I did the same test again with a MS6905-Master Ver.2 that I received. The previous test was with Ver2.3. All other things remained the same.
The system ran for 4 and a half hours with the 3DMark2001SE Demo and game, again with that C3 Nehemiah at 6.0x133=800Mhz. Then after 4 and a half hours it had locked up. I am sure around 4 hours it was still running. Not bad, but not flawless.

Below the slotkets, with the jumper settings that were used for the tests.

I guess the 3 pin J2 (reserved) remains a mystery. 😕

tpowell.ca wrote:

I guess the 3 pin J2 (reserved) remains a mystery. 😕

I will try to trace it when I have some time.

This is pretty neat - a VIA C3 on a 440BX. Has anyone tried dual CPU with the Nehemiah? I realise SMP is not supported, but VIA had originally planned to SMP support, that is, according to a spokesman I conversed pre-production.

feipoa wrote:

This is pretty neat - a VIA C3 on a 440BX. Has anyone tried dual CPU with the Nehemiah? I realise SMP is not supported, but VIA had originally planned to SMP support, that is, according to a spokesman I conversed pre-production.

Funny. I was wondering the exact same thing yesterday. 😀

I would like to add to that, testing a dual Ezra-T configuration.

Also, don't know if this is a new record but...

That's right. A 440BX motherboard with an Ezra-T at 50MHz FSB and a 3x multiplier, and stable.

You might be the first one to take a screenshot of it, but not a new record because I've done this before too. 😀

Make sure to add L1D, L2D, ICD, and BPD on top of this for way more slowdown.

Tenorman wrote:

You might be the first one to take a screenshot of it, but not a new record because I've done this before too. 😀

Make sure to add L1D, L2D, ICD, and BPD on top of this for way more slowdown.

I did that. Was you score even lower? 7.9x was the lowest I could get without using any throttling or using BIOS settings.

7.02 was my lowest.

tpowell.ca wrote:

I did that. Was you score even lower? 7.9x was the lowest I could get without using any throttling or using BIOS settings.

You win.

I get 8.37 running at 50 * 3 with BPD, ICD, L1D, and L2D. I could go in the BIOS and start messing with things like RAM timings, but it probably isn't worth the effort.

I also posted my score at 1.0 Ghz for comparison. Quite the range.

Tenorman wrote:

You win. […]
Show full quote

tpowell.ca wrote:

I did that. Was you score even lower? 7.9x was the lowest I could get without using any throttling or using BIOS settings.

You win.

I get 8.37 running at 50 * 3 with BPD, ICD, L1D, and L2D. I could go in the BIOS and start messing with things like RAM timings, but it probably isn't worth the effort.

I also posted my score at 1.0 Ghz for comparison. Quite the range.

IMG_0396.JPG

IMG_0397.JPG

Interesting. Considering we're using the same cpu settings, that means your board/BIOS settings combo are more optimized than mine (for top speed).
I'm using CL2 PC133 RAM with optimized defaults, precharge control enabled, delayed transaction enabled.
I wonder if having less RAM might have a positive effect on cache hit accuracy on these CPUs with so little L1/L2 cache.

Guys, L1D+BPD gives you a slower speed than using all four L1D+L2D+BPD+ICD - that is what I used with my 50*3 to get 7.02.

infiniteclouds wrote:

Guys, L1D+BPD gives you a slower speed than using all four L1D+L2D+BPD+ICD - that is what I used with my 50*3 to get 7.02.

So RAM size has no effect on SpeedSys CPU scores, but infiniteclouds is right. L1D+BPD is slower than L1D+L2D+BPD+ICD.
Very odd indeed.

And at 994MHz, my score is 520.08 making it basically the same as Tenorman's.

Tenorman, what BIOS version are you using?

I am using the Powerleap BIOS from the driver pack on Vogons Drivers:
5)powleap.bin
Source: PowerLeap Tualatin CPU adapter, by esupport.com 'the BIOS experts'
Package: Powerleap_mr_6bxc_beta

It looks like you are using the same one, judging from the date and "Award v6.00PG".

Using only L1D+BPD, I get 8.16 instead of 8.37. Not a huge difference, but it is indeed slower.

These are my RAM settings. I have not tweaked them manually. I do see that I have Precharge Control disabled and Delayed Transaction disabled.

That's really odd.
Your BIOS settings for CAS latency don't make sense. CAS latency should be either 2 or 3 for SDRAM.

Also, SpeedSys in your screenshots is able to identify your external clock. That line is missing from my SpeedSys results.

Agreed that this is odd. Maybe I will try manually configuring it sometime and see what happens. Either that or reloading the defaults. It seems to work though so I'm not too worried about it.

I will double-check that we are using the same version of SpeedSys later. I am running it in MS-DOS mode, but EMM386 is loaded, so that could also be a difference.