v mod 3/4 for p4b533

[Ardy]

Originally posted by: NanoMem
4:5 ratio above 132MHZ FSB's has always been there on my P4B533-V as far as I can remember...
Peace

Always been there for me too) I'm just not sure that auto defaulted to it..
Sure would like 3:4
Oh well I still get 171 fsb @4-5 with the 2.26 chip,it's just that this board doesn't offer much for the 1.6

Thugsrook,
" i dont bother with the lock myself cause im at or above 166fsb. using the lock also hurts performance slightly"
So why at 166 and above the lock doesn't matter? Does it default at that speed to 33/66 ?

 

[NanoMem]

mlaine
P4B533-V, -VM and -M all have a total of 5 switches in DWS1 while your P4B533-E and the plain P4B533 use 6 switches for DSW1. Either way you should only be concerned with the high switch i.e. in your case that would be #6.

I suggested few things for you to try out in that thread of yours regarding dipswitches 1 through 5 in addition to flipping switch 6 of course and all under "jumper free" mode. I'd still try those out. Also, you might consider falling back to one of the very first BIOS rev's.

Can you check the ratings off your RAM chips to confirm that it is not over-rated but will support 400DDR natively and not just under an overclocked setup? If it's hidden under heat spreaders, then I already know the answer...

Have you been able to take this RAM beyond 320DDR (FSB=160) even under 1:1 cuz your CPU or RAM might be the culprits? Can you run your 2.26 beyond 160 FSB? Try also just one stick of RAM at a time as there are some people who beleive that it might be more difficult to simultaneously oc 2 of them at once.

Peace

 

[mlaine]

Nanomem, thank you for your suggestions. I have only flipped the dip number 6 and left everything else alone (being in the jumper-free mode). I have not tried the jumpered mode/flipping the other switches/older BIOS revisions, but might give them a try when I get to work with the mobo again and have enough time.

The RAM chips don't have heat spreaders and have following text printed on memory chips:
KINGMAX
KDL684T4A2C
-50

The label on the chip reads:
MPXB62D-68KX3
L/N: (didn't write that up)
S/N: (didn't write that up)
DDR-400 256MB CL2.5

I can take RAM further with 1:1 ratio but the CPU or memory becomes unstable (testing with 3DMARK2001) so I've been holding back at 160 MHz FSB. At these greater than 160 MHz speeds I was feeding 2.7 V for memory and 1.6V for CPU. I've also tried a bit higher voltages (up to 2.9V for memory and 1.65V for CPU) but it's not stable and always crashes with 3DMARK2001. I might try to run it with only stick only but I really need 512 MB so I would anyway have to revert back to easier settings to get my work done. Still, I'm happy with my current settings (2.720 GHz/FSB 160) that seem to be stable and provide nice performance. I was just hoping for better memory performance as there should be plenty of room for improvement as far as the specs go.

I've tried these settings with 1004 official and 1007.12 beta BIOSes with same results.

 

[NanoMem]

Sorry, mlaine, I didn't mean for you to disregard your second ram stick for good. I was only trying to help you figure out the hidden potential of your system in terms of RAM/board/CPU oc'ability above fsb=160 so you can zero in on the culprit. And to an extent you have as now you know that your 2.26 CPU might be restricting you from going above FSB=160 and your plan to pursue a higher ratio is the right one since you'd be able to oc RAM higher while keeping your CPU happy underneath your 160 FSB threshold.

It looks like your RAM is truly capable of DDR400 as that -50 suggests 1000/5*2=400 (*2 for DDR). Check this link and see if you can decode the rest of your part numbers on the Kingmax web site to determine the exact CAS settings that will work at or near DDR400.

In terms of dipswitch settings have you tried the following conbinations
DS1,DS2,DS2,DS3,DS4,DS5,DS6
OFF, OFF, OFF, OFF, ON, OFF or
ON, OFF, ON, OFF, ON, OFF
and see if they work in jumper free mode. Start by relaxing your CAS settings to their slowest settings possible and FSB 133 in the CMOS with these and see if you're able to achieve that 3:4 ratio. Also set your AGP/PCI to auto if possible instead of locking them.

Good luck!

Peace

 

[Mikki]

mlaine, if you get a chance, try a couple things for me or let me know you aready have. Nanomem's suggestion of pulling one stick of ram is my first one. I did some tests earlier and found the one stick has a significant advantage of two sticks regarding performance and a slight advantage for overclocking.

Another thing is this-->when I first turned DS6 on, my system would not post at 100fsb, 2.5/3/3/7 timings, and ratio @ 1:1. 113fsb was fine, but not at 100. So that same little freakyness maybe affecting you, I'm only thinking because 100/133fsb speeds are mine/yours default speeds. Please try reversing the ratio and try it again. Also, if you have the time, up your fsb a tad (say...135) and try it with the 1:1 and 3:4 ratios. I'm wondering if it has something to do with a newer, faster processor with the 133fsb.

I've read your posts, I don't remember seeing where you've tried this already, if you have, let me know. Really HTH

 

[mlaine]

Mikki, I have no problem with POST when I'm using 3:4 or 1:1 ratio but Windows XP crashes immediately to blue screen using 3:4 ratio complaning (randomly) about ACPI.SYS or something else... I have tried 133/136/140/144/148 FSB speeds with 3:4 ratio and the system always posts but does not load Windows XP. There are absolutely no problems with 1:1 ratio up to 160 FSB. I haven't tried lower than 133 MHz settings because no way I'm going to underclock my processor even if the memory bandwith might increase somewhat. Maybe something is messed up with PCI/AGP clock causing a card to crap out and I'm going to try NanoMem's suggestion to set dip switches 1-5 and see if that helps.

 

[THUGSROOK]

 

[Mikki]

Did ya try the one stick of ram thing? And I think I read that you had your memory timings all relaxed, yeah? I'd be very interested in if a stick of the CorsairXMS would help. Know anyone that you could borrow from? (On that note, if you're interested in getting some, Overclockers.co.uk carries it at a decent price, and they ship all over Europe. Also, newegg.com does international orders, I don't know much about that, though)

This sucks, I wish I could help. There's no reason you should be having this problem, what we need is someone with a -E board and the same cpu to confirm it's not the cpu.

 

[THUGSROOK]

all they should have to do is this....

1) setup bios for 133fsb 1:1 mem & slow cas.
2) shut down system & enable DSx.
3) bootup to bios and set fsb to desired speed for testing.

WCPUid can correctly show the ddr speed.
Sandra mem test isnt always correct - but the test results are.

this should also work on the P4B533-V for a 3:4 ratio.

 

[mlaine]

Well, I've tried the following with my P4B533-E:

- Set FSB to 133, ratio to 1:1 and memory to 2.5-3-3-7 (slowest possible)
- Set AGP/PCI clock to AUTO
- Shut down system and switch dip6 on
- Power back on

Posts fine. Going to BIOS to see that all values are OK - they are. Windows XP crahses immediately while loading: PAGE_FAULT_IN_NON_PAGED_AREA

- Shut down system and removed one stick of memory. Power back on.

Posts fine. Crashes again. Same error message.

- Shut down system and tried 1-5 dip settings as NanoMem suggested. Power back on.

Both post fine but crash at same error while loading.

- Shut down system and remove NIC and sound card (nothing else remains but the display adapter). Reset all dip switches to OFF except dip 6 ON. Power back on.

Posts fine. Crashes again while loading XP. Same error message.

- Shut down system, turn dip 6 OFF. Reinsert second memory stick and NIC/sound card. Set FSB back to 160 and memory to SPD. Power back on.

Posts and boots fine. No stability problems.

Sigh, seems that the 3:4 ratio is a no-no for me...

 

[Mikki]

Good test...bad results. Sorry mlaine. :frown:

 

[SupermanCK]

mlaine maybe if you can try to test out your memories with some other system with 3:4 mem ratio...
just to let you know:
160mhz w/ 1:1 - mem = 160mhz = 320ddr effective
133mhz w/ 3:4 - mem = 176mhz = 352ddr effective
but your memory should be more than enough to handle it...
have you try locking your agp/pci divider?

 

[NanoMem]

mlaine, Your "ACPI hal" as well as "page fault in physical memory" XP errors suggest that your RAM is getting overclocked but for whatever reason XP doesn't like it. ACPI pbs usually point to a conflict between low level BIOS settings and low level kernel components of the OS usually HAL or its hardware drivers. Physical RAM pbs in your case can only be attributed to a timing conflict between the memory controller and the OS while physical ram areas are being accessed. Have you looked in your XP event logs to get more details about these issues?

I am not saying you have bad hardware (everything works at FSB 160 and 1:1) but it's just that when you flip DS6 it throws a wrench between your RAM and the OS. Have you tried each stick by itself in different ram slots? You said between 160FSB@1:1 --> 160MHZ RAM and 133FSB@3:4 --> 177MHZ RAM speeds is where you have pbs; Have you confirmed that your RAM/OS will work between 160MHZ and 177MHZ RAM clocks by doing what Mikki suggested i.e. trying FSB values between 120 and 132 with 3:4 and no DS6 trick? Yes your CPU would be underclocked in the tests but at least you'd be able to verify that the RAM/OS are not holding you back.

Although ACPI could be rolled back or updated to the "Standard PC" in the XP Device Manager I'd rather install the OS without ACPI support from the get go (massive registry entries and files involved) and activate it down the line if everything is fine. But before you contemplate this route, make sure to get any MS update to your OS first. In "standard mode PC", you only get 16 IRQ's without having to entrust the OS with the management of the PIC. In ACPI mode the OS takes over the functions of the PIC and compensates you with more IRQ's for allowing it to do so(MS tradeoffs:|)

Have you checked if your Kingmax and P4B533-E have been certified to get along at least under 133 FSB and 1:1?

Peace

 

[mlaine]

Thanks for the support guys! I haven't tried different DIMM slots or underclocking to find the exact limit for memory/OS instability. I'll have a look in XP event logs to see if there's any useful info related to crashes.

Nanomem: thanks for the ACPI->Standard PC conversion tip but I'm quite satisfied with the current situation and I think I'm not going to mess with OS as everything works just fine with current settings. The Kingmax website does not include any info about testing with P4B533-E (although a few other non-Asus mobos were mentioned on one page).

It might take a while before I test any further settings as now most of my time outside work is consumed by Neverwinter Nights.

 

[THUGSROOK]

mlaine ~ please keep us informed of any new info and well be doing the same for you

 

[Mikki]

Updated some scores to fill in some holes.....

SciSoft Sandra - CAS . 2/2/2/6 . . . . . . 2/3/3/6 . . . . . . . 2.5/3/3/7
113fsb 1:1 226ddr = 1602 / 1602 . . . 1559 / 1561 . . . 1553 / 1555 DS6 ON
125fsb 1:1 250ddr = 1785 / 1784 . . . 1738 / 1738 . . . 1731 / 1730 DS6 ON
100fsb 3:4 266ddr = 1946 / 1945 . . . 1902 / 1903 . . . 1900 / 1900 DS6 ON
133fsb 1:1 266ddr = 1934 / 1934 . . . 1888 / 1887 . . . 1878 / 1878 DS6 ON
113fsb 3:4 300ddr = 2188 / 2189 . . . 2142 / 2143 . . . 2137 / 2136 DS6 ON
150fsb 1:1 300ddr = 2169 / 2167 . . . 2114 / 2113 . . . 2105 / 2105 DS6 ON
125fsb 3:4 333ddr = 2434 / 2434 . . . 2381 / 2380 . . . 2379 / 2380 DS6 ON
133fsb 3:4 354ddr = * FAILED * . . . . 2577 / 2578 . . . 2573 / 2573 DS6 ON
150fsb 3:4 400ddr = * FAILED * . . . . 2882 / 2882 . . . 2878 / 2877 DS6 ON

100fsb 3:4 266ddr = 1946 / 1945 . . . 1903 / 1901 . . . 1901 / 1901 DS6 OFF
125fsb 3:4 333ddr = 2433 / 2432 . . . 2380 / 2382 . . . 2379 / 2380 DS6 OFF
133fsb 1:1 266ddr = 1939 / 1941 . . . 1891 / 1890 . . . 1890 / 1889 DS6 OFF
113fsb 3:4 300ddr = 2190 / 2189 . . . 2141 / 2140 . . . 2137 / 2137 DS6 OFF
150fsb 1:1 300ddr = 2172 / 2172 . . . 2118 / 2118 . . . 2118 / 2117 DS6 OFF
125fsb 3:4 333ddr = 2433 / 2432 . . . 2380 / 2382 . . . 2379 / 2380 DS6 OFF

After I disabled all USB, LAN, and Sound:
113fsb 3:4 300ddr = 2317 / 2317 . . . 2294 / 2293 . . . 2291 / 2290 DS6 ON
125fsb 3:4 333ddr = 2564 / 2564 . . . 2538 / 2539 . . . 2536 / 2536 DS6 ON
133fsb 3:4 354ddr = * FAILED * . . . . 2736 / 2735 . . . 2733 / 2732 DS6 ON
150fsb 3:4 400ddr = * FAILED * . . . . 3050 / 3050 . . . 3048 / 3048 DS6 ON

 

[SupermanCK]

great job mikki...
i still wonder why you can't run it with ddr354+ with 3:4...cas2226 of course

 

[THUGSROOK]

great update Mikki!

very usefull info for all the P4B533-E owners out there.

 

[Beefy79]

Good Lordy Lordy - it works!!

I'll give you a quick run down of my setup first:
Asus P4B533 rev 1.03
BIOS 1006
P4 1.6A @ 2.15Ghz (set to 133Mhz FSB - 134.5Mhz actual)
2x 256Mb Kingmax PC2700 CL 2.5

With the standard 1:1 ratio i was getting ~2050Mb/s in Sandra with the fastest settings of 2-2-2-5. Now with DS6 on I'm getting ~2700Mb/s with the slowest settings of 2.5-3-3-7. I don't think my RAM will go any higher, but I will play around - as it is I still have to do some stability tests.....

My only concern is my AGP/PCI speed. For example, does the system now think its on an overclocked 100Mhz bus and the memory derives from this, or is it still at 'official' 133Mhz bus, and *just* with hacked memory? Either that, or does anyone know any software that reliably *checks* (not just reports like Sandra tends to do) your AGP/PCI speed?

The reason I ask is that I'd much rather set my AGP/PCI to Auto and it be all 'normal', than force it to 66/33 if it doesn't need to be! By my calculations, if it thinks it is 100Mhz overclocked, my AGP will be at 89.7Mhz and my PCI will be at 44.8Mhz - too high for my liking.

 

[THUGSROOK]

Beefy79 ~ me and Nanomem have confirmed AGP/PCI dividers are uneffected.
AGP/PCI lock is also uneffected if you use it.
at 133fsb you are fine.

that 2x kingmax is actually running well for you, im almost surprised :Q

HTH

 

[Beefy79]

Thanks for that info Thugsrook! I will set it back to 'Auto' and see how I go.... please don't take it as mistrust, but is there anyway I can reliably check the AGP/PCI speeds anyway?

And yes, I'm a little surprised that my memory could handle it as well! But I will reserve my full excitement for a few hours of successful Prime95 testing.......

 

[THUGSROOK]

 

[dbwillis]

Thugs-
Did you try cas 1.5 at all? Im trying it for fun ...266 is on prime for 2 hrs now..

 

[Beefy79]

Alas, my poor little memory could not take the heat...... it was working really well on the cold boot, but as soon as it 'heats up' I guess, I get instability at first (programs crashing to desktop with no error message), then onto BSOD's, then I couldn't even get into Windows

Never mind. I was happy enough with the default performance, and even happier getting an extra 33% for free from the 1.6A! I can't complain about the memory not cutting it. I only paid AU$330 for the 512Mb (Corsair was AU$120 more expensive, and that cash was put towards a Ti4400 ). But for all you people out there that have spent a fair bit on good memory, then you should *definitely* try this mod!

 

[THUGSROOK]

i never tried it that low.

@ 333ddr cas 1.5/3/3/7 i got a no bootski