Date: Tue, 15 Apr 1997 23:46:25 -0400 (EDT) From: Thomas Arnold <tom@inna.net> To: Doug Russell <drussell@saturn-tech.com> Cc: Chuck Robey <chuckr@glue.umd.edu>, Stephen Roome <steve@visint.co.uk>, Dave Alderman <dave@persprog.com>, hardware@FreeBSD.ORG Subject: Re: Pentuim or Pentuim Pro ? Message-ID: <Pine.BSI.3.95.970415233558.10274B-100000@caught.inna.net> In-Reply-To: <Pine.BSF.3.95.970415211641.12530B-100000@586quick166.saturn-tech.com>
next in thread | previous in thread | raw e-mail | index | archive | help
On Tue, 15 Apr 1997, Doug Russell wrote:
>
> Oh... By the way. Has anyone done any benchmarks on the K6 chips under
> FreeBSD yet? If not, let me know what tests you all want done on one,
> because I'm going to be picking up my first K6 for evaluation tomorrow.
> It's only a K6-166 chip, and I'll be running it in an ASUS HX board, but
> it should be some kind of indicator as to what the things can do.
Gee. Ask and ye shall receive!
( Did these last night about 5 minutes after swapping my K5-133 for a
K6-166 ).
Chip : AMD K6-166
Motherboard : Machspeed VX based. 512k Cache. Award BIOS flashed to
support K6-166
FreeBSD : v2.2GAMMA
Set for 66mhz and 2.5x multiplier
Sieve of Eratosthenes (Scaled to 10 Iterations)
Version 1.2b, 26 Sep 1992
Array Size Number Last Prime Linear RunTime MIPS
(Bytes) of Primes Time(sec) (Sec)
8191 1899 16381 0.007 0.007 225.2
10000 2261 19997 0.009 0.009 225.8
20000 4202 39989 0.018 0.018 227.3
40000 7836 79999 0.036 0.068 122.9
80000 14683 160001 0.072 0.244 69.0
160000 27607 319993 0.144 0.537 63.5
320000 52073 639997 0.288 1.288 53.5
640000 98609 1279997 0.575 3.371 41.3
1280000 187133 2559989 1.151 9.341 30.1
2560000 356243 5119997 2.301 21.310 26.6
Relative to 10 Iterations and the 8191 Array Size:
Average RunTime = 0.029 (sec)
High MIPS = 227.3
Low MIPS = 26.6
FLOPS C Program (Double Precision), V2.0 18 Dec 1992
Module Error RunTime MFLOPS
(usec)
1 -4.6896e-13 0.4936 28.3621
2 2.2160e-13 0.3550 19.7199
3 -6.9944e-15 0.3188 53.3219
4 -9.7256e-14 0.2887 51.9496
5 -1.6542e-14 0.6858 42.2846
6 4.3632e-14 0.4933 58.7919
7 -4.9454e-11 0.8783 13.6629
8 7.2164e-14 0.5234 57.3160
Iterations = 32000000
NullTime (usec) = 0.0120
MFLOPS(1) = 24.8368
MFLOPS(2) = 26.2366
MFLOPS(3) = 39.6526
MFLOPS(4) = 56.0262
I am curious what causes nsieve to be so inconsistent across its spectrum.
This is the same behavior the Cyrix 6x86's show. If you look at Tom
Balfe's message from a couple days ago you'll see what the K5's I run do
on these boards. The K5's seems to be VERY consistent. Although the K6
is a screamin machine, the benchmarks bother me. I talked to Machspeed
today and am going to try some motherboard tweaking, but I'd really have
to know more of what would affect nsieve like that.
+-----------------------------------------------+
: Tom Arnold - No relation to Rosanne :
: SysAdmin/Pres - TBI, Ltd ( inna.net ) :
: The Middle Peninsula's Internet Connection :
+-----------------------------------------------+
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?Pine.BSI.3.95.970415233558.10274B-100000>