Re: Too good to believe?

Herbert Wengatz 2850 (herbert.wengatz@ae3.hypo.de)
Mon, 25 May 1998 14:00:50 MET DST

Richard Gooch wrote:
+> root@terrorist.math.ntu.edu.tw writes:
+> >
+> > I had this e-mail exchange with a vendor in a hardware newsgroup-- but
+> > it was a group about ASUS motherboards, not CPUs or linux. Since this
+> > sounded far too good, I am wondering if anyone can confirm or deny the
+> > numbers here [`me' is myself speaking, `him' is Rick Lindsay, a vendor
+> > of decent reputation on the net].
+> >
+> > him1: Complete Linux kernel compile times:
+> > him1: Dual P2-300, 64mb SDRAM, 75 seconds
+> > him1: Dual PPro166/512, 64mb EDO, 45 seconds
+>
+> I have trouble believing this. I have a dual PPro180/256 system (Tyan
+> Titan Pro) with 128 MBytes RAM, and it takes me a few minutes to
+> compile a kernel. And I do:
+>
+> % make MAKE="make -j 3"
+>
+> to make sure my CPUs aren't idle. Granted I have smaller L2 cache, but
+> I don't believe that will change the time by more than 20% or
+> 30%. Certainly not several times faster!
+>
+> Regards,
+>
+> Richard....

I have trouble believing this, also. I myself own a dual PPro Board with
two 180 MHz CPUs overclocked to 233 MHz with 128 MB RAM (60ns EDO) and
I reach a compiletime of about 2 Minutes 30 seconds or so with a

make -j (To have the make start the machine as many processes as possible
at the same time. The overall compiletime seems not to differ
much from a "make -j 3" on my machine. I guess the amount of RAM
has a large influence here.)

What in fact may be a major improvement to the compiletime is the cachesize.
My PPros have "only" 256 Kb each; the above mentioned system seems to have
double the amount - but it still seems to be somewhat extremely short to me.

Regards,
Herbert
________________________________________________________________________
Herbert Wengatz, ConSol * GmbH | Private email: hwe@muc.de.NOSPAM
Job: Herbert.Wengatz@ConSol.DE.NOSPAM | "Excellence is a moving target."
Remove the ".NOSPAM" if you want to email me.

-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.rutgers.edu