boltzgen_examples, branch master Examples on how to use the boltzgen code generator Integrate AA in CUDA example 2020-01-17T20:07:49+00:00 Adrian Kummerlaender 2020-01-17T20:07:49+00:00 682c46c7932f6dfe8d3874bd54e412162a81d216 






Integrate SSS in Open(C,G)L interop example
2020-01-16T21:21:00+00:00

Adrian Kummerlaender

2020-01-16T21:20:27+00:00

dbe19deb0d42ee7c2359d951686c6100d4ec27b3












Add support for SSS propagation in OpenCL example
2020-01-10T23:18:20+00:00

Adrian Kummerlaender

2020-01-10T23:18:20+00:00

74392de73ae6c799486f34eccf5c2c37ff9187ea












Add C++ compiler to CUDA shell expression
2020-01-10T23:16:21+00:00

Adrian Kummerlaender

2020-01-10T23:16:21+00:00

ac226d7ddf6ccfde2d5db3dc4807456ab3d46521

I am not entirely sure why this was not required previously.





I am not entirely sure why this was not required previously.







Share Lattice implementation between plain and interop OpenCL example
2019-11-12T21:59:21+00:00

Adrian Kummerlaender

2019-11-12T21:59:21+00:00

8c1f317010eeba28b3d4ae3ef158b1e760997ed1












Share boltzgen version between expressions
2019-11-12T21:59:03+00:00

Adrian Kummerlaender

2019-11-12T21:59:03+00:00

2dba6db0c94ccb260f3e9345b25146ce75ccea9a












Export density, support indexing parameter in CUDA example
2019-11-12T19:01:35+00:00

Adrian Kummerlaender

2019-11-12T19:01:35+00:00

155c630827bae820412d663e1579929a2832d24b












Use separate Nix environments for each target
2019-11-11T12:41:46+00:00

Adrian Kummerlaender

2019-11-11T12:41:46+00:00

336d20eed737445f3b2aa7d875c90ccc64a97213












Add LDC example for CUDA target
2019-11-10T20:25:07+00:00

Adrian Kummerlaender

2019-11-10T20:25:07+00:00

988818c224197eb506ac2f42edef13d7bd1cc11a












Fix SSS population padding
2019-11-09T22:47:58+00:00

Adrian Kummerlaender

2019-11-09T22:47:58+00:00

52033eeec131fe862ad85dcb52e961b5904330eb

D2Q9 and D3Q27 worked by coincidence, should now work for all lattices.





D2Q9 and D3Q27 worked by coincidence, should now work for all lattices.







Add support for SSS pattern to C++ LDC example
2019-11-09T20:36:04+00:00

Adrian Kummerlaender

2019-11-09T19:44:20+00:00

910c2299f4be8ec1819cddaae508d60cddcb9796












Add OpenCL GL interop example
2019-11-09T15:21:54+00:00

Adrian Kummerlaender

2019-11-09T15:21:54+00:00

20b821bfdf8441b6e2b0e086097a729f006e5b0a












Adapt to upstream function rename
2019-11-09T14:41:56+00:00

Adrian Kummerlaender

2019-11-09T14:41:56+00:00

0a68fc826a80123e3ce98f6696d834d0013b2c43












Add basic AA pattern support for OpenCL example
2019-11-06T20:26:03+00:00

Adrian Kummerlaender

2019-11-06T20:24:08+00:00

4610b50bfe47ed0d75f30279fca69d0dcdc04ee2












Add basic AA pattern support for C++ example
2019-11-05T19:14:29+00:00

Adrian Kummerlaender

2019-11-05T19:14:29+00:00

527e8aa9c33d7c82b548b2280c979f653a7131ae












Adapt to upstream changes
2019-11-04T22:50:41+00:00

Adrian Kummerlaender

2019-11-04T22:50:41+00:00

8285173fa9c69de72020962506eb080a49be0ce5












Adapt to upstream changes
2019-11-02T16:57:23+00:00

Adrian Kummerlaender

2019-11-02T16:57:23+00:00

b708361bf722b2c19f47699ceb931c6efeb8c9d9












Rename folder, add basic README.md
2019-10-30T18:45:39+00:00

Adrian Kummerlaender

2019-10-30T18:45:39+00:00

d540b701836dbcdef727947f52a0ca01430968a6












Extract settings into config file, add documentation
2019-10-30T18:40:56+00:00

Adrian Kummerlaender

2019-10-30T18:40:56+00:00

fc58dab96f27624435a1c08480234a7c03071e27












Move C++ LDC template from upstream, improve build
2019-10-30T15:09:05+00:00

Adrian Kummerlaender

2019-10-30T15:08:19+00:00

2bff37d2d4c0e210fd3e4288bdd0add12b9eb86b












Merge shell environments
2019-10-29T20:07:35+00:00

Adrian Kummerlaender

2019-10-29T20:07:35+00:00

022f427afe9bab50ece321442f6a62899771bbbc












Add example for C++ target
2019-10-29T20:04:53+00:00

Adrian Kummerlaender

2019-10-29T20:04:53+00:00

a725685ddaa2eaa0711ceb51e1f74d09349d618e












Move example into subfolder
2019-10-29T19:50:49+00:00

Adrian Kummerlaender

2019-10-29T19:46:21+00:00

6bf667eeb0c8f40e0f6a28df1ecf4a40824771c3












Adapt to upstream changes
2019-10-29T19:41:33+00:00

Adrian Kummerlaender

2019-10-29T19:41:33+00:00

39aa2832b3d8e1a3e1dd2efb0bf9629d8190ed39












Fix double precision
2019-10-29T15:06:09+00:00

Adrian Kummerlaender

2019-10-29T15:06:09+00:00

12e8d10df34e8def9a5fb3944bff6de25f020d27












Basic 2D LDC using boltzgen for kernel generation
2019-10-28T22:08:39+00:00

Adrian Kummerlaender

2019-10-28T22:07:44+00:00

42f4ae5f67f17ff37b3e95cab3c905668816eee8

Using cell lists as parameters for multiple non-branching kernels seems
to reduce performance by ~50 MLUPS (for single precision D2Q9).
This might be alleviated by padding the cell lists to enable thread
layout control or by improved kernel dispatching.

On the upside this OpenCL program runs not only on GPUs but is also vectorized on Intel
CPUs yielding about 180 MLUPS (single precision) and - anticlimactically - 85 MLUPS for
double precision on a i7-4790K.
However both these values compare well to the performance of established CPU LBM codes.





Using cell lists as parameters for multiple non-branching kernels seems
to reduce performance by ~50 MLUPS (for single precision D2Q9).
This might be alleviated by padding the cell lists to enable thread
layout control or by improved kernel dispatching.

On the upside this OpenCL program runs not only on GPUs but is also vectorized on Intel
CPUs yielding about 180 MLUPS (single precision) and - anticlimactically - 85 MLUPS for
double precision on a i7-4790K.
However both these values compare well to the performance of established CPU LBM codes.