ECLOUD vs Py ECLOUD at 450 Ge V

ECLOUD vs. Py. ECLOUD at 450 Ge. V C. O. Domínguez, G. Iadarola and F. Zimmermann 10 August 2012 - e - cloud meeting

Outline 1) Some build-up remarks 2) Benchmarking pressure with both codes 10 August 2012 - e - cloud meeting

Outline 1) Some build-up remarks 2) Benchmarking pressure with both codes 10 August 2012 - e - cloud meeting

Build-up remarks Always bigger slope for Py. EC Plateau for Py. EC; Slope for EC Same initial seeds 10 August 2012 - e - cloud meeting

Build-up remarks 10 August 2012 - e - cloud meeting

Build-up remarks 10 August 2012 - e - cloud meeting

Build-up remarks e- density (log scale in y) e- density (derivative) Bunch passage 10 August 2012 - e - cloud meeting

Build-up remarks First two bunches Primary e- Secondary e- Flying time of primary e- from the center to the walls 10 August 2012 - e - cloud meeting

Build-up remarks Two bunches in saturation Primary e- are negligible No more plateau 10 August 2012 - e - cloud meeting

Build-up remarks Some questions arise: - Why can’t we see the plateau in EC? - Why the slope is always bigger in Py. EC? 10 August 2012 - e - cloud meeting

Outline 1) Some build-up remarks 2) Benchmarking pressure with both codes 10 August 2012 - e - cloud meeting

Pressure Benchmarking I order to decide which code to use for benchmarking pressure measurements, some tests on the sensitivity to the number of macroparticles have been done: - In EC I varied the parameter “npepb” to 1000 and 3000 (default: 2000). - The mechanism in Py. EC is quite different so I varied different parameters. Default #1 #2 #3 #4 #5 #6 N_mp_max 100 k 130 k 100 k N_mp_regen 50 k 40 k 70 k 50 k 50 k N_mp_regen_low 5 k 5 k 5 k 2. 5 k 10 k 5 k 5 k N_mp_after_regen 10 k 10 k 5 k 20 k 10 k nel_mp_ref_0 (e-/m) 10 10 10 50 10 August 2012 - e - cloud meeting

Pressure Benchmarking I order to decide which code to use for benchmarking pressure measurements, some tests on the sensitivity to the number of macroparticles have been done: - In EC I varied the parameter “npepb” to 1000 and 3000 (default: 2000). - The mechanism in Py. EC is quite different so I varied different parameters. N_mp_max N_mp_regen N_mp_after_regen N_mp_regen_low nel_mp_ref_0 is the number of e- created per macroparticle during the passage of the first bunch 10 August 2012 - e - cloud meeting

Pressure Benchmarking I order to decide which code to use for benchmarking pressure measurements, some tests on the sensitivity to the number of macroparticles have been done: - In EC I varied the parameter “npepb” to 1000 and 3000 (default: 2000). - The mechanism in Py. EC is quite different so I varied different parameters. Default #1 #2 #3 #4 #5 #6 N_mp_max 100 k 130 k 100 k N_mp_regen 50 k 40 k 70 k 50 k 50 k N_mp_regen_low 5 k 5 k 5 k 2. 5 k 10 k 5 k 5 k N_mp_after_regen 10 k 10 k 5 k 20 k 10 k nel_mp_ref_0 (e-/m) 10 10 10 50 10 August 2012 - e - cloud meeting

Pressure Benchmarking ECLOUD Npepb=1000 10 August 2012 - e - cloud meeting

Pressure Benchmarking ECLOUD Npepb=2000 10 August 2012 - e - cloud meeting

Pressure Benchmarking ECLOUD Npepb=3000 10 August 2012 - e - cloud meeting

Pressure Benchmarking ECLOUD Npepb=2000 Npepb=1000 Npepb=3000 10 August 2012 - e - cloud meeting

Pressure Benchmarking ECLOUD 10 August 2012 - e - cloud meeting

Pressure Benchmarking Py. ECLOUD N_mp_max=100000, 130000; #size of allocated vectors #Regen parameters N_mp_regen=40000, 70000; N_mp_regen_low=5000; N_mp_after_regen=10000; nel_mp_ref_0=10. #e-/m 10 August 2012 - e - cloud meeting

Pressure Benchmarking Py. ECLOUD N_mp_max=100000; #size of allocated vectors #Regen parameters N_mp_regen=50000; N_mp_regen_low=2500, 10000; N_mp_after_regen=5000, 20000; nel_mp_ref_0=10. #e-/m 10 August 2012 - e - cloud meeting

Pressure Benchmarking Py. ECLOUD N_mp_max=100000; #size of allocated vectors #Regen parameters N_mp_regen=50000; N_mp_regen_low=5000; N_mp_after_regen=10000; nel_mp_ref_0=10. , 50. #e-/m 10 August 2012 - e - cloud meeting

Pressure Benchmarking Py. ECLOUD N_mp_max=100000; #size of allocated vectors #Regen parameters N_mp_regen=50000; N_mp_regen_low=5000; N_mp_after_regen=10000; nel_mp_ref_0=50. , 100. #e-/m 10 August 2012 - e - cloud meeting

Pressure Benchmarking Summary: - In my opinion, Py. EC seems more robust, at least in the parameters explored - Other studies done on the sensitivity to grid dimensions also show that with Py. EC differences are smaller. - The shape of the different ratio lines is always more “stable” with Py. EC. 10 August 2012 - e - cloud meeting

Thank you for your attention! 10 August 2012 - e - cloud meeting