[Ur] Big drop in Ur performance in latest web framework benchmark

[Adam Chlipala]

On 04/01/2017 05:22 PM, Andy wrote:
> Prelim results for the latest web framework benchmark is out 
> (https://www.techempower.com/benchmarks/previews/round14/) and the 
> performances of Ur suffer pretty some big drops in a few tests 
> (Fortune Single Query, Multiple Query - but improved significantly in 
> Data Updates). Any idea what causes that?

The number of cores per machine has changed, and my best guess is that 
the Ur/Web benchmark's solution hasn't been tweaked properly to set its 
number of threads as a function of the number of cores. (Now there are 
quite different core counts across the three machines in the study, 
while the counts used to match.)
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://www.impredicative.com/pipermail/ur/attachments/20170408/a3f9da13/attachment.html>