[mvapich-discuss] (no subject)

Jonathan Perkins perkinjo at cse.ohio-state.edu
Wed Feb 17 19:34:04 EST 2016


Hi Justin.  The problem has been resolved on our end.  Can you try again?

On Wed, Feb 17, 2016 at 4:45 PM Jonathan Perkins <
perkinjo at cse.ohio-state.edu> wrote:

> That's weird, I'm seeing this too.  I'll check on this and let you know
> once its resolved.
>
> On Wed, Feb 17, 2016 at 3:34 PM Justin Luitjens <jluitjens at nvidia.com>
> wrote:
>
>> Sure but can you check the link.  I’m getting a 404 not found at that
>> link.
>>
>>
>>
>> *From:* Jonathan Perkins [mailto:perkinjo at cse.ohio-state.edu]
>> *Sent:* Wednesday, February 17, 2016 12:34 PM
>> *To:* Justin Luitjens; mvapich-discuss at cse.ohio-state.edu
>> *Subject:* Re:
>>
>
>>
>> Hello Justin.  The package that you're using requires an HCA to use the
>> GDR features.  We do provide a package that does not require an HCA for all
>> the GDR features but unfortunately it is only available for rhel7 at this
>> time.  Can you try using this package with your system?  If it works we'll
>> update the site to point out that it can be used with rhel6 as well
>> (rhel/centos).
>>
>>
>> http://mvapich.cse.ohio-state.edu/download/mvapich/gdr/2.2b/mvapich2-gdr-regular-ofed-2.2b-cuda-7.5.tar.gz
>>
>> On Wed, Feb 17, 2016 at 1:21 PM Justin Luitjens <jluitjens at nvidia.com>
>> wrote:
>>
>> --===============4387389260387728233==
>> Content-Language: en-US
>> Content-Type: multipart/alternative;
>>
>> boundary="_000_631ba403f2354450936ae0e29a0e7365HQMAIL111nvidiacom_"
>>
>> --_000_631ba403f2354450936ae0e29a0e7365HQMAIL111nvidiacom_
>> Content-Type: text/plain; charset="us-ascii"
>> Content-Transfer-Encoding: quoted-printable
>>
>> I'm seeing an odd issue with mvapich2.2b-gdr.  I have installed the RPM p=
>> ackage on CentOS 6.4 and I have also installed the GDRCopy library.
>>
>> I have set the MV2_GPUDIRECT_GDRCOPY_LIB variable and the code runs fine =
>> (I'm not using GPUDirect at the moment).
>>
>> However, as soon as I set the environment variable MV2_USE_CUDA=3D1 my pr=
>> ocess hangs at MPI_Init.
>>
>> For reference I'm only running on a single node which does not have an in=
>> finiband nic on it.  I only want to run on a single node and never need t=
>> o go cross-node.
>>
>> Anyone have a work around for this or know what might be going wrong?
>>
>> Thanks,
>> Justin
>>
>>
>> -------------------------------------------------------------------------=
>> ----------
>> This email message is for the sole use of the intended recipient(s) and m=
>> ay contain
>> confidential information.  Any unauthorized review, use, disclosure or di=
>> stribution
>> is prohibited.  If you are not the intended recipient, please contact the=
>> =20sender by
>> reply email and destroy all copies of the original message.
>> -------------------------------------------------------------------------=
>> ----------
>>
>> --_000_631ba403f2354450936ae0e29a0e7365HQMAIL111nvidiacom_
>> Content-Type: text/html; charset="us-ascii"
>> Content-Transfer-Encoding: quoted-printable
>>
>> <html xmlns:v=3D"urn:schemas-microsoft-com:vml" xmlns:o=3D"urn:schemas-mi=
>> crosoft-com:office:office" xmlns:w=3D"urn:schemas-microsoft-com:office:wo=
>> rd" xmlns:m=3D"http://schemas.microsoft.com/office/2004/12/omml"
>> xmlns=3D=
>> "http://www.w3.org/TR/REC-html40">
>> <head>
>> <meta http-equiv=3D"Content-Type" content=3D"text/html; charset=3Dus-asci=
>> i">
>> <meta name=3D"Generator" content=3D"Microsoft Word 15 (filtered medium)">=
>>
>> <style><!--
>> /* Font Definitions */
>> @font-face
>> =09{font-family:"Cambria Math";
>> =09panose-1:2 4 5 3 5 4 6 3 2 4;}
>> @font-face
>> =09{font-family:Calibri;
>> =09panose-1:2 15 5 2 2 2 4 3 2 4;}
>> /* Style Definitions */
>> p.MsoNormal, li.MsoNormal, div.MsoNormal
>> =09{margin:0in;
>> =09margin-bottom:.0001pt;
>> =09font-size:11.0pt;
>> =09font-family:"Calibri",sans-serif;}
>> a:link, span.MsoHyperlink
>> =09{mso-style-priority:99;
>> =09color:#0563C1;
>> =09text-decoration:underline;}
>> a:visited, span.MsoHyperlinkFollowed
>> =09{mso-style-priority:99;
>> =09color:#954F72;
>> =09text-decoration:underline;}
>> span.EmailStyle17
>> =09{mso-style-type:personal-compose;
>> =09font-family:"Calibri",sans-serif;
>> =09color:windowtext;}
>> .MsoChpDefault
>> =09{mso-style-type:export-only;
>> =09font-family:"Calibri",sans-serif;}
>> @page WordSection1
>> =09{size:8.5in 11.0in;
>> =09margin:1.0in 1.0in 1.0in 1.0in;}
>> div.WordSection1
>> =09{page:WordSection1;}
>> --></style><!--[if gte mso 9]><xml>
>> <o:shapedefaults v:ext=3D"edit" spidmax=3D"1026" />
>> </xml><![endif]--><!--[if gte mso 9]><xml>
>> <o:shapelayout v:ext=3D"edit">
>> <o:idmap v:ext=3D"edit" data=3D"1" />
>> </o:shapelayout></xml><![endif]-->
>> </head>
>> <body lang=3D"EN-US" link=3D"#0563C1" vlink=3D"#954F72">
>> <div class=3D"WordSection1">
>> <p class=3D"MsoNormal">I’m seeing an odd issue with mvapich2.2b-gdr=
>> .  I have installed the RPM package on CentOS 6.4 and I have also in=
>> stalled the GDRCopy library.
>> <o:p></o:p></p>
>> <p class=3D"MsoNormal"><o:p> </o:p></p>
>> <p class=3D"MsoNormal">I have set the MV2_GPUDIRECT_GDRCOPY_LIB variable =
>> and the code runs fine (I’m not using GPUDirect at the moment).<o:p=
>> ></o:p></p>
>> <p class=3D"MsoNormal"><o:p> </o:p></p>
>> <p class=3D"MsoNormal">However, as soon as I set the environment variable=
>> =20MV2_USE_CUDA=3D1 my process hangs at MPI_Init.<o:p></o:p></p>
>> <p class=3D"MsoNormal"><o:p> </o:p></p>
>> <p class=3D"MsoNormal">For reference I’m only running on a single n=
>> ode which does not have an infiniband nic on it.  I only want to run=
>> =20on a single node and never need to go cross-node.<o:p></o:p></p>
>> <p class=3D"MsoNormal"><o:p> </o:p></p>
>> <p class=3D"MsoNormal">Anyone have a work around for this or know what mi=
>> ght be going wrong?<o:p></o:p></p>
>> <p class=3D"MsoNormal"><o:p> </o:p></p>
>> <p class=3D"MsoNormal">Thanks,<o:p></o:p></p>
>> <p class=3D"MsoNormal">Justin<o:p></o:p></p>
>> <p class=3D"MsoNormal"><o:p> </o:p></p>
>> </div>
>>
>> <DIV>
>> <HR>
>> </DIV>
>> <DIV>This email message is for the sole use of the intended recipient(s) =
>> and may=20
>> contain confidential information.  Any unauthorized review, use, dis=
>> closure=20
>> or distribution is prohibited.  If you are not the intended recipien=
>> t,=20
>> please contact the sender by reply email and destroy all copies of the or=
>> iginal=20
>> message. </DIV>
>> <DIV>
>> <HR>
>> </DIV>
>> </body>
>> </html>
>>
>> --_000_631ba403f2354450936ae0e29a0e7365HQMAIL111nvidiacom_--
>>
>> --===============4387389260387728233==
>> Content-Type: text/plain; charset="us-ascii"
>> MIME-Version: 1.0
>> Content-Transfer-Encoding: 7bit
>> Content-Disposition: inline
>>
>> _______________________________________________
>> mvapich-discuss mailing list
>> mvapich-discuss at cse.ohio-state.edu
>> http://mailman.cse.ohio-state.edu/mailman/listinfo/mvapich-discuss
>>
>> --===============4387389260387728233==--
>>
>> _______________________________________________
> mvapich-discuss mailing list
> mvapich-discuss at cse.ohio-state.edu
> http://mailman.cse.ohio-state.edu/mailman/listinfo/mvapich-discuss
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mailman.cse.ohio-state.edu/pipermail/mvapich-discuss/attachments/20160218/66e10ca8/attachment-0001.html>


More information about the mvapich-discuss mailing list