Subject: Re: ppc benchmarks, quick and dirty. 604ev, g3, mips
To: John Klos <john@ziaspace.com>
From: Michael <macallan18@earthlink.net>
List: port-macppc
Date: 03/14/2005 11:11:02
--Signature=_Mon__14_Mar_2005_11_11_03_-0500_nkCkXYVsddfMbS39
Content-Type: text/plain; charset=US-ASCII
Content-Disposition: inline
Content-Transfer-Encoding: 7bit

On Sun, 13 Mar 2005 16:35:49 -0800 (PST)
John Klos <john@ziaspace.com> wrote:

> Just a really quick follow-up. Using pkgsrc/benchmarks/flops (which 
> specifically says that it's small enough for any cache), I did a few 
> comparisons. For simplicity, I am using units of mflops per 10 MHz since 
> the clock speeds are all over the place. Also, flops includes four 
> results; I'm using results 1 and 4, since 1 includes fdiv and 4 does not.
> 
> G5:
> 3.3, 3.42  (OS X)
> 
> G4 7457:
> 1.95, 3.39  (NetBSD 2.0)
> 1.96, 3.53  (OS X)
> 
> G3 750FX:
> 2.27, 4.94  (OS X)
> 
> G3 750:
> 2.11, 4.87  (NetBSD 1.6)
> 2.11, 4.73  (NetBSD 2.0)
> 
> 604ev:
> 2.4, 6.05   (NetBSD 1.6)
> 
> 603e:
> 1.94, 4.2   (NetBSD 1.6)
> 2.03, 3.91  (NetBSD 2.0)
> 
> m68060:
> 1.3, 2.6    (NetBSD 1.6)
> 
> m68040:
> .88, 1.16   (NetBSD 1.6)
> 
> So yes, the 604 beats out everything except the G5, and even beats out the 
> G5 in floating point code without fdiv.

Just for fun I ran the same test on a few more boxes, all compiled with CPU-specific optimizations enabled and -O2. Here are the results:
UltraSPARC IIi, Sun U10:
2.99, 3.98	(NetBSD 2.99.16, 64bit)

MicroSPARC II, Tadpole SPARCbook 3GX
1.28, 2.23	(NetBSD 2.99.16)

604e, Motorola PowerStack II:
2.38, 5.82	(AIX 4.3.2, gcc)

G3 750, UMAX Pulsar:
2.10, 4.72	(NetBSD 2.99.16)

MIPS R10k, SGI Indigo2:
4.96, 16.39	(IRIX 6.5.19, gcc -mabi=n32)
5.20, 16.27	(IRIX 6.5.19, gcc -mabi=64)

Athlon XP:
4.61, 7.06	(FreeBSD 4.11)

I ran the test on the PowerStack with xlc too, but even with 'potentially semantics altering' optimization the results differed less than one percent from those obtained with gcc. As expected the MIPS runs circles around everyone else in raw FLOPS per clock, but the Athlon did quite a bit better than I expected. All tests were done with gcc 3.3.[3|4|5].

have fun
Michael

--Signature=_Mon__14_Mar_2005_11_11_03_-0500_nkCkXYVsddfMbS39
Content-Type: application/pgp-signature

-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.0 (NetBSD)

iQEVAwUBQjW3l8pnzkX8Yg2nAQLXEggAvDwuKe1d2Ya9+c/fhXiia4/iSPQF67zn
X15k6CDnlX3Ya81LOxazAhV4wYk5x3sIIoHwdyez0V8P1Yzf9RSEj+upsO0WfYD8
WFQu1PWJOku9JfZFE5m+yTZqIqGZHkaj6bEdhVJF4uTtFbe++VmkSrbuPV2GFQsV
nmHQkuYCqaIWDLDIghBT5ac2U3Bap4Xsrndjv6UaJzPGT5+k0LE2JNZQMXTj9Yvt
3NhOODMmCyrbP5KoPqcNAvsdOJjx/SnyYTzmn9bK5yz2vrS5w2/IomcvjK8keSjJ
w3GatCpFrgxJDRXiA+QhATpBzVvbRyMOcK/18EfJAcH50xUBzzGLXg==
=fmWj
-----END PGP SIGNATURE-----

--Signature=_Mon__14_Mar_2005_11_11_03_-0500_nkCkXYVsddfMbS39--