64-bit/66MHz SCSI Goodness is GOOD FOR YOU!!!

[MichaelD]

Originally posted by: TerryMathews

Originally posted by: MichaelD
I'm overclocking and running a 150FSB. This doesn't have an effect b/c the NCCH-DL has a working AGP/PCI lock. CPU-Z shows AGP/PCI speeds of 66.6MHz and 33.3MHz exactly, so that's not it.

I cannot run a 200FSB b/c these particular Xeons can't take it. The way it works is that for a split second, the bios powers up the CPUs at DEFAULT multiplier (16x) times whatever FSB you've set. THEN pending a successful boot, it lowers/raises speeds to whatever you've set in the bios.

I cannot post at 12x200 which is the same 2.40GHz I'm running now b/c of the above stated problem.

Did you swap around the CPUs? I've got one that can POST at 16x200 and one that can't. As long as the one that can is in CPU0, you're golden.

Also, did you make sure you were using the 1:1 memory setting and not the faster one? When you jumper and POST at 133FSB, you need to set the memory to 333 to get the proper 1:1 setting at 200. ASUS = stupid BIOS when it comes to that.

And yes, just because you've utilized a PCI lock doesn't mean it's not adversely affecting performance. Running the bus at 66.6MHz and the FSB at 150MHz means they're no longer synchronized - for large transfers the processor is going to have to wait for a 66.6MHz clock cycle to transmit data to/from the SCSI card. It'd be more of a latency issue than a transfer bandwidth one - but it still might be causing some other gremlins.

If you really want to check if it's a hardware problem or software one, download a Knoppix LiveCD that supports your SCSI card and try one of the Linux HD benchmarking programs.

Hmm. I AM running a 4:5 mem/sys ratio. Let me see if I can toy aroudn in the bios w/o killing it. BRB.

*edit*

Well, I changed the memory setting in the bios from Auto to DDR333. No diff. I may try for 200FSB again.

 

[MichaelD]

TTT for CPU SPEED update.

 

[imported_Lucifer]

Though I am not much of a PC person, congratulations for building a sweet computer. I think it was you who posted those pics of the Dual Xeon you built? Very nice. I hope it serves you well. I like it a lot.

Good luck to you!!

 

[MichaelD]

Bump for update. See orig post, top of thread. Now at stable 2.66GHz.

 

[imported_Phil]

Originally posted by: MichaelD
Bump for update. See orig post, top of thread. Now at stable 2.66GHz.

Git :beer:

Nice going, though

 

[TerryMathews]

Originally posted by: MichaelD
Bump for update. See orig post, top of thread. Now at stable 2.66GHz.

If you're at 2.66, $5 says you could be running at 200FSB any time you want to...

 

[Markfw]

Bump for my buddy, good luck..........................

 

[sharkeeper]

I would be more concerned about the i/o than the overclock but that's me...

Cheers!

 

[trevor2k]

MichaelD- Why stop at 166??? My rig is running stable at 2.8 each, and after switching some case fans around, whalaa, 10 degrees cooler....I hit 47 max. Come on, your processors can do it!!! Are you using pc3200? then you can do 1:1.

 

[MichaelD]

Mwuahahaha! All the mad overclockers have checked in.

I still haven't been able to boot into Windows at 200MHz FSB, no matter what. I can BIOS boot at any speed I damn please, but into Windows (where it matters) is a diff story.

It's NOT the memory, as it's Corsair PC3200 CAS2 stuff, that's run at 210 CAS2 before.

I ran at 2.66GHz just fine; until I tried running four instances of P95TT. It errored out in 10 mins. Interestingly, all SANDRA/3DMark/Aquamark/ATTO/gaming benchmarks and real world usage resulted in no errors at all.

But, me being the picky overclocker that I am, if it can't run P95TT at least 12 hours straight, then it's not a stable overclock.

I like my 166FSB, 1:1 ratio; thanks very much! So, I lowered the multiplier from 16x to 15x.

I'm now at 2.50GHz. I will attempt the four instances of P95 overnight and we'll see what happens.

 

[sharkeeper]

Yes you MUST have ALL FOUR virtual CPU in task mangler pegged to the max or your processors are not being stressed.

You should run it for a week to be sure.

Cheers!

 

[MichaelD]

Originally posted by: sharkeeper
Yes you MUST have ALL FOUR virtual CPU in task mangler pegged to the max or your processors are not being stressed.

You should run it for a week to be sure.

Cheers!

Ah, sarcasm from the Master on The Mountain!

I know you hate the fact that I'm overclocking Xeons...but hey, more MHz for the money is where it's at!!! I wish I had $2K to drop on a brace of 3.6GHz, 2MB cache Xeons....but I don't.

 

[Markfw]

Michaeld, I love my opterons ! When I was gone , and they were quiet, I got killed in F@H ! And my 3200@3600 only does 1 out of 2 units withoutpuking !!! (only figured this out tonight) OC'ing doesn't allwayss work ? I am now down to 2.3 ghz on my 90nm OC test (3x HTT, memory is UNDERCLOCKED) 230 at 5:6 but actual rated at 217 !

 

[TerryMathews]

Originally posted by: MichaelD
I still haven't been able to boot into Windows at 200MHz FSB, no matter what. I can BIOS boot at any speed I damn please, but into Windows (where it matters) is a diff story.

Have you tried switching the processors? IE what in CPU1, move to CPU0? That's what I had to do - but now it's 100% stable.

 

[MichaelD]

Bump for anyone that's interested. See orig post. Had to lower my overclock to 2.50GHz. It is now rock stable. P95TT for 17hours, 46 mins. No errors or warnings.

 

[MichaelD]

Some new SCSI results. I've been messing with the settings on the card; change one thing, retest. Put first thing back, change another, retest, etc.

What I finally came up being the best mix of settings kind of surprised me. The card is set to: No read ahead (Normal), Write-Thru (not cached), Direct I/O.


My reads went down 7.3MBs, but my writes went up by 31.3MB/s! A considerable improvement and a good tradeoff IMO. Here's an ATTO screenie of the new results:

New ATTO results

Here's the old ones, for comparisons' sake:

Old results

 

[andnow]

What is the "166 secret"?
Does it work with both the NCCH-DL and the PC-DL?

 

[sharkeeper]

MichaelD:

Despite what ATTO may indicate, you will have better real world performance with write back caching ON, cached i/o and adaptive read ahead ON. Flush time of 10sec is acceptable.

Cheers!

 

[remagavon]

Originally posted by: sharkeeper
MichaelD:

Despite what ATTO may indicate, you will have better real world performance with write back caching ON, cached i/o and adaptive read ahead ON. Flush time of 10sec is acceptable.

Cheers!

What would be the best settings for using this card in a 33mhz pci slot? I have 2 15.3's on the way and am considering picking up one of the Elite's for $200 but I don't have a 66mhz slot to use (for the time being). I know I won't get the most out of it but it will be faster than a single IDE drive that I'm using now

 

[sharkeeper]

Those settings are recommended as listed above for best overall performance.

Cheers!

 

[MichaelD]

The "166 Secret" is to hold down the Insert key upon booting up. This forces a boot at 166MHz with no jumpers on the FSB selector pins on the mobo itself.

There's a bug in the board that even though there is a jumper position (i.e. no jumpers at all, all pins open) it won't boot at 166MHz, unless you hold down the insert key.

AFAIK, it works for the PC-DL as well as the NCCH-DL.

I would consult the PC-DL thread over at the 2cpu.com forums just to be sure; they would know.




RE: The SCSI. The card is in a 64-bit, 66MHz slot.

SHARKEEPER: I may change the settings back; I'm giving it a few days at these new settings to see if I feel a difference; I've not had much free time since I changed them to experiment. Should I consider changing the 10sec flush time to something else? Shorter? Longer? As always, thanks.

 

[sharkeeper]

Try 2 sec flush instead of 10. You'll choose one or the other...

10 second flush is good if you have fast disks but I find with the fastest 15k disks, long flush times with 512MB of cache can make the system pause every once in a while but I'm usually requesting a LOT from the array ALL THE TIME so...

Cheers!