While I think Misters Samborski and Yorgey have fixed the performance problem, I&#39;m still curious about the apparent synergy effect:<br><br><div class="gmail_quote"><blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex">
<div><div>
&gt; First example [1] just builds a HashTable of ~1 500 000 ByteStrings.<br>
&gt; The program takes about 11 seconds to execute. </div></div></blockquote><div> </div><blockquote class="gmail_quote" style="margin:0pt 0pt 0pt 0.8ex;border-left:1px solid rgb(204,204,204);padding-left:1ex"><div><div>
&gt; Reading data itself is not a<br>
&gt; problem, as I&#39;ve measured the execution time of a program which just<br>
&gt; reads the file (and prints its last line to actually force the read,<br>
&gt; hence &quot;handle&quot; Haskell&#39;s laziness) and it takes no more than a second.<br></div></div></blockquote><div> </div><blockquote class="gmail_quote" style="margin:0pt 0pt 0pt 0.8ex;border-left:1px solid rgb(204,204,204);padding-left:1ex">
<div><div>
&gt; Coupled together reading and building, however, takes unbearably long.<br>
&gt; I didn&#39;t wait till the program stopped and just terminated it, but it<br>
&gt; had been running for ~10 minutes already.<br>
</div></div></blockquote></div><br>???<br><br><br>