[Top][All Lists]

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [DotGNU]Could someone test this on all platforms ?

From: Gopal V
Subject: Re: [DotGNU]Could someone test this on all platforms ?
Date: Fri, 7 Feb 2003 19:39:02 +0530
User-agent: Mutt/1.2.5i

If memory serves me right, Paolo Molaro wrote:
> I guess in the computer science courses they don't teach how measuring is
> supposed to be done? ;-)

We're mostly taught that there are "lies , damned lies and stastics"
and told to extrapolate that theory to benchmarks and release dates ;-).

> 1) the first DateTime.Now call will make the jit compile lots of methods
> (initialization of some types and so on). 

Hmm...This also applies for ilrun (CVM conversion, unrolling, inlining) ?.

> 2) the Millisecond property has, well, a millisecond resolution:-)

Well that's obtained after the two DateTime.Now calls.. So that should
not interfere .

> I just added a "t1=DateTime.Now;" before the loop and increased the
> number of iterations to 10240000. Results are below:

Hmm.. I was comparing the performance of x86 unrolled ilrun doing the
FooBar.Test() and full JIT doing FooBar.Test() .. When I got a higher
speed for the simpler ilrun , I added 1024 runs to balance out any
effects of conversion.

This looks a lot like the code I'm working on , where a for loop over 
the entries checking for true vs false ... (it now has ~7000 iterations, 
but 1024 was a decent approximation).

> about two times slower than mono in your test.

hehe, after you doctored it to give that :) ....

I was just playing around with the unroller when I checked this . This 
was a simpler case to check whether my computer had gone haywire .  My 
comparison led to the last mail.

Well, let's get back to work ...

The difference between insanity and genius is measured by success

reply via email to

[Prev in Thread] Current Thread [Next in Thread]