Comments (3)
I'm not a HPCG main developer but I know the local size should be divisible by 8 because HPCG uses 4-level multigrid preconditioner.
How about 432x432x432 or 424x424x424 instead of 430x430x430?
from hpcg.
HPCG uses multigrid preconditioner with three levels. Each level requires division by 2. So your original problem size has to be divisible by 8 (=222).
Performance observed with HPCG is not meant to track HPL. In fact, HPCG achieves a fraction of HPL achieves. Please see HPCG web site and look at the results pages that show what that fraction is for machines around the world.
from hpcg.
Hi,
Thanks for the quick reply
With hpcg.dat as:-
424 424 424
1200
the simulation completed within 115m7.580s. I got the following data:-
GFLOP/s Summary:
Raw DDOT: 9.54469
Raw WAXPBY: 4.74248
Raw SpMV: 6.57449
Raw MG: 0.665184
Raw Total: 0.792704
Total with convergence overhead: 0.792704
Total with convergence and optimization phase overhead: 0.773687
...................
__________ Final Summary __________:
HPCG result is VALID with a GFLOP/s rating of: 0.773687
HPCG 2.4 Rating (for historical value) is: 0.792703
Though with HPL i get approx 7.62620e+02 Gflops (for N=83328 NB=192 P=1 Q=1 505.81 )
HPCG number(0.773Gflops) are way too off from what i usually get via HPL (~780Gflops)
Or am i wrong on my observations here?
from hpcg.
Related Issues (20)
- compile error HOT 2
- Unit tests in `unittesting` directory fail to compile HOT 2
- I get a problem in the build step HOT 1
- If the --rt parameter is read from file, it does not get used HOT 1
- There exits code bug in graph multicoloring in OptimizeProblem.cpp
- Matlab example HOT 1
- HPCG Memory Output HOT 4
- Volta-enabled HPCG compilation HOT 1
- Undefined data attribute in parallel region with default(none) HOT 2
- HPCG Cuda Binary with MPI support not working properly for multiple hosts HOT 1
- HPCG crash when nx=440 ny=440 nz=424 HOT 4
- Does having warnings invalidate the benchmark results ? HOT 1
- Visualizing computational results ? HOT 1
- Formula to derive HPCG problem size - that fits in system memory HOT 1
- Number of Smoother Steps HOT 1
- HPCG Cuda Binary with multiple GPUs? HOT 5
- Number of OMP Threads for MPI_GCC_OMP HOT 2
- ComputeResidual.cpp:60:5: error: variable 'n' must have explicitly specified data sharing attributes HOT 4
- Loop upper bound implicitly shared by GCC causes error in ComputerResidual.cpp:60
- Hello. How to generate trace for HPCG? HOT 2
Recommend Projects
-
React
A declarative, efficient, and flexible JavaScript library for building user interfaces.
-
Vue.js
🖖 Vue.js is a progressive, incrementally-adoptable JavaScript framework for building UI on the web.
-
Typescript
TypeScript is a superset of JavaScript that compiles to clean JavaScript output.
-
TensorFlow
An Open Source Machine Learning Framework for Everyone
-
Django
The Web framework for perfectionists with deadlines.
-
Laravel
A PHP framework for web artisans
-
D3
Bring data to life with SVG, Canvas and HTML. 📊📈🎉
-
Recommend Topics
-
javascript
JavaScript (JS) is a lightweight interpreted programming language with first-class functions.
-
web
Some thing interesting about web. New door for the world.
-
server
A server is a program made to process requests and deliver data to clients.
-
Machine learning
Machine learning is a way of modeling and interpreting data that allows a piece of software to respond intelligently.
-
Visualization
Some thing interesting about visualization, use data art
-
Game
Some thing interesting about game, make everyone happy.
Recommend Org
-
Facebook
We are working to build community through open source technology. NB: members must have two-factor auth.
-
Microsoft
Open source projects and samples from Microsoft.
-
Google
Google ❤️ Open Source for everyone.
-
Alibaba
Alibaba Open Source for everyone
-
D3
Data-Driven Documents codes.
-
Tencent
China tencent open source team.
from hpcg.