<div><span class="gmail_quote">On 10/24/07, <b class="gmail_sendername">Thomas Lindgren</b> <<a href="mailto:thomasl_erlang@yahoo.com">thomasl_erlang@yahoo.com</a>> wrote:</span><blockquote class="gmail_quote" style="margin:0;margin-left:0.8ex;border-left:1px #ccc solid;padding-left:1ex">
<br>--- Steve Vinoski <<a href="mailto:vinoski@ieee.org">vinoski@ieee.org</a>> wrote:<br><br> Anders, thanks for collecting and posting these.<br>> I've just performed a set<br>> of new timings for all of them, as listed below. For
<br>> each, I just ran this<br>> command:<br>><br>> time erl -smp -noshell -run <test_case> main<br>> o1000k.ap >/dev/null<br>><br>> where "<test_case>" is the name of the tbray test
<br>> case file. All were<br>> looped ten times, and I took the best timing for<br>> each. All tests were done<br>> on my 8-core 2.33 GHz dual Intel Xeon with 2 GB RAM<br>> Linux box, in a local<br>> (non-NFS) directory.
<br><br>So, looking at Steve's results on his 8-core system,<br>we have:<br><br> real user tbray5/real user/real<br>tbray5 9.8 -- 1.0 --<br>tbray14 6.63 34.53 1.48
5.21<br>tbray15 4.12 25.14 2.38 6.10<br>tbray16 3.16 16.15 3.10 5.11<br>tbray_tuple 2.28 8.61 4.30 3.78<br>tbray_ets 1.87 7.42 5.24 3.97
<br>tbray_blkr 1.55 7.34 6.32 4.74<br><br>tbray5/real is the speedup versus the baseline, while<br>user/real is the speedup for each version due to<br>parallelization.<br><br>Thus, the latest version is
6.3 times faster than the<br>first one. The parallel speedup is about the same in<br>tbray5 and tbray_blkr, a very decent utilization of<br>>50%, but the amount of work (user) has shrunk from<br>(presumably more than)
34.53 seconds to 7.34 seconds.<br><br>Tim Bray's original Erlang number on "his macbook"<br>appears to be 34.16 seconds user (probably about the<br>same real?). How does this compare to Ruby? Tim Bray<br>reported that it needed
3.46 seconds real, again on<br>his macbook. (As I understand it, all results here are<br>for the big data set.)<br></blockquote></div><br><div>Yes, all results are for o1000k.ap, Tim's original large dataset. As for Ruby, I just ran Tim's original code on the 8-core, and out of ten attempts the best was:
</div><div><br class="webkit-block-placeholder"></div><div><div>real 0m2.210s</div><div>user 0m2.095s</div><div>sys 0m0.109s</div><br class="webkit-block-placeholder"></div><div>--steve</div>