Another Sandra issue....

[oceaneer]

Hi All....

Played with Sandra last night, and she said that I was running a 23mHz PCI bus speed, or a 1/4 modifier. My BIOS says 1/3, or 33mHZ (I am running 6x90FSB) What's up? (She had the AGP speed right)

Also, she said my WD Caviar was not putting out, so to speak. How can I tell if UDMA 33 is being used by the drive?

Thanx

------------------
Celeron 400@540 on ABIT BE6 r2, 96M RAM, G400 32m DH oem, voodoo II SLI, Montego A3D Xstream, 300W PS

[Jake]

Hi,

I'm getting the same results with wrong PCI bus freq. divider in Sandra. I think it is a problem with Abit SoftMenuIII boards (BF6,BE6-2). I'm not sure if it's Sandra that detects the divider wrong or SoftmenuIII that programs the divider wrong. It only happens at off-spec FSB speeds. It happens to me when I try to run 124 Mhz and I manually force a divider of 1/3. My peripherals are able to handle 41Mhz bus speed but sandra reports it to be 31 Mhz. It could be an error in SoftmenuIII that selects the "best" divider no matter what you choose in manual setup...

Anyone know of a program actually able to detect correct PCI speed or divider or these boards?

Regards,

Jake

[Brian R.]

Either you guys see my post?
http://forums.murc.ws/ubb/Forum3/HTML/001729.html

Any comments?

B

[Jake]

Comments? What post?

There is nothing in it! Could you post it again here, please.

Cheers,

Jake

[Brian R.]

Post? What post? We don't need no stinking posts!

I have no idea...

Anyway, there was a post...which said something like:

I have a BE6-2 mb and I have tested it with oc'd fsb setting applied through both the Softmenu and the mb dipswitches. On my mb, using a 112 MHz fsb (PIII 600), the dipswitches give a much faster computer than the Softmenu fsb setting of 112 MHz, or anyother Softmenu setting for that matter. Not a small difference, but more like 40% faster or so.

It's not that the dipswitches are fast, it's the Softmenu settings that are much slower than normal. I was hoping someone would check their mb to see if anyone else could duplicate this. You don't need a finely tuned benchmarking program to detect this, any program which tests your processor speed with calculations will do. I used SETI calculations as a benchmark. Even 3DMark showed this difference. In fact, with 3DMark, it showed that a higher fsb slowed down the processor when applied with Softmenu settings.

PS. I have the latest bios installed for the BE6-2.

Let me know your results if you have a chance to compare the Dipswitch fsb settings with the same ones in the SOftmenu.

[This message has been edited by Brian R. (edited 13 April 2000).]

[Rags]

Hi Brian,

I have a BF6, and if I get time tonight, I will check that out.

Rags

[Jake]

Thanks Brian!
Of course we don't need no stinking posts!

This is very interesting. I will certainly look into it by doing some tests myself. I tried using the dip-switches when I first got this board, but dismissed it as it is not possible to run off spec fsb speeds by using the dips. (125-132). But as my PIII-450 is not stable at 600 anyway, maybe I should try running 124 using the dips (Computer nerd lingo for dip-switches. I like it! Does that make me a nerd?). How about you guys run some Sandra benchmarks (memory, multimedia, HDD test and CPU) and post the results on this thread for comparison? 3Dmark 2000 CPU test results too if possible. That would make me very happy! (I just love benchmarking!).

On a sidenote I e-mailed with Adrian (Creator of Sandra) and he told me that the QJ version bios has a different bios ID string which causes lack of information in Sandra's motherboard page. I don't know if this also is the cause of the 'wrong FSB to PCI divider detection'. It could be...

Unfortunatly my granddad passed away monday this week and I'm going to his funeral this weekend, so testing and benchmarking on my part will have to wait until sunday evening or sometime monday. Have a nice weekend ya'all!!

Cheers,

Jake


------------------
Who is General Failiure and why is he reading my drive?
----------------------
MGA-G400DH 32Mb Mill., Abit BF6 mobo bios ver. QJ, PIII-450@558, 128Mb PC-133 SDRAM, 17" Hitachi monitor, Plextor 40TS CDROM, Plextor 8/20 CDR. Diamond MX300 A3D PCI soundcard, Adaptec AHA-2940U2W, Seagate Cheetah 9.1 GB LVD HDD, Quantum Fireball ST 6.4 GB UDMA33 HDD

[Brian R.]

Jake - Here are my Sandra results:

600 @ 672 (Dips)
CPU 1815 MIPS
FPU 900 MFLOPS

600 @ 672 (Softmenu)
CPU 1828 MIPS
FPU 907 MFLOPS

Evidently, this CPU benchmark is not reflective of the slow performance I've seen.

The time to perform a SETI wu is slower by 40% or so when softmenu is used for fsb setting.