Re: gratuitous optimization and benchmarking soo 08 Apr 2006 03:56 UTC
* From: "Taylor R. Campbell" <email@example.com>
* Date: Sat, 8 Apr 2006 02:13:36 +0000
* Subj: Re: gratuitous optimization and benchmarking
| I see the disparity now. You are using Scheme48 1.3, from which an
| important compiler optimization of CWV was removed; I was using
| Scheme48 0.57, which still had the optimization.
| Anyway, this thread is getting rather belaboured with irrelevant
| details. My original point remains: performance does not enter into
| the design rationale.
I'd like to know whether it was intentional that you performed
the test with such an old version instead of new one.
Are you sure that VALUES/CALL-WITH-VALUES is always faster than
MU under any circumstances? (The circumstances also include all
implementations that exist at present.) The rationale only said
"... and somewhat slow under some circumstances.". How about "
... and somewhat slow in some implementations."?