docs: fix some misspellings

Summary: Closes facebookincubator#87 Differential Revision: D5835831 Pulled By: pietern fbshipit-source-id: 68d315a575dc1c1f79e50777b4e317223f5c7da3
H-Huang · Sep 14, 2017 · d7d8d2c · d7d8d2c
1 parent 5f64820
commit d7d8d2c
Show file tree

Hide file tree

Showing 3 changed files with 4 additions and 4 deletions.
diff --git a/docs/algorithms.md b/docs/algorithms.md
@@ -100,7 +100,7 @@ at each step. At each process and step, the portion of the buffer that was being
 sent in the reduce-scatter is received in the allgather, and the portion that was
 being received in the reduce-scatter is now sent.
 
-Across the steps of the reduce-scatter, data is received intto different buffers
+Across the steps of the reduce-scatter, data is received into different buffers
 and there is no potential for race conditions. However, mirrored steps of the
 reduce-scatter and allgather (e.g. last step of the reduce-scatter and first
 step of the allgather) write into the same buffers. To prevent race conditions,

diff --git a/docs/latency.md b/docs/latency.md
@@ -31,7 +31,7 @@ In no particular order:
 
 Make sure it is enabled if your NIC supports it. For high bandwidth
 NICs, this is absolutely necessary to achieve line rate on a single
-connection (some anecdotal evindence: 10Gb/s without TSO at 100% CPU
+connection (some anecdotal evidence: 10Gb/s without TSO at 100% CPU
 usage versus 40Gb/s (line rate) with TSO at 30% CPU usage).
 
 ```

diff --git a/docs/rendezvous.md b/docs/rendezvous.md
@@ -58,14 +58,14 @@ rendezvous to a particular namespace.
 
 There are many more key/value stores that can be used for rendezvous
 (e.g. [etcd](https://coreos.com/etcd) or [ZooKeeper](https://zookeeper.apache.org/)).
-As long as a C or C++ interface for your store of choice is availabe,
+As long as a C or C++ interface for your store of choice is available,
 is relatively easy to hook it up to the Gloo rendezvous process.
 See the `gloo::rendezvous::Store` abstract base class for the interface to implement.
 
 ## Using MPI
 
 If you are already using MPI to run jobs across machines, getting started with
-Gloo should be staightforward. Instead of using a separate key/value store for
+Gloo should be straightforward. Instead of using a separate key/value store for
 rendezvous, the existing MPI communicator is used to create contexts.
 Make sure to compile Gloo with `USE_MPI=ON`.