new

diff --git a/meta b/meta
index 5de2df84c5da4385d16101760b3ad977f9de5089..bfd9d5aba0ac812a64101d92ecc249e42c747c4d 100644 (file)
--- a/meta
+++ b/meta
@@ -1,6 +1,6 @@
 Version: 1
-Previous: 338bfebceffcc75bd1dc021319f1d89d2e19448d
-Head: 4c75264faedfa3aa871d7a2326f268fc8e34cd14
+Previous: a6eecc16e7021daf7f3f088de328bf833795ffc0
+Head: 96d204753f2d1a48cb8ee01a32ab09ae411cb03f
 Applied:
   destroy_cqs: 66410c265b2661e16b7f6126bcc4ea5c7e1af1bd
   conn-err: ac46465ae6061b06f005ba9b49b371c9eb0d2325
@@ -12,5 +12,6 @@ Applied:
   rstream-bufsize: 388f3bd699ec0d761a5c1bbf0ee0a1b1d4dff3d2
   rs-def-qpsize: 2fa74d987f14bf98bc599287a3eea8e008275f34
   rs-1sge: 4c75264faedfa3aa871d7a2326f268fc8e34cd14
+  rs-def-inline: 96d204753f2d1a48cb8ee01a32ab09ae411cb03f
 Unapplied:
 Hidden:

diff --git a/patches/rs-def-inline b/patches/rs-def-inline
new file mode 100644 (file)
index 0000000..72ff952
--- /dev/null
+++ b/patches/rs-def-inline
@@ -0,0 +1,28 @@
+Bottom: caec06cc45592384f2a93faf3ca4f4571b1481ff
+Top:    caec06cc45592384f2a93faf3ca4f4571b1481ff
+Author: Sean Hefty <sean.hefty@intel.com>
+Date:   2012-05-24 18:18:43 -0700
+
+rsockets: Reduce the default inline size
+
+Inline data consumes the same space used by the SGL.  Since
+we reduced the default number of SGEs per SQ entry to 1,
+also reduce the default inline data size to 16 bytes.
+Otherwise, the SQ size won't actually be reduced.
+
+Although this increases the latency of small messages over
+16 bytes, tests show that decreasing the inline data size
+from 64 bytes to 32 or 16 bytes improves large message
+bandwidth 8-10%.
+
+Sample rstream 64k_bw test results:
+
+inline size 64: 24.85 Gbps
+inline size 16: 26.5  Gbps
+
+Signed-off-by: Sean Hefty <sean.hefty@intel.com>
+
+
+---
+
+