K6-2 OC, and questions

[NewBlackDak]

I got a replacement K62-400(CTX core)for the one I burned up. Cost me a whoppin $17.
Installed it, and immediately went for the OC(it's a sickness, I know). Well it's running at 560(112x5)@2.6v. Pretty good. Other system specs are Epox MVP3G-2, 128MB PC-133 RAM, Voodoo3 2K OCed to 173MHz, WD 10-Giger, slow el-cheapo DVD/burner combo.
3D Mark 2K got 1472 Marks. Is this about right?

A question though. This board has the 1MB L2 cache on it. Sandra see it as 1024kB write through cache in motherboard information. However in CPUinfo, and in h-oda's WCPUID it says I have no L2 cache. The latest chipset drivers have been installed, and and I've tried cache on/off fast/fastest settings in BIOS. None of these changes what Sandra and wcpuid tells me. Is there something I have to do in the registry to enable the L2, or somthing else? Any help is appreciated.

Thanx NBD

------------------
Who made that man a gunner.
I did sir, he's my cousin.
Who is he?
He's an ******* sir.
I know that, what's his name?
That is his name sir, *******, Major *******.
And his cousin?
He's an ******* too sir. Gunners Mate First Class Phllip *******

[deltree_windows]

would it be referring to the on-chip L2 cache which the k6-3 had? the 1MB on board was called L3 in that case

------------------
Spirit- Blind kids can find lost kittens too

Ripcord- Maybe you stink at baseball because you need glasses

Now you know - And Knowing is Half the Battle!

[Todd a]

Sorry the oldest I got on the online results brouser is 4000 for my Voodoo3 2000 @166 and a Duron 600@900. According to some old benchmark list I was able to dig up I ran about 1700. I was using a K6-2 450 running at 112x4=448Mhz. The Voodoo3 2000 was running at the same 166Mhz as above. The memory was 2 sticks of 64MB PC133 locked to AGP speed (75Mhz) at cas2 (very unstable at 100Mhz or higher). I had a lot of Quake2 benchmarks too, playing with the 3Dnow! patch running Massive1 and Crusher demos. Those were the days.

------------------
AthlonXP 1600+ 1.65v 42'C
FOP32-1 Mod w/80mm 30db
Epox 8KHA+ KT266A
256MB Crucial DDR266 cas2
GeForce4 Ti4200 (300/551)
Sound Blaster Audigy
Antec SX-630 Case (mod'd)
TTGI TT-350-SS 370w PS

[fallguy]

NBD,
I don't know if this will help alot, but as an interesting point I always found that the memory benchmarks in Sandra went up slightly if I disbled the L2 cache, and down when it was enabled.

Its not exactly a definative test, but it is something you might look at.

Fallguy

[Todd a]

That is possible. The CPU will not waste cycles looking in the L2 cache, but for most applications the L2 is a significant boost in performance for these systems.

------------------
AthlonXP 1600+ 1.65v 42'C
FOP32-1 Mod w/80mm 30db
Epox 8KHA+ KT266A
256MB Crucial DDR266 cas2
GeForce4 Ti4200 (300/551)
Sound Blaster Audigy
Antec SX-630 Case (mod'd)
TTGI TT-350-SS 370w PS

[dfgsd]

i am sure wcpuid shows no l2 chace becasue it is lookin for it on the chip not on the board.


fallguy did u have k6-2 or k6-3..disableing the onboard l2(l3) on a k6-3 casued an improvemnt i know, but on a k6-2?

[fallguy]

I had a K6-2.

And it did cause a slight increase the artificial memory benchmarks, but a notable decrease in the real world performance. Just another hint that benchmarks that rely on bandwidth do not entirely correlate to real world performance.....

Fallguy