Changes between Version 9 and Version 10 of doc/tec/gpu

Timestamp:

Sep 10, 2013 11:18:56 AM (11 years ago)

Author:

raasch

Comment:

--

doc/tec/gpu

@@ -42 +42 @@
 In single-core mode, lateral boundary conditions completely run on device. Most loops in {{{pres}}} ported. Vertical boundary conditions ({{{boundary_conds}}}) ported.
 
-'''Results for 512x512x64 grid (time in micro-s per gridpoint and timestep):''' \\
-||.1 ||2*Tesla, quadcore, pgi             ||0.32606 ||
-||.2 ||1*Tesla, single-core (no MPI), pgi ||0.42138 ||
-||.3 ||quadcore, pgi   (acc)              ||0.78062 ||
-||.4 ||quadcore, pgi   (vec)              ||0.64060 ||
-||.5 ||quadcore ??     (cache)            ||0.67272 ||
-||.6 ||quadcore ??     (cache)            ||0.79969 ||
-||.7 ||quadcore ??     (vec)              ||0.77608 ||
-||.8 ||quadcore ??     (acc)              ||1.00139 ||
+r1221 \\
+Reduction operations in {{{pres}}} and {{{flow_statistics ported}}}.
+
+'''Results for 256x256x64 grid (time in micro-s per gridpoint and timestep):''' \\
+||.1 ||1*Tesla, single-core (no MPI), pgi13.6             ||0.33342 ||r1221 ||
+||.2 ||single-core (no MPI), pgi13.6   (cache, Temperton) ||2.34144 ||r1221 ||
 
 The initialization time of the GPU (power up) can be avoided by running {{{/muksoft/packages/pgi/2013-136/linux86-64/13.6/bin/pgcudainit}}} in background.
 
+For current PGI compiler version 13.6, use "-ta=nocache" and set environment variable {{{PGI_ACC_SYNCHRONOUS=1}}}. Otherwise, there will be a significant loss in performance (factor of two!).
+
 '''Next steps:'''
 
-* testing the newest PGI 13.2 compiler version, porting of reduction operations ({{{timestep}}}, {{{flow_statistics}}}, divergence in {{{pres}}}), check the capability of parallel regions (can IF-constructs be removed from inner loops?)
+* porting of MIN/MAXLOC operations in ({{{timestep}}}, porting of {{{disturb_fields}}} (implement parallel random number generator)
+* check the capability of parallel regions (can IF-constructs be removed from inner loops?)
 * for MPI mode update ghost boundaries only, overlapping of update/MPI-transfer and computation
 * overlapping communication in pressure solver (alltoall operations)
-* porting of remaining things (disturbances, calc_liquid_water_content, compute_vpt, averaging, I/O, etc.)
+* porting of remaining things (calc_liquid_water_content, compute_vpt, averaging, I/O, etc.)
 * ...