Wiki source code of Benchmarking

Last modified by paulapopa on 2022/10/04 08:19

author	version	line-number	content
		1	== TVB-INVERSION 1.0.0 ==
		2
		3	=== Sampling priors - Remote execution ===
		4
		5	==== ====
		6
		7	==== 1. Execution times ====
		8
		9	In this section we provide some benchmarks regarding the execution times for the sampling priors step within tvb-inversion workflow, where we need to run a large number of simulations.
		10
		11	These have been computed on DAINT-CSCS HPC, on a single node, with different configurations for the number of simulations and the number of workers.
		12
		13	\|=(% scope="row" %)(((
		14	Model
		15	)))\|=(((
		16	Sim length (s)
		17	)))\|=Regions\|=Nr simulations\|=Nr workers\|=Execution time (hh:mm)
		18	\|=MontbrioPazoRoxin\|30\|100\|(((
		19	30
		20	)))\|30\|00:17
		21	\|=MontbrioPazoRoxin\|30\|100\|200\|20\|01:08
		22	\|=MontbrioPazoRoxin\|30\|100\|300\|30\|01:10
		23	\|=MontbrioPazoRoxin\|30\|100\|400\|40\|01:18
		24	\|=MontbrioPazoRoxin\|30\|100\|500\|50\|01:34
		25	\|=MontbrioPazoRoxin\|30\|100\|500\|55\|01:30
		26	\|=MontbrioPazoRoxin\|30\|100\|600\|55\|01:45
		27	\|=MontbrioPazoRoxin\|30\|100\|600\|60\|OOM
		28	\|=MontbrioPazoRoxin\|60\|100\|500\|40\|(((
		29	03:07
		30	)))
		31	\|(% scope="row" %)MontbrioPazoRoxin\|60\|100\|500\|55\|OOM
		32	\|= \|= \|= \|= \|= \|=
		33	\|= \| \| \| \| \|
		34
		35	==== ====
		36
		37	==== 2. Limitations ====
		38
		39	* Reaching the memory limit on the CSCS node.
		40	** For a simulation of 30 seconds, we can fit 55 parallel workers in the available memory
		41	** For a simulation of 60 seconds, we can fit 40 parallel workers in the available memory
		42	* Reaching the maximum connections on CSCS could happen during a run
		43	** connect once to launch the job
		44	** connect multiple times during the monitoring step to check the status of the job
		45	** connect once to stage out results

Wiki source code of Benchmarking

TVB Inversion