tAdd more content - brcon2020_adc - my presentation for brcon2020
(HTM) git clone git://src.adamsgaard.dk/.brcon2020_adc
(DIR) Log
(DIR) Files
(DIR) Refs
(DIR) LICENSE
---
(DIR) commit da0a4b7ed98cd134c4053d43613ec69832244e74
(DIR) parent 11961efe351b5aaa8bad5a292a69abb4b08be295
(HTM) Author: Anders Damsgaard <anders@adamsgaard.dk>
Date: Tue, 28 Apr 2020 18:36:27 +0200
Add more content
Diffstat:
M brcon2020_adc.md | 44 ++++++++++++++++++++++++-------
1 file changed, 34 insertions(+), 10 deletions(-)
---
(DIR) diff --git a/brcon2020_adc.md b/brcon2020_adc.md
t@@ -239,12 +239,15 @@ Academic interests:
* Performance through massively parallel deployment (MPI, GPGPU)
- * NOAA/NCRC Gaea cluster
+ * NOAA/DOE NCRC Gaea cluster
* 2x Cray XC40, "Cray Linux Environment"
* 4160 nodes, each 32 to 36 cores, 64 GB memory
* infiniband
* total: 200 TB memory, 32 PB SSD, 5.25 petaflops (peak)
+% National Climate-Computing Research Center
+% as much power as a small city
+
## Scaling problem
New algorithms hard to implement in HPC codes
t@@ -264,15 +267,35 @@ NO!
%## Measuring computational energy use
+## Example: Ice-sheet flow with sediment/fluid modeling
-## Algorithm matters
+
+ --------------------------._____ ATMOSPHERE
+ -----> ```--..
+ ICE `-._________________ __
+ -----> ------> |vvvv| |vvv
+ _________________| |__|
+ -----> ,'
+ ,' <>< OCEAN
+ ----> / ><>
+ ____________________________________/___________________________________
+ SEDIMENT -->
+ ________________________________________________________________________
* example: granular dynamics and fluid flow simulation for glacier flow
+* 90% of Antarctic ice sheet mass driven by ice flow over sediment
+
+* need to understand ice-basal sliding in order to project sea-level rise
+
+
+## Algorithm matters
+
sphere: git://src.adamsgaard.dk/sphere
C++, Nvidia C, cmake, Python, Paraview
massively parallel, GPGPU
detailed physics
+ 20,191 LOC
#pause
3 month computing time on nvidia tesla k40 (2880 cores)
t@@ -285,6 +308,7 @@ NO!
C99, makefiles, gnuplot
single threaded
simple physics
+ 2,348 LOC
#pause
real: 0m00.07 s on potato laptop from 2012
t@@ -296,16 +320,16 @@ NO!
for numerical simulation:
* high-level languages
- * easy
- * produces results quickly
- * does not develop low-level programming skills
- * no insight into numerical algorithm
- * realistically speaking: no direct way to HPC
+ * easy
+ * produces results quickly
+ * does not develop low-level programming skills
+ * no insight into numerical algorithm
+ * realistically speaking: no direct way to HPC
* low-level languages
- * require low-level skills
- * saves electrical energy
- * directly to HPC, just sprinkle some MPI on top
+ * require low-level skills
+ * saves electrical energy
+ * directly to HPC, just sprinkle some MPI on top
## Thanks