Hi, Roland:
What is the Max # of cards OFED driver/library can support on a
single node ?
Thanks.
--CQ
-Original Message-
From: [EMAIL PROTECTED]
[mailto:[EMAIL PROTECTED] On Behalf Of Roland Dreier
Sent: Tuesday, February 20, 2007 6:12 PM
To: Dotan Barak
Cc:
Thanks for your good point. --CQ
-Original Message-
From: Roland Dreier [mailto:[EMAIL PROTECTED]
Sent: Thursday, February 15, 2007 11:59 AM
To: Tang, Changqing
Cc: Michael S. Tsirkin; openib-general@openib.org
Subject: Re: How heavy to resize a CQ ?
In dynamic process
1. process A and process B is connected with QP. A first
post a send
to B, B does not post receive. Then A and B are doing a long time
RDMA_WRITE each other, A and B just check memory for the RDMA_WRITE
message. Finally B will post a receive. Does the first
pending send in
A
[mailto:[EMAIL PROTECTED]
Sent: Monday, February 05, 2007 5:03 PM
To: Tang, Changqing
Cc: Michael S. Tsirkin; openib-general@openib.org
Subject: Re: Immediate data question
Changqing Thank you. Other than using immediate data to send
Changqing notification from one end to the other
ChangqingDoes this pending SEND_WITH_IMM message
affect the
Changqing performance of the receiver process ? Is this message
Changqing buffered in the receiver's HCA, or the sender retry and
Changqing get RNR ack until receiver posts a receive ?
If no receive is
Roland:
If I only want to send/recv 4 bytes with immediate data:
On sender side:
opcode = IBV_WR_SEND_WITH_IMM;
imm_data = my_4_bytes_data;
Do I still need to specify sg_list and num_sge ?
On receiver side, because the immediate data is inside the completion
Message-
From: Roland Dreier [mailto:[EMAIL PROTECTED]
Sent: Monday, February 05, 2007 4:09 PM
To: Tang, Changqing
Cc: Michael S. Tsirkin; openib-general@openib.org
Subject: Re: Immediate data question
If I only want to send/recv 4 bytes with immediate data:
I assume you mean
I have trouble to get the latest OFED 1.2 source because of the firewall
or other reason I don't know. Is there a checkout version somewhere I
can access, for example. Daily build version ?
Thanks.
--CQ
-Original Message-
From: [EMAIL PROTECTED]
[mailto:[EMAIL PROTECTED] On Behalf
PM
To: Tang, Changqing
Cc: Tziporet Koren; EWG; OPENIB
Subject: Re: [openib-general] Minutes for January 15, 2007
teleconference about OFED 1.2 development progress toward code freeze
We need this compatible feature very much so I hope you
double confirm this. Can app build with 1.0
when RDMA is used, a message is transferred from card A (in node
A) to card B (in node B), card B delivers the message to
to user buffer, and sends ACK to card A, but ACK is lost
due to switch fail. So process on node A get fail for this
transfer, but process on node B check the
-Original Message-
From: [EMAIL PROTECTED]
[mailto:[EMAIL PROTECTED] On Behalf Of Roland Dreier
Sent: Tuesday, January 16, 2007 1:11 PM
To: Tziporet Koren
Cc: EWG; OPENIB
Subject: Re: [openib-general] Minutes for January 15, 2007
teleconference about OFED 1.2 development
+ libibverbs 1.0 to be included as well
I believe I have a way to make libibverbs 1.1
binary-compatible with applications (but not drivers) built
against libibverbs 1.0. So it will probably not be necessary
to include libibverbs 1.0.
We need this compatible feature
Where do you put these new user functions ? Do you create a new library
? I hope not to create a new library, we already have so many libraries
now, it is hard to manage
For users using dlopen().
Also when is it available ?
Thanks.
--CQ
-Original Message-
From: [EMAIL PROTECTED]
: Or Gerlitz [mailto:[EMAIL PROTECTED]
Sent: Monday, January 08, 2007 1:18 AM
To: Tang, Changqing
Cc: openib-general@openib.org
Subject: using IB on a port without IPoIB running NIC
Tang, Changqing wrote:
We understand that, but we hope to have a connect/accept style IB
connection setup, without
: Monday, January 08, 2007 8:18 AM
To: Tang, Changqing
Cc: Or Gerlitz; openib-general@openib.org
Subject: Re: using IB on a port without IPoIB running NIC
Thank you for the information, I may change my mind to
require IPoIB
to run newer version of HP-MPI on OFED 1.2, if I don't find
What I need is that, without IPoIB, how do I wire IB connection ?
Currently with Verbs API, it is an alltoall QP number
exchange. I want
to remove the alltoall QP number exchange in MPI dynamic process.
Well, does your MPI implementation currently use librdmacm?
No, we don't use both
As I said, the problem is the alltoall QP number exchange.
I hope that
a process can only provide one piece of information(such as
ip/port in
TCP/IP) so that all other processes have the same piece of info and
can make connection to it.
Well, start with a socket, each time a
How do you know that the old path is back ? Do you have a notification
handler called, or you need to query periodically ?
Thanks.
--CQ
another question: If one brings back the old path (after the first
failure) and use the old path record
to do lap/apr then reenabling migration using
? If path 1 is down, we can not set
it as alternate path, right ?
If we can not bring a path back on fly, the usage of APM is limited.
--CQ
-Original Message-
From: Sean Hefty [mailto:[EMAIL PROTECTED]
Sent: Wednesday, October 25, 2006 11:41 AM
To: Tang, Changqing
Cc: [EMAIL PROTECTED
Has the registration site been set up ?
--CQ
-Original Message-
From: [EMAIL PROTECTED]
[mailto:[EMAIL PROTECTED] On Behalf Of Jeff Squyres
Sent: Tuesday, October 17, 2006 6:57 AM
To: Bill Boas
Cc: Open Fabrics; openib-general@openib.org; [EMAIL PROTECTED]
Subject: Re:
What kernel are you testing? system() should work (in non threaded
apps at
least) with modern kernel.
--
Gleb.
___
From the OFED release notes:
3. Fork support from kernel 2.6.12 and above is available provided
3. Fork support from kernel 2.6.12 and above is available
provided that applications do not use threads. The fork() is
supported as long as parent process does not run before child
exits or calls exec().
After fork(), in child, before exec(), can we call printf(), putenv(),
or even re-direct
Thanks for the clarification. --CQ
You need to make a difference between full fork support that
will be available only in libibverbs1.1 and the system /fork
exec fork support that is depend on the kernel only and
available from kernel 2.6.12.
See also the explanation from Gleb on this
We tested RC7, but fork() does not work:
1. system() causes IB to fail.
2. fork(), child calling exit(0) immediately also causes IB to fail.
Anyone has tested fork() related issue ?
--CQ Tang, HP-MPI
-Original Message-
From: [EMAIL PROTECTED]
[mailto:[EMAIL PROTECTED] On Behalf Of
No, a PD belongs to a specific device. However nothing
prevents you from creating one PD for each device, and two MRs
(one for each device, each using one of those two PDs) that
cover the same memory.
Roland:
I did exactly what you said with two cards on a node, however,
if I use the
'retries exceeded' means that the transport retry count was
exceeded, so most likely your timeout is set too low.
Is there a common recommended value for this timeout ? I use 18, which
represents 1 second.
Without seeing your code, I couldn't begin to say why you
don't see a send
Changqing Is there a common recommended value for this timeout ?
Changqing I use 18, which represents 1 second.
18 should be OK I guess, unless you have congestion in your
fabric, in which case you have other problems anyway.
Changqing It is very hard to reproduce this error with
structure, the other copy all data to provided
dest buffer.
--CQ
-Original Message-
From: Ralph Campbell [mailto:[EMAIL PROTECTED]
Sent: Wednesday, August 23, 2006 1:16 PM
To: Tang, Changqing
Cc: Caitlin Bestler; openib-general@openib.org
Subject: RE: [openib-general] basic IB doubt
28 matches
Mail list logo
