User:Arash - Revision history

Arash: /* compiler_bincuda.qrcuda/mwcol/bcacol */

2011-08-21T11:51:04Z

compiler_bincuda.qrcuda/mwcol/bcacol

← Older revision		Revision as of 11:51, 21 August 2011
Line 233:		Line 233:

	These test functions are solving {{srcbranch\|arash\|models/test/bintok/bincuda/larg_distil.a4c}}, {{srcbranch\|arash\|models/test/bintok/bincuda/mwcolumn.a4c}} and {{srcbranch\|arash\|models/test/bintok/bincuda/bcacolumn.a4c}} respectively.		These test functions are solving {{srcbranch\|arash\|models/test/bintok/bincuda/larg_distil.a4c}}, {{srcbranch\|arash\|models/test/bintok/bincuda/mwcolumn.a4c}} and {{srcbranch\|arash\|models/test/bintok/bincuda/bcacolumn.a4c}} respectively.
			[[Category:GSOC2011]][[Category:ASCEND Contributors]]

Arash: /* Running the test */

2011-08-21T11:41:28Z

Running the test

← Older revision		Revision as of 11:41, 21 August 2011
Line 203:		Line 203:
	END c5_10_demo_column;</source>		END c5_10_demo_column;</source>

	== Running the ~~test~~ ==		== Running the tests ==

	A CUnit test ~~case~~ was prepared to test ~~BinCUDA generation~~ and ~~execution~~.		A CUnit test suite was prepared to test the QRCUDA solver and the generated CUDA model evaluator objects (i.e. BinCUDAs). The test suite code is located in test_bincuda.c({{srcbranch\|arash\|ascend/compiler/test/test_bincuda.c}}) and contains six test functions; gen, satpnt, multivec, qrcuda, mwcol and bcacol.
	The code is located in test_bincuda.c({{srcbranch\|arash\|ascend/compiler/test/test_bincuda.c}}).		You can run the test by executing "test/test compiler_bincuda.[test function name]" at the top level ASCEND directory.
	You can run the test by executing "test/test compiler_bincuda.~~gen~~" at top		For more information about how QRCUDA and BinCUDAs are interacting please refer to ({{srcbranch\|arash\|ascend/bintokens/bincuda/BinCUDA_Readme.txt}}).
	level ASCEND directory.		To change the current benchmark model, you can change the macro DEF_FILENAMESTEM
			in the code. [Please note that if your model includes any specific
			ASCEND function (e.g. asc_ipow) the function should be defined in the
			btcudapl.cu ({{srcbranch\|arash\|ascend/bintokens/bincuda/btcudapl.cu}}) file.]

	~~The~~ test ~~case~~ outputs the CPU-based evaluation time, GPU-based evaluation		=== compiler_bincuda.gen ===
			This test function outputs the CPU-based evaluation time, GPU-based evaluation
	time and the number of equations in the model.		time and the number of equations in the model.
	It generates the code in the "/tmp" directory and the Makefile located in the same directory		It generates the code in the "/tmp" directory and the Makefile located in the same directory
	is responsible for building the shared binary object for BinCUDAs. The CUDA		is responsible for building the shared binary object for BinCUDAs. The CUDA
	build and compile commands are provided in the Makefile({{srcbranch\|arash\|ascend/bintokens/bincuda/Makefile}}).		build and compile commands are provided in the Makefile ({{srcbranch\|arash\|ascend/bintokens/bincuda/Makefile}}).

			=== compiler_bincuda.satpnt ===

	~~To change~~ the ~~current benchmark~~ model, ~~you can change~~ the ~~macro FILENAMESTEM~~		In the multi-vector residual evaluator, the model is concurrently evaluated for multiple input vectors. As the GPU parallel architecture is used, the evaluation time for multiple inputs is equal to the evaluation time for a single input. The "satpnt" test function is responsible for determining the ''saturation point'' for a specific model. We define the saturation point as the maximum number of vectors where the computational time for concurrent residual evaluation is equal to that time measured for a single input vector evaluation.
	in the ~~code~~. [Please note that ~~if your model includes any specific~~
	~~ASCEND~~ function (e.g. ~~asc_ipow)~~ the function ~~should be defined~~ in the		Please note that this test function is only measuring the computational time and the time for data transfer between CPU and GPU is not provided in the results.
	~~btcudapl~~.~~cu ({{srcbranch\|arash\|ascend~~/~~bintokens~~/~~bincuda/btcudapl.cu}}) file.]~~
			=== compiler_bincuda.multivec ===

			In the "multivec" test function, the results achieved from the multi vector evaluators is verified against the standard CPU based implementation provided in ASCEND framework and then the computational performance of multivector evaluators are measured.

			=== compiler_bincuda.qrcuda/mwcol/bcacol ===

	~~[[Category:GSOC2011]]~~		These test functions are solving {{srcbranch\|arash\|models/test/bintok/bincuda/larg_distil.a4c}}, {{srcbranch\|arash\|models/test/bintok/bincuda/mwcolumn.a4c}} and {{srcbranch\|arash\|models/test/bintok/bincuda/bcacolumn.a4c}} respectively.
	~~[[Category:ASCEND Contributors]]~~

Arash: /* Progress */

2011-08-18T11:55:52Z

Progress

← Older revision		Revision as of 11:55, 18 August 2011
Line 92:		Line 92:
	** Concurrent kernel launcher (streaming) is implemented for residual evaluator kernels. (In a model with ~80000 relations, the evaluators are now executed 4x faster compared to the previous version that used sequential kernel launcher)		** Concurrent kernel launcher (streaming) is implemented for residual evaluator kernels. (In a model with ~80000 relations, the evaluators are now executed 4x faster compared to the previous version that used sequential kernel launcher)
	** Multi-vector evaluators were tested and the results were identical to normal CPU based evaluators		** Multi-vector evaluators were tested and the results were identical to normal CPU based evaluators
			** Multi-vector evaluators were integrated to the line-search algorithm of QRCUDA

	== Ideas and Issues ==		== Ideas and Issues ==

Arash: /* Progress */

2011-08-12T11:19:23Z

Progress

← Older revision		Revision as of 11:19, 12 August 2011
Line 91:		Line 91:
	* After 6-August		* After 6-August
	** Concurrent kernel launcher (streaming) is implemented for residual evaluator kernels. (In a model with ~80000 relations, the evaluators are now executed 4x faster compared to the previous version that used sequential kernel launcher)		** Concurrent kernel launcher (streaming) is implemented for residual evaluator kernels. (In a model with ~80000 relations, the evaluators are now executed 4x faster compared to the previous version that used sequential kernel launcher)
			** Multi-vector evaluators were tested and the results were identical to normal CPU based evaluators

	== Ideas and Issues ==		== Ideas and Issues ==

Arash: /* Progress */

2011-08-07T11:40:45Z

Progress

← Older revision		Revision as of 11:40, 7 August 2011
Line 89:		Line 89:
	*** The normal calculator is extended to calculate the minimum square normal value for multi vector evaluation		*** The normal calculator is extended to calculate the minimum square normal value for multi vector evaluation
	** A unit test created for testing multi vector evaluators		** A unit test created for testing multi vector evaluators
			* After 6-August
			** Concurrent kernel launcher (streaming) is implemented for residual evaluator kernels. (In a model with ~80000 relations, the evaluators are now executed 4x faster compared to the previous version that used sequential kernel launcher)

	== Ideas and Issues ==		== Ideas and Issues ==

Jpye at 05:05, 6 August 2011

2011-08-06T05:05:56Z

← Older revision		Revision as of 05:05, 6 August 2011
Line 218:		Line 218:

	[[Category:GSOC2011]]		[[Category:GSOC2011]]
			[[Category:ASCEND Contributors]]

Arash: /* Project Plan */

2011-08-04T05:02:46Z

Project Plan

← Older revision		Revision as of 05:02, 4 August 2011
Line 21:		Line 21:
	** Optimise the CUDA code		** Optimise the CUDA code
	*** Change kernels memory access pattern to coalesced access		*** Change kernels memory access pattern to coalesced access
	*** Store the mapping information into fast texture constant memory		<!-- *** Store the mapping information into fast texture constant memory -->
	*** Change Memory management model from standard model to PINNED memory management. This makes the memory transfer between host and device faster.		*** Change Memory management model from standard model to PINNED memory management. This makes the memory transfer between host and device faster.
	** Implement hybrid CPU/GPU based evaluation instead of GPU-based evaluation. By doing this, the CPU can be used for the small equation groups while the GPU is busy evaluating the large groups.		** Implement hybrid CPU/GPU based evaluation instead of GPU-based evaluation. By doing this, the CPU can be used for the small equation groups while the GPU is busy evaluating the large groups.
	*** support for models containing 'external relations'		<!-- *** support for models containing 'external relations'-->
	** Prepare a multi platform Makefile to compile and build BinCUDAs		** Prepare a <!-- multi platform --> Makefile to compile and build BinCUDAs
	** Complete the external functions in “btcudapl.cu”		<!-- ** Complete the external functions in “btcudapl.cu”-->

	* Implement the batch multi-vector residual evaluator		* Implement the batch multi-vector residual evaluator

Arash: /* Progress */

2011-08-02T12:34:05Z

Progress

← Older revision		Revision as of 12:34, 2 August 2011
Line 81:		Line 81:
	** Extensive search carried out to create large and solvable models (larger than current 30000). During this search, QRCUDA was tested with different models and several bugs were identified in QRCUDA and fixed.		** Extensive search carried out to create large and solvable models (larger than current 30000). During this search, QRCUDA was tested with different models and several bugs were identified in QRCUDA and fixed.
	** The next step is to create GPU-based line search.		** The next step is to create GPU-based line search.
			* After 16-July
			** The heuristic formula for multi-vector residual evaluator is defined (Armijo rule)
			** Research on different variation of Armijo rule was completed and I decided to use (0.5) as the coefficient, the main reason behind this decision is that we can calculate (0.5) ^N with a combination of shift-left and divide operators which has a great performance advantage over any other coefficient.
			* After 26-July
			** The evaluator kernels are converted from 2D kernels into 3D kernels (the extra dimension is used for input vectors created with Armijo rule)
			** A parallel kernel was implemented to calculate square normal of residuals
			*** The normal calculator is extended to calculate the minimum square normal value for multi vector evaluation
			** A unit test created for testing multi vector evaluators

	== Ideas and Issues ==		== Ideas and Issues ==

Arash: /* Test models */

2011-07-14T03:43:55Z

Test models

← Older revision		Revision as of 03:43, 14 July 2011
Line 172:		Line 172:

	=== Number of Equations ===		=== Number of Equations ===
	The model originally has 128 unique equation symbolic forms and 19959 equation instances.The number of relations in the model can be adjusted by changing two parameters, 100 and 51, by a multiplicative factor. For example in {{srcbranch\|arash\|models/test/bintok/larg_distil.a4c}},		The model originally has 128 unique equation symbolic forms and 19959 equation instances.The number of relations in the model can be adjusted by changing two parameters, 100 and 51, by a multiplicative factor. For example in {{srcbranch\|arash\|models/test/bintok/bincuda/larg_distil.a4c}},

	<source lang="a4c">REQUIRE "column.a4l";		<source lang="a4c">REQUIRE "column.a4l";
Line 182:		Line 182:
	</source>		</source>

	Alternatively, multiple columns can be used instead of single column ({{srcbranch\|arash\|models/test/bintok/larg_distil_2.a4c}}),		Alternatively, multiple columns can be used instead of single column ({{srcbranch\|arash\|models/test/bintok/bincuda/larg_distil_2.a4c}}),

	<source lang="a4c">REQUIRE "column.a4l";		<source lang="a4c">REQUIRE "column.a4l";

Arash: /* Progress */

2011-07-14T03:41:58Z

Progress

← Older revision		Revision as of 03:41, 14 July 2011
Line 68:		Line 68:
	** More clean-ups in the BinCUDAs.		** More clean-ups in the BinCUDAs.
	** The active block evaluation mechanism was added to the batch evaluator.		** The active block evaluation mechanism was added to the batch evaluator.
	** QRCUDA is now using GPU-based model evaluation for the residual evaluation in large blocks (the code was tested on {{srcbranch\|arash\|models/test/bintok/test2.a4c}}, more testing is required).		** QRCUDA is now using GPU-based model evaluation for the residual evaluation in large blocks (the code was tested on {{srcbranch\|arash\|models/test/bintok/bincuda/test2.a4c}}, more testing is required).
	** QRCUDA was tested with {{srcbranch\|arash\|models/test/bintok/larg_distil.a4c}} and after some bug fixes, the GPU evaluator results are now identical to the same results achieved from standard calc_residuals method.		** QRCUDA was tested with {{srcbranch\|arash\|models/test/bintok/bincuda/larg_distil.a4c}} and after some bug fixes, the GPU evaluator results are now identical to the same results achieved from standard calc_residuals method.
	* After 26-June		* After 26-June
	** The testcase was modified to solve the distillation model in both mass balance and energy balance mode.		** The testcase was modified to solve the distillation model in both mass balance and energy balance mode.
	** Performance analyses with valgrind and gprof.		** Performance analyses with valgrind and gprof.
	** Bug fix in PyGTK so now the system is re-analyzed after execution of the methods.		** Bug fix in PyGTK so now the system is re-analyzed after execution of the methods.
	** QRCUDA solved its first large model (31733 equations) in mass balance and energy balance mode, the results are identical to the QRSlv results. Both solvers are converged and the self_test method was executed without any error.		** QRCUDA solved its first large model (31733 equations) in mass balance and energy balance mode, the results are identical to the QRSlv results. Both solvers are converged and the self_test method was executed without any error ({{srcbranch\|arash\|models/test/bintok/bincuda/mwcolumn.a4c}}).
	* After 6-July		* After 6-July
	** QRCUDA was integrated to PyGTK.		** QRCUDA was integrated to PyGTK.