update intro with SSNI equation, fix references and spelling

lanl · Sep 15, 2023 · 03a108d · 03a108d
1 parent 1a92ee1
commit 03a108d
Show file tree

Hide file tree

Showing 4 changed files with 28 additions and 14 deletions.
diff --git a/doc/sphinx/0_intro/introduction.rst b/doc/sphinx/0_intro/introduction.rst
@@ -216,11 +216,25 @@ Optional results:
 
 Scaled Single Node Improvement
 ==============================
-One element of evaluation will focus on scaled single node improvement: 
+One element of evaluation will focus on scaled single node improvement (SSNI). SSNI is defined as follows: 
+
+Given two platforms using one as a reference, SSNI is defined as a weighted geometric mean using the following equation. 
+
+.. math::
+
+   SSNI = N(\prod_{i=1}^{M}(S_i)^{w_i})^\frac{1}{\sum_{i=1}^{M}{W_i}}
+
+
+Where: 
+
+*	N = Number of nodes on ATS-5 system / Number of nodes on reference system,
+
+*	M = total number of Benchmarks,
+
+*	S = application speedup; Figure of Merit on ATS-5 system / Figure of Merit on reference system (Crossroads); S must be greater than 1, 
+
+*	w = weighting factor. 
 
-* Weighted average of Single node performance improvement * number of nodes  
-* Multiple node configurations can be proposed and may vary on single node performance improvement and number of nodes 
-* Baseline will be SSNI of Crossroads SPR-HBM 
 
 
 Approvals

diff --git a/doc/sphinx/1_branson/branson.rst b/doc/sphinx/1_branson/branson.rst
@@ -9,7 +9,7 @@ This is the documentation for the ATS-5 Benchmark Branson - 3D hohlraum single n
 Purpose
 =======
 
-From their [site]_:
+From their [Branson]_:
 
 Branson is not an acronym.
 
@@ -191,4 +191,4 @@ Verification of Results
 References
 ==========
 
-.. [site] Alex R. Long, 'Branson', 2023. [Online]. Available: https://github.com/lanl/branson. [Accessed: 22- Feb- 2023]
+.. [Branson] Alex R. Long, 'Branson', 2023. [Online]. Available: https://github.com/lanl/branson. [Accessed: 22- Feb- 2023]
diff --git a/doc/sphinx/3_vibe/vibe.rst b/doc/sphinx/3_vibe/vibe.rst
@@ -7,7 +7,7 @@ This is the documentation for the ATS-5 Benchmark, Parthenon-VIBE.
 Purpose
 =======
 
-The Parthenon-VIBE benchmark [site]_ solves the Vector Inviscid Burgers' Equation on a block-AMR mesh.
+The Parthenon-VIBE benchmark [Parthenon-VIBE]_ solves the Vector Inviscid Burgers' Equation on a block-AMR mesh.
 This benchmark is configured to use three levels of mesh resolution and mesh blocks of size 16^3. This AMR configuration is meant to
 mimic applications which require high resolution, the ability to capture sharp and dynamic interfaces, while balancing global memory footprint and the overhead of "ghost" cells.
 This configuration should not be changed as it would violate the intent of the benchmark.
@@ -227,4 +227,4 @@ Verification of Results
 References
 ==========
 
-.. [site] Jonah Miller, 'Parthenon', 2023. [Online]. Available: https://github.com/parthenon-hpc-lab/parthenon. [Accessed: 20- Mar- 2023]
+.. [Parthenon-VIBE] Jonah Miller, 'Parthenon', 2023. [Online]. Available: https://github.com/parthenon-hpc-lab/parthenon. [Accessed: 20- Mar- 2023]
diff --git a/doc/sphinx/5_mlmd/mlmd.rst b/doc/sphinx/5_mlmd/mlmd.rst
@@ -2,12 +2,12 @@
 MLMD
 ******
 
-This is the documentation for the ATS-5 MLMD Benchmark for HIPPYNN driven kokkos-Lammps Molecular Dynamics. 
+This is the documentation for the ATS-5 MLMD Benchmark for HIPPYNN [HIPPYNN]_ driven kokkos-Lammps Molecular Dynamics. 
 
 Purpose
 =======
 
-To benchmark perfromance of Lammps driven Molecular Dynamics. The problem configured in this test is a small Ag model built using the data included in the Allegro paper (DOI: https://doi.org/10.1038/s41467-023-36329-y)   
+To benchmark performance of Lammps [Lammps]_ driven Molecular Dynamics. The problem configured in this test is a small Ag model built using the data included in the Allegro paper (DOI: https://doi.org/10.1038/s41467-023-36329-y)   
 
 Characteristics
 ===============
@@ -16,15 +16,15 @@ Problem
 -------
 This is a set of simulations on 1,022,400 silver atoms at a 5fs time step. 
 
-This benchmark is comparible to the Allegro Ag simulation: https://www.nature.com/articles/s41467-023-36329-y
+This benchmark is comparable to the Allegro Ag simulation: https://www.nature.com/articles/s41467-023-36329-y
 
 This is a general purpose proxy for many MD simulations. 
 
 This test will evaluate Pytorch performance, GPU performance, and MPI performance. 
 
 Figure of Merit
 ---------------
-The figure of merrit is the throughput of the MD simulations, whch is reported by lammps as 'Matom-step/s'. 
+The figure of merit is the throughput of the MD simulations, which is reported by Lammps as 'Matom-step/s'. 
 
 Building
 ========
@@ -241,5 +241,5 @@ Verification of Results
 References
 ==========
 
-.. [site] Nicolas Lubbers, "HIPPYNN" 2021. [Online]. Available: https://github.com/lanl/hippynn. [Accessed: 6- Mar- 2023]
-.. [site] Axel Kohlmeyer et. Al, "Lammps". [Online]. Available: https://github.com/lammps/lammps. [Accessed: 6- Mar- 2023]
+.. [HIPPYNN] Nicolas Lubbers, "HIPPYNN" 2021. [Online]. Available: https://github.com/lanl/hippynn. [Accessed: 6- Mar- 2023]
+.. [Lammps] Axel Kohlmeyer et. Al, "Lammps". [Online]. Available: https://github.com/lammps/lammps. [Accessed: 6- Mar- 2023]