Fix the wrong definition of avnorm in applyExp

[mingruyang]

According to page 17 of this paper, avnorm which is used for error estimate should be Av_{m+1} instead of Av_m.

[mtfishman]

Hi Mingru,

Sorry for the slow response. Have you seen that this change makes a difference in calculations you are doing (not doubting that it is the correct thing to do, just curious)?

Cheers, Matt

[mingruyang]

Hi Matt,

I only tested it with the sample code in the TDVP repo. Actually I didn’t see any difference in the results but it runs slightly slower. I think it doesn’t matter in the current setting but might affect the time step adjusting.

Best, Mingru

On Jan 16, 2020, at 11:44 AM, mtfishman [email protected] wrote:

Hi Mingru,

Sorry for the slow response. Have you seen that this change makes a difference in calculations you are doing (not doubting that it is the correct thing to do, just curious)?

Cheers, Matt

— You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub https://github.com/ITensor/ITensor/pull/315?email_source=notifications&email_token=AEHX2T5VNJ76DSV2P7CT5T3Q6C2JTA5CNFSM4KDLP7LKYY3PNVWWK3TUL52HS4DFVREXG43VMVBW63LNMVXHJKTDN5WW2ZLOORPWSZGOEJFJNFY#issuecomment-575313559, or unsubscribe https://github.com/notifications/unsubscribe-auth/AEHX2T7TBUOGSAM6IWF43E3Q6C2JTANCNFSM4KDLP7LA.