XBR2D96D@DDATHD21.BITNET (Knobi der Rechnerschrat) (12/28/88)
Hallo, to complete my previous requests on the Personal Iris: we made the observations concerning the ethernet performance on a fully equipped 4D/20 (inl. FPU). We also run 'arena' with the same result. Everybody looses on the 4D/20 when playing against a 4D/70G. We switched machines between the players and again the 4D/20-player lost. Regards Martin Knoblauch TH-Darmstadt Physical Chemistry 1 Petersenstrasse 20 D-6100 Darmstadt West-Germany BITNET: <XBR2D96D@DDATHD21>
archer@elysium.SGI.COM (Archer Sully) (12/29/88)
In article <8812280909.aa04738@SMOKE.BRL.MIL>, XBR2D96D@DDATHD21.BITNET (Knobi der Rechnerschrat) writes: > Hallo, > > to complete my previous requests on the Personal Iris: we made the > observations concerning the ethernet performance on a fully equipped > 4D/20 (inl. FPU). We also run 'arena' with the same result. Everybody > looses on the 4D/20 when playing against a 4D/70G. We switched machines > between the players and again the 4D/20-player lost. > > Regards > Martin Knoblauch > > TH-Darmstadt > Physical Chemistry 1 > Petersenstrasse 20 > D-6100 Darmstadt > West-Germany > > BITNET: <XBR2D96D@DDATHD21> Its been my experience that the 4D/20 runs dog at about 12 frames per second, while a 4D/70 runs around 20-25 fps. This results in a significant advantage for the pilot on the '70. The same is true of the GT, since it runs a different version of dog that has different models for the planes. It does about 15 fps. All of these figures are fundamentally unrelated to ethernet performance, but have more to do with fill rates. Dog compensates by adjusting the flight characteristics according to the frame rate. This keeps the '70 from (literally) flying circles around the PI pilot, but doesn't address the problem that one pilot gets twice as many frames in which to react, shoot, etc... as the other. If anyone has a suggestion for this (other than fixing the frame rate so that it runs at exactly the same speed on all workstations), let me know. --