Dear group members,<br><br>For those of you that are interested in the recent paper by Martin Plenio: Here are some comments by Steven White, the person that invented DMRG. As you guys know, I am here amongst the DMRG gods, and asked him to tell me his impressions. They are below.<br>
<br>I am very happy to hear that Ville already has DMRG in his fantastic toolkit, so I think we could potentially have t-DMRG soon and use similar models if needed. <br><br>Best,<br>A.<br><br clear="all">Alán Aspuru-Guzik | Assistant Professor<br>
Harvard University | Department of Chemistry and Chemical Biology<br>12 Oxford Street, Room M113 | Cambridge, MA 02138<br>(617)-384-8188 | <a href="http://aspuru.chem.harvard.edu">http://aspuru.chem.harvard.edu</a><br><br>
<br>
<br><br><div class="gmail_quote">---------- Forwarded message ----------<br>From: <b class="gmail_sendername">Steven White</b> <span dir="ltr">&lt;<a href="mailto:srwhite@uci.edu">srwhite@uci.edu</a>&gt;</span><br>Date: Tue, Mar 30, 2010 at 2:29 AM<br>
Subject: Re: Paper<br>To: Alan Aspuru-Guzik &lt;<a href="mailto:alan@aspuru.com">alan@aspuru.com</a>&gt;<br><br><br>Dear Alan,<div>The mapping looks similar to things I&#39;ve seen in solid state.  I have thought the near-neighbor</div>
<div>coupling could be generated with a fair amount of generality by a Lanczos type procedure--so</div>
<div>I&#39;m not sure how this polynomial version is different.</div><div>Regarding the t-DMRG, the paper is missing a section on the calculational details, and references.</div><div>This might be like saying &quot;using DFT, we solved this problem and here are the energies&quot; without</div>

<div>ever mentioning which DFT, which basis, etc etc.  It is all probably fine, but here are some questions</div><div>needing answers in any reasonable paper:</div><div>Which version of t-DMRG (reference), which breakup (2nd order Trotter?  4-th order Trotter?)  What</div>

<div>time step?  How big a lattice? How many time steps?  How many states kept?  Was the truncation</div><div>controlled by a truncation error cutoff or a maximum number of states?  What was the accumulated</div><div>truncation error? What was the maximum in the number of states kept if the a truncation error cutoff was used?</div>

<div>Can they show any plots with different cutoff criterion for comparison?  What do they mean by</div><div>numerically exact? (This phrase is sometimes used for regular DMRG in a 1D system, but I&#39;ve</div><div>never heard it used for time dependent DMRG.) How long did the calculations take?</div>

<div><br></div><div>Without these sorts of details one can&#39;t really evaluate the t-DMRG part.</div><div>Best regards,</div><div>Steve</div><div><div></div><div class="h5"><div><br><br><div class="gmail_quote">On Tue, Mar 30, 2010 at 1:15 AM, Alan Aspuru-Guzik <span dir="ltr">&lt;<a href="mailto:alan@aspuru.com" target="_blank">alan@aspuru.com</a>&gt;</span> wrote:<br>

<blockquote class="gmail_quote" style="border-left: 1px solid rgb(204, 204, 204); margin: 0pt 0pt 0pt 0.8ex; padding-left: 1ex;">Dear Steven<br><br>This is the paper I was talking to you about. Let me know your thoughts on the time evolution aspect.<br>

<br>Best and thanks!<br><font color="#888888">Alan<br>
</font></blockquote></div><br></div>
</div></div></div><br>