rec.autos.simulators

by **Richard Mile** » Tue, 13 Aug 1996 04:00:00

I thought some of you might find this interesting...

Firstly please treat this as 'thinking out loud' rather than some
arsehole who thinks he knows everything about CPUs, motherboards, vid
cards, memory etc. I know that these figures are extremely dodgy and
subject to many variables. Where many results have been received the
highest performance scores have been used. However the CPU/MHz ratings
are actually very accurate. Please do not flame me to death, instead
contribute some results and refine the figures.

Over the last couple of weeks I've been keeping a note of GP2 log:on
processor and video results for different types of CPU and Vid Card.

By simply dividing the processor score by the processor MHz you can
arrive at a fairly consistent figure of GP2 CPU power/MHz..

Here are some results:
AMD 5x86 - 1.376
Cyrix 5x86 - 1.775
PPro - 1.895
Pentium - 1.94
Cyrix 6x86 - 2.68 * **

*Note: Calculated from actual MHz. If calculated from Pxxx+ rating the
score is 2.15.

**Note: This is based on only one score for a P166+. Thus other scores
are required to confirm this high rating.

Now onto Video Scores:

Stingray 64/V - 417
S3 with S3spdup - 474
Millenium - 487
ET6000 - 500

Millenium - 637

Observations:
PPro with fastvid yields excellent PCI video performance. PPro has
mediocre CPU/MHz rating.

Cyrix 6x86 has fasntastic CPU/MHz performance with identical video
performance. Unlike Quake it seems that GP2 is not heavily FPU
dependant.

Speculation:
Summarise overall system performance by adding CPU score to video score
(equally weighted).

PPro 200, Millenium = 1016 (real score)
Ppro 200, ET6000 = 1033

P200, ET6000 = 888

6x86 P166+, ET6000 = 857
6x86 P200+, ET6000 = 970 (75Mhz bus)

Conclusions:
None yet as some more 6x86 scores are required before I am convinced to
buy one. PPro does in reality provide highest scores (at a price). 6x86
could provide best performance/$$.

by **Robert Mull** » Tue, 13 Aug 1996 04:00:00

>I thought some of you might find this interesting...
>Firstly please treat this as 'thinking out loud' rather than some
>arsehole who thinks he knows everything about CPUs, motherboards, vid
>cards, memory etc. I know that these figures are extremely dodgy and
>subject to many variables. Where many results have been received the
>highest performance scores have been used. However the CPU/MHz ratings
>are actually very accurate. Please do not flame me to death, instead
>contribute some results and refine the figures.
>Over the last couple of weeks I've been keeping a note of GP2 log:on
>processor and video results for different types of CPU and Vid Card.
>By simply dividing the processor score by the processor MHz you can
>arrive at a fairly consistent figure of GP2 CPU power/MHz..
>Here are some results:
>AMD 5x86 - 1.376
>Cyrix 5x86 - 1.775
>PPro - 1.895
>Pentium - 1.94
>Cyrix 6x86 - 2.68 * **
>*Note: Calculated from actual MHz. If calculated from Pxxx+ rating the
>score is 2.15.
>**Note: This is based on only one score for a P166+. Thus other scores
>are required to confirm this high rating.
>Now onto Video Scores:

>Stingray 64/V - 417
>S3 with S3spdup - 474
>Millenium - 487
>ET6000 - 500

>Millenium - 637
>Observations:
>PPro with fastvid yields excellent PCI video performance. PPro has
>mediocre CPU/MHz rating.
>Cyrix 6x86 has fasntastic CPU/MHz performance with identical video
>performance. Unlike Quake it seems that GP2 is not heavily FPU
>dependant.
>Speculation:
>Summarise overall system performance by adding CPU score to video score
>(equally weighted).
>PPro 200, Millenium = 1016 (real score)
>Ppro 200, ET6000 = 1033
>P200, ET6000 = 888
>6x86 P166+, ET6000 = 857
>6x86 P200+, ET6000 = 970 (75Mhz bus)
>Conclusions:
>None yet as some more 6x86 scores are required before I am convinced to
>buy one. PPro does in reality provide highest scores (at a price). 6x86
>could provide best performance/$$.

This is a cool concept and one that I am very interested in but is it
possible to throw in some frame rate/processor occupancy figures as
well. I think these give you a better feel for true performance. I
think the log:on is a flawed approach due to watching it increase over
the course of video card upgrades and overclocking on my system.

by **Badrul Hish** » Wed, 14 Aug 1996 04:00:00

>>I thought some of you might find this interesting...

[Snip]

>>Conclusions:
>>None yet as some more 6x86 scores are required before I am convinced to
>>buy one. PPro does in reality provide highest scores (at a price). 6x86
>>could provide best performance/$$.

>This is a cool concept and one that I am very interested in but is it
>possible to throw in some frame rate/processor occupancy figures as
>well. I think these give you a better feel for true performance. I
>think the log:on is a flawed approach due to watching it increase over
>the course of video card upgrades and overclocking on my system.

I collected a lot of log:on's results and having examined them, I agree
with you that the numbers don't really represent the best performance of
GP2 for any particular system. The estimated fps rate is a more accurate
representation of GP2 performance in the system as it takes in both
processor speed & video card to come out with such a value.

The problem is people report fps rates but with different level of
graphic details & features turned on & off, so it's a bit hard to
really compare performance. That will also apply to occupancy - the figures
were reported from various tracks.

So if we could agree on a standard level of graphic details & decide
a part of a track for the value of occupancy, then we can start comparing
performances.

regards,
hisham

rec.autos.simulators

GP2 log:on - comparison of 6x86, P5, PPro etc...

GP2 log:on - comparison of 6x86, P5, PPro etc...

GP2 log:on - comparison of 6x86, P5, PPro etc...

GP2 log:on - comparison of 6x86, P5, PPro etc...