From ca371a4fa83941c2a8e7f4d114e8b6e63994dd0f Mon Sep 17 00:00:00 2001
From: chun9l <97897047+chun9l@users.noreply.github.com>
Date: Tue, 21 Jul 2026 23:31:58 +0100
Subject: [PATCH 1/3] Add early draft for shmem docs

---
 docs/source/shared_memory.rst | 22 ++++++++++++++++++++++
 1 file changed, 22 insertions(+)
 create mode 100644 docs/source/shared_memory.rst

diff --git a/docs/source/shared_memory.rst b/docs/source/shared_memory.rst
new file mode 100644
index 0000000..bb1349c
--- /dev/null
+++ b/docs/source/shared_memory.rst
@@ -0,0 +1,22 @@
+.. _shared-memory:
+=============================================
+Client-Server Communication via Shared Memory
+=============================================
+
+UM-Bridge uses HTTP to communicate between a UQ client and a model server
+by default. However, this is not the most performant option (in HPC) as it 
+carries around 1 millisecond of overhead per request. We provide a quicker
+method that uses shared memory (RAM) to transfer data.
+
+Currently, only the C++ and Python implementations are supported; instructions 
+are documented in the respective sections below. When enabled, UM-Bridge will 
+first perform a data exchange test, and will fallback to HTTP if this fails.
+
+
+Enabling Shared Memory Communication in C++
+===========================================
+
+
+
+Enabling Shared Memory Communication in Python
+==============================================
\ No newline at end of file

From 1411e02ca158a70d052900d81a18c148a35908bb Mon Sep 17 00:00:00 2001
From: chun9l <97897047+chun9l@users.noreply.github.com>
Date: Thu, 23 Jul 2026 17:39:32 +0100
Subject: [PATCH 2/3] add more text on shmem

---
 docs/source/shared_memory.rst | 25 +++++++++++--------------
 1 file changed, 11 insertions(+), 14 deletions(-)

diff --git a/docs/source/shared_memory.rst b/docs/source/shared_memory.rst
index bb1349c..3cf3c4d 100644
--- a/docs/source/shared_memory.rst
+++ b/docs/source/shared_memory.rst
@@ -3,20 +3,17 @@
 Client-Server Communication via Shared Memory
 =============================================
 
-UM-Bridge uses HTTP to communicate between a UQ client and a model server
+UM-Bridge uses HTTP to communicate between an UQ client and a model server
 by default. However, this is not the most performant option (in HPC) as it 
 carries around 1 millisecond of overhead per request. We provide a quicker
-method that uses shared memory (RAM) to transfer data.
+alternative that uses shared memory (RAM) to transfer data. This approach 
+works by using allocated memory buffers in RAM as the transfer medium for 
+UM-Bridge inputs and outputs. To prevent data races in a parallel setting, 
+each memory buffer is tagged with the ID of the thread making the request 
+in the client, and this ID is also shared (through a JSON entry) so that 
+the model can access the correct buffer. 
 
-Currently, only the C++ and Python implementations are supported; instructions 
-are documented in the respective sections below. When enabled, UM-Bridge will 
-first perform a data exchange test, and will fallback to HTTP if this fails.
-
-
-Enabling Shared Memory Communication in C++
-===========================================
-
-
-
-Enabling Shared Memory Communication in Python
-==============================================
\ No newline at end of file
+Currently, only the C++ and Python implementations are supported, and this
+feature is enabled automatically. The code will first perform a test 
+transfer using shared memory when the client connects to the server; it 
+reverts back to HTTP if this test fails.

From ebb605828a9c2580fcb8f6eb7e374e4218d7b0ca Mon Sep 17 00:00:00 2001
From: chun9l <97897047+chun9l@users.noreply.github.com>
Date: Fri, 24 Jul 2026 16:25:26 +0100
Subject: [PATCH 3/3] update docs with shmem

---
 docs/generate_documentation.sh | 3 +++
 docs/source/shared_memory.rst  | 9 +++++----
 2 files changed, 8 insertions(+), 4 deletions(-)

diff --git a/docs/generate_documentation.sh b/docs/generate_documentation.sh
index 361a498..d373331 100755
--- a/docs/generate_documentation.sh
+++ b/docs/generate_documentation.sh
@@ -131,6 +131,9 @@ for f in $(find ../../umbridge/ -name 'README.md'); do
    $NAME
 EOF
 done
+
+# Extra docs. Added manually
+echo "   ../shared_memory" >> docs_output/source/umbridge/index.rst
 echo "   ../mpi_container" >> docs_output/source/umbridge/index.rst
 echo "   ../math_description" >> docs_output/source/umbridge/index.rst
 
diff --git a/docs/source/shared_memory.rst b/docs/source/shared_memory.rst
index 3cf3c4d..1cc9a2e 100644
--- a/docs/source/shared_memory.rst
+++ b/docs/source/shared_memory.rst
@@ -1,4 +1,5 @@
 .. _shared-memory:
+
 =============================================
 Client-Server Communication via Shared Memory
 =============================================
@@ -10,10 +11,10 @@ alternative that uses shared memory (RAM) to transfer data. This approach
 works by using allocated memory buffers in RAM as the transfer medium for 
 UM-Bridge inputs and outputs. To prevent data races in a parallel setting, 
 each memory buffer is tagged with the ID of the thread making the request 
-in the client, and this ID is also shared (through a JSON entry) so that 
+in the parallel client, and this ID is shared through a JSON entry so that 
 the model can access the correct buffer. 
 
 Currently, only the C++ and Python implementations are supported, and this
-feature is enabled automatically. The code will first perform a test 
-transfer using shared memory when the client connects to the server; it 
-reverts back to HTTP if this test fails.
+feature is enabled automatically when available. The code will first 
+perform a test transfer using shared memory when the client first connects 
+to the server; it reverts back to HTTP if this test fails.