Messages in this thread | | | Date | Sat, 27 Jun 1998 11:24:13 -0700 | From | David Heitmann <> | Subject | Re: bogomips in 2.0.34 |
| |
Kurt Garloff wrote: > > On Tue, Jun 23, 1998 at 08:16:53PM +0200, Henrik Edlund wrote: > > On Mon, 22 Jun 1998, David Heitmann wrote: > > > bogomips for my Pentium 200/MMX had gone from 398.95, > > > down to 299.01. Has anyone else noticed this, or have > > > any ideas what the cause might be? > > > > That's funny. My BogoMIPS increased 16 points from .33 to .34... :-) > > BogoMIPS may depend on the actual alignment of the loop.
But the code in arch/i386/lib/delay.S uses the ENTRY macro that is defined in include/linux/linkage.h, which, if I'm reading correctly, is using an ".align 16, 0x90" assembler directive in my case. Why would the code be misaligned? Is there a problem with the assembler or the linker?
That is kind of scary to think that there could be a 25% decrease in instructions executed per unit time regardless of the .align directive. What if some important piece of code that gets executed a lot happens to be poorly aligned? Could I see a major difference in kernel performance?
> > -- > Kurt Garloff, Dortmund > <K.Garloff@ping.de> > PGP key on http://student.physik.uni-dortmund.de/homepages/garloff
- To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majordomo@vger.rutgers.edu
| |