As a design point, I strongly prefer not to play optimization games for the embedded crowd, for a number of reasons: 1. When I worked in embedded systems, we...
Sears, Steven
sjs@...
Mar 1, 2004 1:06 pm
2758
... There is always the famous specmanship that P.J. Plauger used for some of the standard C library -- that the library could implement the following function...
Hello, This email message is a notification to let you know that a file has been uploaded to the Files area of the dat-discussions group. File :...
dat-discussions@yahoo...
Mar 3, 2004 2:10 pm
2760
We will have a DAT Collaborative conference call this Wednesday March 3, 1:00-2:00pm EDT (10:00am-11:00am PDT). Moderator: Arkady Kanevsky Phone: 888-827-8686 ...
Hello, This email message is a notification to let you know that a file has been uploaded to the Files area of the dat-discussions group. File :...
dat-discussions@yahoo...
Mar 3, 2004 4:49 pm
2762
Hello, This email message is a notification to let you know that a file has been uploaded to the Files area of the dat-discussions group. File :...
dat-discussions@yahoo...
Mar 3, 2004 4:49 pm
2763
The SRQ APIs (kdapl and udapl) are part of DAPL-1.2 has been uploaded to DAT reflector (files) today. The formal vote on it will take place 2 weeks from now on...
We will have a DAT Collaborative conference call this Wednesday March 10, 1:00-2:00pm EDT (10:00am-11:00am PDT). Moderator: Arkady Kanevsky Phone: 888-827-8686...
The vote for non coherent memory is scheduled for 3/24. Please, cast you vote (and comments) prior to that on reflector or at the call. Arkady Kanevsky...
Sherman, I have a couple of comments. 1. Do we really need 2 separate calls or single call is sufficient? The call just sync lmr_triplet array. What it is used...
... The dat_lmr_sync_rdma_read() is to invalidate the I/O cache before an incoming RDMA read so that remote client could pick up the modified data in RAM. On...
Sherman Pun
sherman.pun@...
Mar 11, 2004 6:59 pm
2771
... I agree that both "polarities" are required. Write is a "flush", Read is an "invalidate". However, we might consider a flag word to a single API rather...
... dat-discussions-owner@yahoogroups.com ... Stipulated, in fact there is a single Solaris DDI call which takes a "direction" flag. Since ITAPI has this...
Sherman Pun
sherman.pun@...
Mar 16, 2004 12:00 am
2773
I am traveling and will not be able to make it. Caitlin told us last week that she will not be able to make it. Next call is March 24 with votes on SRQ and...
... From: Sears, Steven Sent: Thursday, February 26, 2004 4:16 PM To: dat-discussions@yahoogroups.com Subject: [dat-discussions] Handle Proposal: request for...
Sears, Steven
sjs@...
Mar 17, 2004 4:58 pm
2775
... Steven, Are these header changes planned for uDAPL 1.2 ? The reason I'm asking is that, as you may know, we have uDAPL 1.1 support in our MPI software and...
These are the header files planned for version 1.2. As you are probably aware, most of the reference implementation is currently at 1.2, but the majority of...
Sears, Steven
sjs@...
Mar 17, 2004 8:30 pm
2777
Hello, This email message is a notification to let you know that a file has been uploaded to the Files area of the dat-discussions group. File :...
dat-discussions@yahoo...
Mar 18, 2004 6:36 pm
2778
Hello, This email message is a notification to let you know that a file has been uploaded to the Files area of the dat-discussions group. File :...
dat-discussions@yahoo...
Mar 18, 2004 6:37 pm
2779
... you may know, we have uDAPL 1.1 support in our MPI software and since these header changes isn't ... header structure. I guess this really depends on when...
... Thanks for clearing this up for me Caitlin. If I understand correctly, our MPI which is compiled with the old redirection mechanism (macros in the header...
... You correctly understand the proposed changes. It will result in a small change for providers, in that a new include files will be added where there is...
Sears, Steven
sjs@...
Mar 22, 2004 1:48 am
2782
... The typedefs that support DAT_PROVIDER are indeed the key, and since they are used only by the re-direction code and the initializer of the DAT_PROVIDER...
The following is proposed to tighten our cleanup semantics for multithreaded applications. I basically says that if something bad happens to a thread in the...
... Taking a minimalistic interpretation of this, it essentially says that a Provider should be organized so that it can close all the objects for an IA even...
There is an error in the header file for old (1.1) kdat_vendor_specific.h file. For the DAT_PROVIDER_ATTR_MASK DAT_PROVIDER_FIELD_EP_CREATOR = 0x01000, ...
Here is the list of errata items for DAPL-1.2 and proposed spec changes for them: Errata # proposed change 73 Spec 1.1 05/08, page 163 (or pg 173 in pdf), ...
Having problems with message search?
Fill out this form to ensure your group is one of the first to be migrated to the new message search system.
