P4 1.8A + PC800 RDRAM

[SafeZone]

Hey guys,
I foolishly bought the Asus p4t533-c (I850E) with 512mb pc800 samsung rdram, with a 1.8A northwood. The max stable OC i have hit with onboard sound disabled is 125FSB ~2.25ghz. At 130mhz fsb the system posts, but windows crashes on boot. This is a pretty crappy overclock when i've been told that the 1.8A can hit 2.4 no problem. Do you guys have any suggestions on how I could get higher? I'm guessing the RDRAM is holding me back. I've read that there are ways to modify voltage for rdram chips, anyone know anything about that?? The Mobo has no AGP/PCI lock, however at 133fsb the agp and pci are set at normal 66/33, but the system will not even post at 133. Also would better cooling help or is that not necessary? my temps haven't really hit higher than 50C, although i have not done any stress-testing.

Should I try to sell this thing and get a new I845E board? or maybe just go for the pc1066 rdram? (even though that is hella expensive)

thanks for the help

 

[Duvie]

I am not necessarily familiar with your rdram mobo, but isn't there a way to change the memory multiplier for the rdram??? I mean at default 100fsb it should have been 4. If you can change it to 3x if it is the rdram you should be able to push ahead until the cpu limits you...

Ie 4 x 125fsb = 500 or pc1000...you have pc800 and maybe this is too much for your particular modules...

if 3 x 125 = 375 or pc750

3 x 150 = 450 or pc900 and that should be close to the 1.8a's limits of 2.7ghz on averagae of recent 1.8's....


If you want to change I would hold off until Granite Bay Dual DDR mobos arrive in mid Nov....If they turn out to be as good as initial thought they will beat pc1066 system and if oc'd with lower multiplier chips should pul away from rdram systems...

 

[Big Lar]

Sounds like Duvie hit the nail on the head, more than likely your ram is the culprit/ I have the same board and havn't had Any issues with it.

 

[SafeZone]

Yea that was a good suggestion, taking the FSB/RDRAM ratio down to 3X allowed me to get to 130mhz stable. However, at 133mhz, even at the highest voltage (1.85v) I could not get windows stable. However, one interesting thing to note, is even if I set the voltage to 1.5v, the BIOS power monitor and Asus Probe report Vcore consistently as 1.648v consistently. It seems that this board overvolts by .15v! I've seen boards undervolt .05v before, but I never seen a board overvolt .15v. This is very strange to me, and maybe this is a problem. Even if I lower the voltage to the lowest setting in bios (1v) it still comes up as 1.648v. However, if I increase to say 1.7v, it will read around 1.8v. ANyways, any input about that would be appreciated.
thanks

 

[Duvie]

Ok...

Asus mobos at least the i845 and sis645 mobos overvolt and do anywhere form .03-.06v..the reason you have .15v may be the fact the vcore booster jumper may be enabled. This often adds .1v on top of the basic overvolting....

As for the culprit I don't think so....

I am concerned about what the limit is now...It can't be the rdram as when you were running it at 125fsb with a 4x multiplier you were running effectively pc1000 and now at 133x3 it wouild only revert back to basic pc800 and should do that fine....

Also the 130 should have possibly been far more unstable as it should have had pci way out of spec at 43.33/86.66....

What vcore gets you stable at 130fsb or 2346mhz???

What are the specs to the power supply you are running???

Are you using p4 connector or ez plug with the standard 4 pin molex??? If using one try the other and then try both at same time....

Also 50c with no stress tesing means idle temp at desktop??? cause 50c plus is high for asus mobos which read at the low end of the spectrum...Try taking off side of case and/or aim large fan into the case to see if it is heat related to the cpu or the chipset...

 

[SafeZone]

at default 1.8ghz idle temps are 35C. At 130FSb, idles would get to 40 C maybe. Under load, say playing a game for an extended period of time, I can get up to 50C. I'm using a 420W PSU from directron.com, its a dual-fan PSU thats modeled after the antec one, only its a lot cheaper. Um my case is pretty well ventilated, front intake fan, rear exhaust + psu fan, plus a side intake fan blowing over cpu. I do have the overvolt jumper set to ON, should I set this to off to eliminate that + .1v or maybe that would not allow me to change voltage anymore...I believe I was able to get 130 stable (didnt do extensive testing, but no apps were crashing) at 1.7v which was around 1.8v in reality. I'll try your suggestion about the p4 power connector and the ez plug.

 

[SafeZone]

Okay, no luck on the EZ plug, tried that, and the P4 plug at same time, same result. Im currently using the p4 plug, Im guessing this is the same thing really. Anyway, I was able to get 125fsb stable @ 1.6v and 130fsb stable @ 1.65v. One strange thing to note, is when I get to 133/33 I get this box that comes up in windows, (if windows boots) saying "oops another error on the adp, you won't see this dialogue in the release version" with an OK box. What the hell is that supposed to mean? Does this mean my board is a beta version or something? It only does this apparently when there is not enough voltage (for instance if I try to run 130fsb @ 1.6v) UGh anyways I can't get 2.4ghz which is very distressing. Maybe heat is the problem, I'm using default intel heatstink+fan with that thermal pad that came with it, maybe I need to get a new hsf or that thermal pad sucks ass. However, my temps are okay, I dunno...

 

[SafeZone]

Help Please !

 

[Adul]

Originally posted by: SafeZone
Help Please !

it could be that your cpu wont go any higher.

 

[SafeZone]

I suppose that is a possibility, makes me angry when other people have gotten to 2.7ghz though

 

[GrumpyMan]

What bios are you using? Since that mobo is made for PC1066 Rdram it has timing issues with PC800 RDram when people try to overclock it and only certain bios versions will let you o/c PC800. And yes, your cpu just may not be able to handle it.

 

[fluxquantum]

my 1.6a is overclocked to 2.3 GHz. i am using the swiftech mcx4000. believe it or not i am using a p4t-e mobo flashed with the p4t533-c bios. i believe your CPU is just limited to what you can overclock. i have beem trying myself to get to 2.4 but i windows will just crash.

 

[wicktron]

It also helps to know what RAM you have. PC800 what? Samsungs? Kingstons? Generic? Etc... Also would be helpful to know the ns rating. 256mb 16-device 45ns PC800 is probably the best OC'ing RD you can find... It can usually do 140-144x4 or 165-180x3 depending on your configuration. Head on over to the Hard|Forums, the guys at Strictly Intel really know the deal when it comes to RD.

 

[tbates757]

The problem is just the fact that you are using RDRAM, it doesn't overclock nearly as well as DDR therefore your CPU overclock is pretty limited. Getting 1066 would probably help but you aren't gonna achieve the kind of overclock that you would with something like Corsair XMS pc3200 ddr, it's just not gonna happen. RDRAM already runs hot enough at stock!

 

[SafeZone]

According to what I am hearing, the problem is not RDRAM related, since with a 3X multiplier, I can thoeretically hit 165FSB with this multiplier and I am not even hitting 133FSB. I'm guessing its the processor, I bought it when the northwoods were pretty young, maybe that is why, is there a serial number somewhere, I don't think I still have the box...

 

[Duvie]

pretty much the main thing for us is the packing date....An early packing date of Dec 01 to Feb 02 and 2.4ghz or 133fsb was hardly happening....May 02 chips were hitting 2.4ghz regularly and often at default vcore....Aug 02 chips arrived with SL68Q chips that hit 2.6ghz often at default vcore and often go up to 2.8ghz....


I agree I don't think it is the rdram...we ruled that out once we changed the multlier (underclocked it severely) and it still didn't let you get but about 5 more mhz out of the fsb...133x3 would have only been 400mhz or pc800 and you were doing 125x4 at 500mhz or pc1000 already with the earlier 4x multiplier....

 

[SafeZone]

i bought it in late may from newegg.com, however that does not necessarily mean it was packed in may correct?

 

[Duvie]

Correct...I purchased my 1.6a at newegg in late june and it was a mid may chip...So that being said and newegg being a high volume dealer there is a good chance the packing date was within 2 months so march or april...If that is true to assume then it could just be the limit of the 1.8 plus the fact that the rdram platform chipsets are not as good in ocing situations...with a bucnh of tweaking and tricks 2.4ghz could possiblly be stable at 1.7v or less....