[openfabrics-ewg] Test results so far from Cisco for OFED 1.0 RC4

Jeff Squyres (jsquyres) jsquyres at cisco.com
Mon May 15 20:30:41 PDT 2006


59: fixed, awaiting Open MPI 1.1a5 (by early tomorrow)
74: still under investigation
80: still under investigation
53: still under investigation
 
There were some others that were also closed.
 
I keep delaying replying on these because they're under active
investigation and testing, but such is the life of an alpha version.
1.1a5 should be available tomorrow morning; I will commit a new SRPM
first thing in the morning.
 


________________________________

	From: openfabrics-ewg-bounces at openib.org
[mailto:openfabrics-ewg-bounces at openib.org] On Behalf Of Tziporet Koren
	Sent: Monday, May 15, 2006 9:37 AM
	To: Scott Weitzenkamp (sweitzen); openfabrics-ewg at openib.org
	Subject: RE: [openfabrics-ewg] Test results so far from Cisco
for OFED 1.0 RC4
	
	

	Update for RC5:

	Bugs:

	49 - fixed

	51 - fixed

	57 - fixed

	62 - on work (FMR are in, still wait that Roland will check-in
the tunable parameters patches )

	81 - on work (Pasha fro  Mellanox)

	 

	Jeff - please update on status of Open MPI bugs

	 

	DSP: 

	The code started to pass tests with data movement. We still see
failures thus we focus on debug and fix issues.

	We will not do RC5 till SDP is usable in the main flows.

	 

	Tziporet

	 

	 

	 

	 

	-----Original Message-----
	From: openfabrics-ewg-bounces at openib.org
[mailto:openfabrics-ewg-bounces at openib.org] On Behalf Of Scott
Weitzenkamp (sweitzen)
	Sent: Thursday, May 11, 2006 11:41 PM
	To: openfabrics-ewg at openib.org
	Subject: [openfabrics-ewg] Test results so far from Cisco for
OFED 1.0 RC4

	 

	Cisco has tested IPoIB, SDP, MPI (MVAPICH 0.9.7, Open MPI 1.1a3,
and Intel MPI 2.0.1), and uDAPL on RHEL4 U3 i686, x86_64, ia64, and
ppc64.  We haved tested all the Mellanox SDR HCAs we support: Cougar,
CougarCub, LionCub, LionMini, and Cheetah.

	 

	For MPI, we have run OSU and Pallas benchmarks, next up is NPB.
I've enclosed MPI perf data.

	 

	Many bugs were opened, I consider the following still-open bugs
to be must-fix for RC5 (plus overhauled SDP):

	 

49 <http://openib.org/bugzilla/show_bug.cgi?id=49>  

2006-05-03 

nor 

P2 

All 

huanwei at cse.ohio-state.edu 

NEW 

 

OFED 1.0: MVAPICH won't compile on ppc64 

51 <http://openib.org/bugzilla/show_bug.cgi?id=51>  

2006-05-05 

nor 

P2 

All 

tziporet at mellanox.co.il 

NEW 

 

OFED 1.0 rc4: SRP not available for RHEL4 U3 

57 <http://openib.org/bugzilla/show_bug.cgi?id=57>  

Fri 14:01 

nor 

P2 

All 

vlad at mellanox.co.il 

NEW 

 

OFED 1.0 rc4: rdma_cm does not work for uDAPL 

59 <http://openib.org/bugzilla/show_bug.cgi?id=59>  

Fri 15:05 

nor 

P2 

All 

jsquyres at cisco.com 

NEW 

 

OFED 1.0: Open MPI not configured correctly to find shlibs 

62 <http://openib.org/bugzilla/show_bug.cgi?id=62>  

Mon 04:09 

maj 

P2 

All 

rolandd at cisco.com 

NEW 

 

OFED 1.0 rc4: too many SRP patches, get this code checked in 

64 <http://openib.org/bugzilla/show_bug.cgi?id=64>  

Mon 00:36 

maj 

P2 

All 

jsquyres at cisco.com 

NEW 

 

OFED 1.0 rc4: Open MPI fails when host has more than one ... 

74 <http://openib.org/bugzilla/show_bug.cgi?id=74>  

Wed 11:59 

nor 

P2 

All 

jsquyres at cisco.com 

NEW 

 

OFED 1.0 rc4: Open MPI Pallas test hangs 

80 <http://openib.org/bugzilla/show_bug.cgi?id=80>  

09:35:02 

nor 

P2 

PPC 

jsquyres at cisco.com 

NEW 

 

OFED 1.0 rc4: Open MPI fails on RHEL4 U3 ppc64 

81 <http://openib.org/bugzilla/show_bug.cgi?id=81>  

10:55:49 

maj 

P2 

All 

huanwei at cse.ohio-state.edu 

NEW 

 

OFED 1.0 rc4: MVAPICH max throughput on PCI-X Mellanox HC... 

	 

	Jeff Squyers has asked me to try Open MPI 1.1a4, it should have
fixes for some of the Open MPI bugs.

	 

	Scott Weitzenkamp

	SQA and Release Manager

	Server Virtualization Business Unit

	Cisco Systems

	 

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.openfabrics.org/pipermail/ewg/attachments/20060515/cd330497/attachment.html>


More information about the ewg mailing list