Optimization and inlining under MC++

[Steve McLellan]

Hi,

We've got a fairly computationally expensive application compiling under the
/clr option. Most of the processor intensive code is legacy C++. Do the
optimization settings have any effect on code that is NOT compiled with the
#pragma unmanaged directive? We're getting incredibly slow performance from
inner loop stuff that under non-CLR compilation is inlined away. Using
#pragma unmanaged alone has resulted in a big performance increase just for
a couple of functions. Is there any stuff anywhere about how performance is
affected?

Additionally, presumably the #unmanaged directive will also apply to any
templated code (since that's compiled up as it's needed) but NOT any
non-templated code that is compiled without the #unmanaged directive (and
therefore defaults to being converted to IL by the compiler).

Thanks for any help, and sorry if this is a bit hard to read - staring at
profiler output is grilling my eyeballs :-)

Steve