OGR folks: Is the 469 client slower for you than the 463 client?

[MereMortal]

I had a 12% drop after I installed the new client! Needless to say I'm back to 463.

900 MHz Tbird benchmarks
463: 7.29 Mnodes/s
469: 6.40 Mnodes/s (5.13A)
469: 5.99 Mnodes/s (5.13B)

Anybody else drop this much? Or am I just having some sort of configuration problem?

 

[ikke]

Yep OGR is still slower on a AMD CPU then then 463 version.

 

[crYnOid]

Yup, my k6-3 dropped about 30% in ogr :Q. But rc5 appears to be a fraction faster. I am still using 463

 

[Nohr]

OGR benched a lot slower using 469 on my Tbird and my K6-III+.

AMD Athlon Thunderbird-C 1377MHz
463b: 11,061,237n/s
469: 9,746,557n/s

AMD K6-III+ 550MHz
463b: 3,583,856n/s
469
Core #0: 2,547,573n/s
Core #1: 2,607,403n/s
Note: 469 autoselected the slower OGR Core #0

 

[MereMortal]

Well, that sucks.

<Nimoy voiceover> in Search of...Faster OGR Cores</Nimoy voiceover>

 

[Viztech]

I was seeing about a 5% loss with 469 on a Celeron 2 on OGR.

What is really disapointing is that we had a 10% increase in OGR going with the 464g pre-release clients on the P2/P3/Celeron CPUs. I don't know what happened there, but it's too bad we're heading in the wrong direction with the OGR cracking speed.

The 463b still is the client to go with on all CPUs on the Windows platform.

viz