What a bunch of MORONS!

[Crashman]

166MHz at Cas3 usually gives LESS performance than 133MHz at Cas2 So the tolutin was handicapped by being FORCED to use CAS3, even at the higher bus speed. I wanted to adress everyone in the last Tolutin post who claimed that the 166MHz bus was an "unfair" advantage, however there where too many people saying the same thing to adress everyone. Sure it SUCKS that you cannot adjust the multiplier on the Tolutin. But what sucks even more is that Tom did not (or could not, depending on BIOS) use the 166/133 memory timing to allow the system to run at Cas2. Chances are the scores may have been BETTER using 166/133/Cas2.

Cast not thine pearls before the swine

 

[Bud]

So then accoring to your logic, a stick of pc133(cas3) ram, should rather be run as pc100(cas2) ???
That's the logic I'm hearing from you....true??

...I'm just curious.

I thought I was wrong once, but I was mistaken.

 

[Guest]

actually I think that 166mhz cas3 is pretty much equal to 133mhz cas2. besides, the athlon and P3 aren't nearly as bandwidth dependant as the P4.

 

[Ncogneto]

Err Crash, you are forgetting some very important things.

1) This still does not account for the overclcocked AGP bus comparing 3d scores to the p4 and athlon running there agp bus at stock speeds.

2) Does not take into account the increased bandwith allotted to the cpu at an increased FSB ( this
one should be minor however )

3) Are you absolutely sure of your claim that cas2 133 is in fact superiour to cas3 at 166? Somehow I doubt it, but you may be right, it would be interesting to see. Cas 3 has one more wait cycle then cas 2, however the "cycle", when running at 166 is a faster cycle then when running at 133.

A little bit of knowledge is a dangerous thing!

 

[rcf84]

Well i guess true CAS 2 PC150 would work good. Maybe alitte PC2700 (333mhz) DDR-SDRAM test with the the p3-t.

Nice Intel and AMD users get a Cookie.... Yummy

 

[Crashman]

Cas2 at around 105MHz gives about the same performance as Cas3 at 133 in many benchmarks. That's a 26% difference in speed. To get the same results, Cas2@133 would be about the same as Cas3 at 168MHz. So even if the 2MHz difference is not worth dealing with, it proves that the Tolutin did not have a performace advantage do to it's 166@Cas3 memory setting.

Cast not thine pearls before the swine

 

[Ncogneto]

But you are forgetting the overclocked agp bus at least you can surely see that the 3d scores were totally whacked and not a fair comparison. Not trying to start an AMD vs p3 flame fest here either, I addressed this in another post. I still think that the platforms Tom choose for a cross comparison were silly to say the least.
1) Why not start with a t-bird 1.1 and overclock it the same as well useing the same identical ram....this would keep things nearly equall.
2) Where are the p4 scores at 1.3-1.5?
3) why not at least overclock the t-bird to 1.4 gig or use a 1.4 git t-bird MP?

Hey, this looks like a good CPU don't take me wrong, I am ridiculing the tester not the test subject. What we have here is worthless for comparing across architecures as the way the test was conducted. Its only worth is showing us that the new p3 has some headroom and can be overclocked ( if you can manage to find all your pci cards that will run at that speed)

A little bit of knowledge is a dangerous thing!<P ID="edit"><FONT SIZE=-1><EM>Edited by ncogneto on 06/12/01 08:32 PM.</EM></FONT></P>

 

[Crashman]

Overclocking the AGP does not usually increase performance by a significant amount in my testing. But I do know that many CUSL2 users running at 166MHz+ FSB have so that they could increase their memory timing. It's also important to remember that there is a BIOS code for the i815 chipsets that reduces memory to it's slowest settings (a hidden protection feature that you don't see in BIOS settup) and AGP from 4x to 2x when you overclock beyond 140MHz, which affects most motherboards.

Cast not thine pearls before the swine

 

[Ncogneto]

Perhaps but this being a newer reference motherboard that is nothing but pure speculation.

A little bit of knowledge is a dangerous thing!

 

[juin]

Toms was clear that he dont have the time for this and want to release the benchmark as fast as he can.

 

[AmdMELTDOWN]

Tom was in a hurry to get this scoop, he did a good job considering.

more test will soon come up, maybe you'll get to see a 1.13 tully bend over and molest the 1.13 tburnt when Tom is done.

"AMD/VIA...you <i>still</i> are the weakest link, good bye!"

 

[Ncogneto]

Care to place a wager on that? maybe you ought to look closely at the t-bird scores @1.33 and stock fsb compared to the tuly's scores at 1.41 and an overclocked FSB and it is already well proven that the t-bird 1.1 will run easily at 1.4 gig.

A little bit of knowledge is a dangerous thing!

 

[Unolocogringo]

I'v played with the 815 a lot . Right now I run a P3 550e@158 fsb=869 CAS 2/2/2/-7/9 . It's the best my memory can do.
I can run at 160fsb -880 3/2/2-7/9 or 166fsb-914 3/3/3- 7/9.
I also run STUFFPCR to force AGP4X and Fast memory timings and 4/way interleave at these speeds.
Of the above setttings the slowest mhz 869 has the highest scores in 3Dmark and Sandra. This is because the faster memory timings.In Sandra i'm rightt on par with a Athlon 1 gig in the memory benchmark.On mad onion Im number 3 in the P3 853- 902mhz class and on the front page of the of the P3/Athlon 853-902 mhz category.
So a little slower cpu and better memory timings make a big difference.
this is not a flame or a brag ,ok a little , but a statement of my first hand experience and many tweeks and benchmarks to achieve 3279 3Dmark 2k1 score with a
Geforce 2 32mb.
I have to side with Crashman as my experiences are the same as his.

I aint signing nothing!!!

 

[Ncogneto]

The variences in your three FSB settings are only 8 mhz ( 158-166) so yes what would you expect? Of course your fastest memory timings would give you the highest score even at a slighlty less FSB. However, we are talking a difference of 33 mhz not 8, which is 4x as much , things start to even out at this rate.

A little bit of knowledge is a dangerous thing!

 

[Guest]

look the fsb of the processor is overclocked tooooooo like the other guys said. look at an old athlon with fsb 100MHz en fsb 133MHz whats the diference have you seen it???????
i'll answer it its a big difference the processor is a lot faster in every benchmark. so you [-peep-] intelpoepers don'yt say me that it is a fair comparison it would be fair to give the athlon the same ram AND the same fsb and then the total fsb of the athlon would be 333MHz in stead of 266MHz its another prestatie winst of 5%(+-)for the athlon and then at same CLOCK speeds the athlon will leave the P3 behind in a fiar test machine

computer

 

[Unolocogringo]

I was trying to say that these scores are about right. If you had a 1.4 gig P3 running 133 fsb with 2/2/2-5/7 memory timings the scores would still be pretty close to what they are now.Memory is the bottleneck here.If he had slowed down the 8mhz fsb as I do and used 2/2/2-7/9 timings the scores would have been very different. All in all I think this is a close to real world comparison.

I aint signing nothing!!!

 

[Ncogneto]

I still disagree, if he would have taken a 1.1 gig 133 fsb athlon or athlon mp and then increased the fsb the same as the p3 then that would have been a fair comparison. Doing so he could have used exactly the same memory and kept the agp and memory bus at the exact same settings. All this talk aboutan overclocked memory bus and AGP bus not affecting scores "much" is relative.

A little bit of knowledge is a dangerous thing!

 

[lhgpoobaa]

i doubt what he is saying.
my new rig has pc150 kingmax in it and from the sandra 2001 memory benchmark results even cas 3 150Mhz just beats cas 2 133Mhz.

just to be safe though im currently running at CAS 2,2,2 @ 150Mhz *cheezy grin*

i even got my kingmax up to 160 at cas 2, but one of my pci devices kept dying due to the 40mhz bus speed

This behavior must not continue! Feel the BURNING STARE of my HAMSTER... and change your ways!