Alan,
On 1 September 2006 at 17:11, Alan King wrote:
|
| <br><font size=2 face="sans-serif">Most banks and computer companies look
| to SPECfp2000 in order to rate computer and networking equipment for financial
| computations. But there are no financial codes in SPECfp2000. </font>
| <br>
| <br><font size=2 face="sans-serif">Now, it takes a big effort to get a
| proof of concept performed on a bank's proprietary software. IMHO
| this inhibits progress in the financial industry. A project based
| on QuantLib would leverage the open source community to get over this problem.
| </font>
Yup and yup.
| <br><font size=2 face="sans-serif">For these reasons, we are building a
| benchmark based on QuantLib. We will run it on a variety of platforms,
| including IBM's Blue Gene. I can't *guarantee* that IBM would contribute
| this code, but our recent history suggests that IBM is very open to doing
| that (Apache, COIN-OR, etc). </font>
| <br>
| <br><font size=2 face="sans-serif">So --- is there any community interest
| in a QuantLib Benchmark Project? </font>
Nice idea! [ I also tend to run the bermudan swap pricer as an openMosix test
on my Quantian (
http://dirk.eddelbuettel.com/quantian/) builds. ]
As a first approximation, building QL and running the unit tests should work.
It may make sense to continue this on the QL developer (rather than user)
list. And, if I may, please don't post html mail.
Cheers, Dirk
--
Hell, there are no rules here - we're trying to accomplish something.
-- Thomas A. Edison