home *** CD-ROM | disk | FTP | other *** search
- In article <3euq3f$r1r@xmission.xmission.com>, dingebre@xmission.com (David Ingebretsen) writes:
- > CamCollect (camcollect@aol.com) wrote:
- > : LightwavePro, December 1994 issue, The PC Primer, Mojo writes
- > : :
- > : "At the 'low' end of the spectrum, the older yet mature Intel 486 (66MHz)
- > : is roughly twice the speed of the 40MHz 040 Warp engine the fastest Amiga
- > : available.>>>>>>>>The DEC Alpha:..One of these machines...is calculated at
- > : 18 times faster than a 33MHz Amiga 040."
- >
- > : Where did you get these numbers? According to my calculations the 486
- > : DX2-66 runs 20.5 MIPS and my slow A2000- 68040-33MHz (GVP) runs 21.5 MIPS.
- > : (my A4000/040/40MHz Warp is about 35% faster than the 33MHz.)
- >
- > : I've found the DEC Alpha 275MHz. to be 10 times faster than the 33MHz
- > : Amiga 040. These tests are of an average scene, no ray tracing. Scenes
- > : using raytracing will see even faster times with the DEC Alpha.
- >
- > I've seen results and have participated in informal benchmarking comparing
- > a 90Mhz pentium, 275 Mhz Dec alpha, and various Amiga configurations. From
- > what I saw, the 90 Mhz pentium was just more than twice as fast as my 33
- > Mhz 040. the Dec Alpha came it at around 10 to 20 times faster than my 33
- > Mhz 040. These test were done using ray-tracing, by the way.
- >
- > I can't believe that if the 90 Mhz Pentium is about twice as fast as my
- > 040, that the 66 Mhz 486 is also twice as fast.
- >
- > Yes, yes; I know what you're thinking. The benchmarking was NOT done with
- > Lightwave. It was done with Real3D (I certainly don't have LIghtwave on a
- > PC, do you?).
- >
- > When I was involved in benchmarking systems for some real-time vehicle
- > dynamics, I quickly learned that that the only true test for speed is
- > running the code on the different systems. One can compare MIPS, MFLOPS,
- > and all the other "benchmarks" on paper til you die of boredom. Until you
- > take the actual compiled code, and run carefully constructed test cases,
- > you won't be able to really compare speed.
- > --
- > David
- >
- > David M. Ingebretsen *** Binary Illusions / 3D Physics ***
- > dingebre@xmission.xmission.com *** Animation and more ***
-
- I read somewhere that Real3d was never optimized for the 040, which is why the
- benchmarks are unbalanced.
-
- AC
-
-
-
-
-
- kr
-
-
-
-