As we all know; the i440BX remains a remarkably efficient chipset - the attached image shows a BP6 at 100MHz FSB compared with various SMP PIII setups/chipsets, most at 133FSB.
SSSandra 'scores' are of course only meaningful as comparisons using the same version: the version used here is: 2004.10.9.89
Optimisations using WPCRSET are [registers:values:meaning]:
52 = 04 = Host Bus Fast Data Ready [enabled]
76 = 0F = 2-2-2 & Leadoff Command Timing = 3
77 = 00 = [probably - reserved/illegal value] do not add delay to DRAM Leadoff Timing
78 = 27 = DRAM Idle Timer = 128 [value 'infinite' didn't work for me]
- they are worth in sum around 8% over maximally BIOS-optimised values & this BP6 remains 24/7 stable using 3 sticks of very ordinary PC133 3-3-3 stuff.
- wbxtune v2.30 would not have set these optimisations, so what's the point?
wbxtune - why bother?
wbxtune - why bother?
You do not have the required permissions to view the files attached to this post.
WPCRSET-optimised as above, a reliable & fairly memory-sensitive benchmark [ScienceMarkMP v1.0] sees around 1.5% improvement overall for this BP6 at 550.
- as a reality-check, this optimised 'score' is just about exactly 2/3 the performance of a uniprocessor Athlon 1200/DDR 2100 platform:
- as a reality-check, this optimised 'score' is just about exactly 2/3 the performance of a uniprocessor Athlon 1200/DDR 2100 platform:
You do not have the required permissions to view the files attached to this post.
Derek - Hi - bought a couple of knackered old BP6' off Ebay for fun, having wanted one for many years - but having [sigh 
] to struggle on w/ dual PII's/PIII's/Athlons/Bartons & etc & etc . . . . this thing isn't strictly the 'main' system; but in fact is having to fill in while I await an Iwill DK8X sometime in the new year [am reviewing the thing for burningissues].
The matched-batch late-week retail 366' in this one at the moment do 100MHz comfortably at default vcore [ran them at this for ages in a MSI 694D Pro-A] - from experience, they're likely to want mebbe another 0.1vcore to get up to [say] 105 . . . so what's the point? I've a couple of 333's coming, so will try for a more realistic chance at 110FSB then.
It's fun having a crack at this mobo when a sad old geek - only shock is how dire the BIOS is compared to proper server/workstation BIOS' . . . . tho' the soft-FSB stuff is amusing.
Performance at 5.5x100 is quite adequate if any demanding applications are multithreaded - Nero Recode2.x [which I now use a fair bit for quick 'n dirty DVD backup instead of its DVDShrink cousin] works at damn near 100% both CPU's & deep-analyses/transcodes a DVD9 in around 90-120 minutes with all the bells 'n whistles turned on - around three-anna-bit times as slow as a dual Barton, which ain't bad at all.
Joy of the thing is you can make it really quiet: with Speedfan & Sanyo fans on Kanie hedgehogs at 65% you can't hear the thing at all from the next room - 50-60W is a heck of a difference from the 150-200W [CPU's alone] horrors I've been using recently . . . . even watercooled you can't make the sheer commotion of such things go away . . .
Glad the fun I'm having now shows - hopefully as much as has shown the fun you & your colleagues & regulars have had with their BP6' in the past.
] to struggle on w/ dual PII's/PIII's/Athlons/Bartons & etc & etc . . . . this thing isn't strictly the 'main' system; but in fact is having to fill in while I await an Iwill DK8X sometime in the new year [am reviewing the thing for burningissues].The matched-batch late-week retail 366' in this one at the moment do 100MHz comfortably at default vcore [ran them at this for ages in a MSI 694D Pro-A] - from experience, they're likely to want mebbe another 0.1vcore to get up to [say] 105 . . . so what's the point? I've a couple of 333's coming, so will try for a more realistic chance at 110FSB then.
It's fun having a crack at this mobo when a sad old geek - only shock is how dire the BIOS is compared to proper server/workstation BIOS' . . . . tho' the soft-FSB stuff is amusing.
Performance at 5.5x100 is quite adequate if any demanding applications are multithreaded - Nero Recode2.x [which I now use a fair bit for quick 'n dirty DVD backup instead of its DVDShrink cousin] works at damn near 100% both CPU's & deep-analyses/transcodes a DVD9 in around 90-120 minutes with all the bells 'n whistles turned on - around three-anna-bit times as slow as a dual Barton, which ain't bad at all.
Joy of the thing is you can make it really quiet: with Speedfan & Sanyo fans on Kanie hedgehogs at 65% you can't hear the thing at all from the next room - 50-60W is a heck of a difference from the 150-200W [CPU's alone] horrors I've been using recently . . . . even watercooled you can't make the sheer commotion of such things go away . . .
Glad the fun I'm having now shows - hopefully as much as has shown the fun you & your colleagues & regulars have had with their BP6' in the past.