpirun_rsh.<o:p></o:p></P>
<P class=MsoNormal><o:p> </o:p></P>
<P class=MsoNormal><SPAN
style="FONT-FAMILY: 'Arial','sans-serif'; FONT-SIZE: 10pt">--</SPAN><SPAN
style="FONT-FAMILY: 'Times New Roman','serif'; FONT-SIZE: 12pt"><o:p></o:p></SPAN></P>
<P class=MsoNormal><SPAN
style="FONT-FAMILY: 'Arial','sans-serif'; FONT-SIZE: 10pt">Michael
Heinz</SPAN><SPAN
style="FONT-FAMILY: 'Times New Roman','serif'; FONT-SIZE: 12pt"><o:p></o:p></SPAN></P>
<P class=MsoNormal><SPAN
style="FONT-FAMILY: 'Arial','sans-serif'; FONT-SIZE: 10pt">Principal Engineer,
Qlogic Corporation</SPAN><SPAN
style="FONT-FAMILY: 'Times New Roman','serif'; FONT-SIZE: 12pt"><o:p></o:p></SPAN></P>
<P class=MsoNormal><SPAN
style="FONT-FAMILY: 'Arial','sans-serif'; FONT-SIZE: 10pt">King of Prussia,
Pennsylvania</SPAN><o:p></o:p></P></DIV></BLOCKQUOTE></BODY></HTML>