P5W - FSB Wall testing and conclusion

[simonmaltby]

FSB WALL TESTING RESULTS AND A CONCLUSION

I did some work last night regarding FSB wall on this board....

P5W DH deluxe with Bios 1101
E6700 OEM Step 6 B2
2GB OCZ 6400 Plat 4,5,4,15

Results:
I have been having real issues at 380FSB and Above. I tried out an E6600 and that worked up to 400FSB, but no higher.

Another post of the foum put me onto crystal http://crystalmark.info/download/index-e.html which lets you change the multi in windows on this board.

So I tried the E6700 at 9x and 8x to see what happened.
The only Bios setting that helped at all was to set the ram to SPD - no voltage changes let me go higher than 379FSB. With SPD on i could do 410FSB before it would go no further. I should point out that SPD sets my ram at 5,6,6,20 - miles out from what it should be.... this got me thinking...

Next I booted with SPD settings and used memset http://rv.page.cegetel.net.perso.cegetel.net/MemSet.zip to see all the ram settings. I took a screen shot and saved the memset settings.

Then I rebooted changing back to manual RAM settings in the bios 4,4,5,15

When in windows I opened memset and changed all the settings apart from 4,4,5,15 to be the same as the screen shot at SPD.

I was then able to go right up to 440FSB before i suffered any issues.

Summary:
1. FSB Wall at 380FSB with manual Ram settings in Bios
2. FSB Wall at 400FSB with SPD ram settings in Bios
3. FSB Wall at 440FSB with manual Ram settings in Bios but changed to match SPD in windows with memset.

Conclusion:
Asus need to provide access to far more RAM settings in the Bios. This might well be the cause of the FSB wall and would explain why different people get the wall at different points... down to when the RAM fails due to its settings.
Wonder uis Asus read this forum... maybe a bios with access to all ram and some multi adjustment like the Gigabyte boards would be good.

[eva2000]

interesting got memtest screenies for those 3 scenarios ?

[simonmaltby]

interesting got memtest screenies for those 3 scenarios ?

here is the memset (yes i know its the old version) screen shot i took at SPD


And here are the memory settings when set manually in the bios at 4,5,4,15


the change i make is to change all settings from tRFC downwards to match the first pic.

Sometimes the bottom three settings vary on boot without changing anything else. Also slight FSB changes in the bios change the tRFC and down, sometimes dramatically. I did have tRFC at 13 after one boot eeek

[eva2000]

hmmm sounds like maybe tRFC or tRD related.. try only loosening these 2 values and see how far fsb go

this was all tested with maxed MCH volts right of 1.65v ?

[eva2000]

thank you so much seems your findings seem to apply to my Intel D975XBX Badaxe too! i loosened tRFC and tRD from 35/6 to 40/7 and went from 433fsb max fsb to now benching at 440fsb on 1067fsb strap (will try higher later )



Maybe all motherboard manufacturers need to do is open their bioses to the other memory timing options that memset has available so we can further fine tune how high the fsb can go ?

[Lestat]

do those two values effect real world memory performance ? or is it one of those settings that really only effects stability ?

hhmm odd i do not have tRD on my 965 chipset with the memset i use ? hhmm

[Durzel]

I would've thought it would have an effect on performance, but apparently memory timings are less relevant to Conroe than they were with other CPUs so I doubt changing those "extra settings" will make a huge difference.

[eva2000]

they do effect super pi times if that's what you mean

[Lestat]

ok using th elatest memset
wow my tRFC is 0 by default! maybe that can explain why i can not make my memory run 925+ stable.
still can not find tRD, hhmm

*edit*

nope that dont work lol

hhmm something odd with this DS3 and ES chip and this Ram
8x multi 950+mhz 4-5-4-12 ram = 32m stable but can not do 1000
9x multi 925mhz 4-5-4-12 ram = no good.

8x multi 950 i think is ok for 4-4-4-12 ill check my screenshots see if it worked.

so what is the max bootable from the bios you guys can do.? best i can do from my DS3 is 400mhz 1:1 4-4-4-12. w/ 2.0v on ram. have not tried 1.9v 2.1 2.2 2.3 and 2.4v on the ram does not matter.
infact 2.4v wont even POST.

hey eva. my HZ's are 0631 these should be D9's yes ? they sure dont act like it. but again. you said before you think maybe my mobo is bad for high dram freq. i believe that is probably correct.

[i found nemo]

do 900 4-4-4-12. that should be stable and still give good b/w. or maybe you can tighten some timings since ram freq would be lower. keep it efficient ya know?

[eva2000]

yeah 0631 should be micron d9s mine are 0629s

i'm trying to tighten the timings down to see how tight i can do at 438-440fsb heh. I noticed one thing if i have memset open there's instability in benching/testing stability. Close memset after making your changes, and more stable

have you tried remounting chipset heatsink again.. my 1st mount i dropped to max 432mhz 4-4-3-8 at 2.2v on my ram, 2nd remount better contact and manage 488mhz 4-4-3-5 at 2.2v!

[Lestat]

my chipset never gets above 38c. i have a 40mm fan on it.
plus this DS3 uses pushpins to hold the chipset in i would have to pull the board out ot do that. and eww alot of work lol.

900 4-4-4-12 is impossible on 9x multi. its not stable enough i think.
8x multi yes. but the FSB of 450 is not bootable from the bios.

this board is quirky for sure. im hoping gigabyte will be able ti fix the little quirks.
but i am not complaining at all $145 very well spent. i can still overclock higher than i did with 2 different p5w dh boards lol.
and i can not do tighter @ 800mhz 2.0 2.1 or 2.2v the system wont boot. 4-3-4 or 4-3-3 nothing works.

[eva2000]

yeah when first got badaxe my PC6400PHU2 HZ wouldn't even boot at 400mhz 4-3-3-12 at 2.2v but after remounting chipset block can do 400mhz 4-3-3-8 at 2.2v now

[Extera]

got the FSB up from 420 tot 433, still need to try the last method of matching de SPD timings using memset. Thanks for this info!

[Lestat]

yeah when first got badaxe my PC6400PHU2 HZ wouldn't even boot at 400mhz 4-3-3-12 at 2.2v but after remounting chipset block can do 400mhz 4-3-3-8 at 2.2v now

was your chipset getting too hot though ? or did you remount it to try and make it run cooler ? meaning, were the temps ok already and you just remounted it just because.?
i also have a 80mm fan blowing on the ram, and also another blowing on the entire PWM area.

maybe ill pull it out later and try it. right now i am doing Laundry hehehe.

[Shpoon]

Interesting...

So, if you take the SPD settings (other than timings), and tweak the timings (tighten them up) are you still able to acheive this higher FSB, or does it start to fail again?

[FELIX]

->Lestat:You say that your tRFC is 0,is your others values are correctes?

[eva2000]

was your chipset getting too hot though ? or did you remount it to try and make it run cooler ? meaning, were the temps ok already and you just remounted it just because.?
i also have a 80mm fan blowing on the ram, and also another blowing on the entire PWM area.

maybe ill pull it out later and try it. right now i am doing Laundry hehehe.

my chipset was cool as a cucumber even passively.. but seems remounts or getting the right mount helps as 432mhz to 488mhz is a huge diff!

Decided to see how far i can tighten the additional timings knowing the effects tRD has on higher fsb.

Software used: Systool dual 32M Pi tests as 99.99% of time if you can pass dual Systool 32M pi you can pass Super Pi 32M.

Using the memset values for the following:

tRFC | tRD | Write to Precharge Command | Write to Read Command | Read-Write Turnaround clocks | Write-Read Turnaround clocks | tRTP

Managed to pass with following settings:

• 440fsb = 40 | 7 | 12 | 10 | 8 | 5 | 4
• 438fsb = 36 | 7 | 12 | 9 | 8 | 5 | 4
• 438fsb = 35 | 7 | 12 | 10 | 8 | 5 | 4
• 438fsb = 34 | 7 | 12 | 11 | 8 | 5 | 4
• 438fsb = 33 | 7 | 12 | 11 | 8 | 5 | 4
• 438fsb = 32 | 7 | 12 | 11 | 9 | 5 | 4
• 438fsb = 30 | 7 | 12 | 11 | 9 | 5 | 4
• 438fsb = 28 | 7 | 12 | 11 | 9 | 5 | 4

@438fsb = 28 | 7 | 12 | 11 | 9 | 5 | 4


@438fsb = 36 | 7 | 12 | 9 | 8 | 5 | 4

[Lestat]

->Lestat:You say that your tRFC is 0,is your others values are correctes?

since i have updated to the new version it does not show 0.
yeah everything else is ok i think but how can we tell ? ther eisnt anything else that can read the settings is there?

Felix where is the tRD for the 965 chipset ? or the DS3 mobo ?

here are some pictures be sure to look at the description for each one.

PICTURE 3
Here is my Bios. what do you guys recommend for these settings ? i only ask cuz the names used in the DS3 are different
PICTURE 1
this is memset showing the manual settings plus all the other ones
PICTURE 2
This is memset showing the settings for AUTO SPD

[FELIX]

-I don't find tRD in 965 datasheet,it's probably exist,but I don't know where.

-ACT to ACT delay->Act. to Act. delayed (tRRD)
-rank Write to Read Delay->Write to Read Delayed
-Refresh to ACT Delay->Refresh cycle time (tRFC)

[Lestat]

Felix any chance of giving them proper names so stupid people like me understand what they are lol ?

[Cocacola]

so the problem that P5wdh can not do high FSB only because RAM timings stuff?
if so im sure someday with right bios it will rocks.. LOL

[eva2000]

so the problem that P5wdh can not do high FSB only because RAM timings stuff?
if so im sure someday with right bios it will rocks.. LOL

if that's the case maybe higher vdimm + higher mch volts = higher fsb too ?

also would explain why some memory can take your board's max fsb higher than other memory i.e. PC5400UL micron fat body d9s seem to take my fsb higher than the newer micron D9GMH where fsb limit is an issue

[eva2000]

thank you simonmaltby.. you've given me an extra 56-62mhz on my E6600 cpu

Well so far 3954mhz is max 32M.

Here's the break down of results for max 32M

@3898mhz - 9x433fsb max
with tRD at board default 6


@3934mhz - 9x437fsb
@tRD of 7 with tRFC 28 and looser additional timings


@3934mhz - 9x437fsb
@tRD of 7 with tRFC 32 and tighter additional timings


@3954mhz - 9x439fsb
@tRD of 7 with tRFC 36 and additional timings needed for 439fsb


Couldn't get 32M @440fsb for now probably more tweaking needed or more vcore/mch volts.

But here's 8M


All 32M times run from same session of windows, so probably not optimal for fastest times yet

[mackanz]

What speed did you boot at Eva?