Calling P4T533 motherboard owners.

[JBELL]

AS I am working on this write up I have become curious what results you have achieved..

Right now I am sitting 2.53GHz @ 142fsb stable with temperatures sitting at 23c/73f MB Temp, 30c/86f CPU Temp.

[Tedinde]

I sold mine a week after buying it. I gave up on it. More people have had bad luck than good with it.

[TechTones]

I'm running Prime right now at 146/4x.

I expected much better. I can do 150/4x but it can't run Prime.
I tried the vdimm mod, and it didn't help at all. Maybe one mhz more on the ram. I don't get it. I was hoping for like 165/4x

I think it's the ram as I can run at 170/3x and 169/3x all day.

I'm going to try and burn it at 1 mhz at a time. But I think it's the rdram that's holding me back. If the board can run Prime at that particular vdimm, it's stable as a rock.

[JBELL]

im gonna down play the ram and try to crank the fsb more...

[JBELL]

142/4x 560fsb - stable
148/3x 444fsb - stable
151/3x - BSOD
Dropped down to:
150/3x *testing* - BSOD

[TechTones]

When did you buy the board? Was serial number did it have?
Mine was 2BZxxxxx.

I'm shocked at only 145-146/4x
I thought 160/4x was very doable. Even the volt mod didn't help.

I still say it's the crap yields from Samsung.

**edit** could it be your CPU?

[JBELL]

no cpu is b0 stepping and did 2.8 on air... my temps are 30c/86f full load .. .it is definitly the ram!

[Klosterbruder]

I run my 24/7 with 133 Mhz, don't want to to stress my little Cpu to much. But with LN2 setup it also runs stable with 197*3, 198*3 should also works fine, but more not, same limit as Holicho get, without any mod. Best *4 mode i haven't test so often, but my last test shows that 150*4 was even hard to get.

[TPR- Darwin]

http://www.wallis-phillips.co.uk/HelloBoyz.jpg

Rambus @ 3v instead of 2.5v

Motherboard Serial 2AZG******

Samsung Part Number MD16R1628AF0-CN9DF Supplied with M/B
Yellow Asus Sticker says 10/02

wont bench @ anything above 157fsb x4
above screen shot was taken with PC in freezer
currently stable on air @ 3302mhz

[TechTones]

I have a phase change freezer but I need to get it fixed.

I had my 2.8 at 3591. I know my stick of ram is very bad. I need a new one. I was doing 170/3x OK. but only 145/4x Prime stable.
I like the Asus board. I think it's a quality board.

[TPR- Darwin]

Originally posted by TechTones
I have a phase change freezer but I need to get it fixed.

I had my 2.8 at 3591. I know my stick of ram is very bad. I need a new one. I was doing 170/3x OK. but only 145/4x Prime stable.
I like the Asus board. I think it's a quality board.

There is meant to be a Tantilum? capacitor mod but I have never been able to find it.
The only parts of this board I am unhappy with is a) My 12v line drops to 11.58 (550w Enermax), b) My Vcore still fluctuates by 0.07v 100% loaded MBM 1.68v <> 1.75v.

[Jeff]

Hello Darwin! :wave: Nice to see you stop by! :D Let me be the first to welcome you to Xtreme! :toast:

[TechTones]

The ong thread about the P4T533 over at Hardocp
[ let them drop dead :D ] has the mod in there somewhere.
Do a search on Tantilum.

I did a volt mod for the rdram on the Asus but it didn't help me one bit. Perhaps bad rdram.

[TPR- Darwin]

Originally posted by TechTones
The ong thread about the P4T533 over at Hardocp
[ let them drop dead :D ] has the mod in there somewhere.
Do a search on Tantilum.

I did a volt mod for the rdram on the Asus but it didn't help me one bit. Perhaps bad rdram.

I can only find people talking about it, cant find the actual link :mad:

[TPR- Darwin]

Originally posted by Jeff
Hello Darwin! :wave: Nice to see you stop by! :D Let me be the first to welcome you to Xtreme! :toast:

:stick: I know u dont I ;)

Perhaps I should do a bit of folding or something

My Promethia aint turned up yet :-(

[TechTones]

Well here's the thread. It's more like a book ;)

http://www.hardforum.com/showthread.php?threadid=454638

Takes a WHILE to load. They suck what can I say?

[Clevor]

I have two 27Z boards. One tested OK at all even settings over 150/5x and is sitting back in the box. The other wasn't stable at all the even settings over 150/4x so I sent it back to the vendor (here in Japan). He offered me a 2B board made in Nov. by removing the packaged RIMM, but I told him I can wait till January in hopes of getting a 2C board.

I have done 156/4x with some 0228 RIMM4200 and 162/4X with 0236 RIMMs. I run two 256 MB sticks, nothing less. The 162/4X RIMMs did pass Memtest clean and the 16M SuperPI run, and all my 3D stress tests. Buffered bandwidth was 4050. I only use the 1.002 BIOS because the RIMMs overclock higher with it, and I don't use the RAID.

I hear rumors the newer, 'fixed' boards are slower than the original 27Zs. I'll flash back to the 1.002 BIOS and check it out once I get my replacement board.

[TechTones]

Clevor,

I'm testing some new Samsung sticks on my P4T533 and I get
around 152/4x Prime stable at default Vrimm. I did the vrimm mod and at 2.62 things are better but no 160/4x like some other people. I find many people only doing 145-148/4x or not even that.


You say the 1.002 bios is the best for O/C'ing? But that bios doesn't support C1 P4's. I think it's the 1004a that does that.

How much of difference on the sticks did the 1002 bios make compared to others?

[JBELL]

I recently swiched my b0 to a c1 stepping p4 - why doesn't the bios support it TechTones? seems to run fine other than the fact C1 steppings and much hotter than B0

[TechTones]

The board does support C1 and Hyper Threading { 1005 bios} but I saw one of the bios improvements on one that said now supports C1 stepping P4's. I don't know. I'm using the 1004a bios right now.

I'd love to get a stick that does 162/4x and better. Too bad some place doesn't sell tested Rambus that OC's high.

[JBELL]

in a few days I will see where my C1 can go - if it same as B0 I will know my samsung is garbage.

[TechTones]

Man Bell

I'm at 153/4x Turbo off at 2.62v right now running Prime OK.

I'll tell you one thing, I think Rambus benefits a lot from burning in.
I mean I saw a difference after a 5 hour burn in with Memtest 3.0

They have a ISO file for Memtest now. Just boot to CDROM and let it rip.

[JBELL]

what did you burn in with?


I have gotten to 142 or such but not much better - I post @ 150 but reboot shortly into loading windows xp.

[TechTones]

Right here on a floppy or make a CD.



http://www.memtest86.com/

[JBELL]

very nice I will try this when I get home today.

[TechTones]

Just like it run for 5 hours or better yet, overnight. The more often
the better. I burn in every day for at least a couple hours or more.

[JBELL]

Originally posted by TechTones
Just like it run for 5 hours or better yet, overnight. The more often
the better. I burn in every day for at least a couple hours or more.

is there burning TOO MUCH??

[TechTones]

No way.

ram was made to be used for a lifetime when using recommended
voltages.

[jdmcnudgent]

i am currently at 133+turbo. board is so unstable, really dont think the bios upgrade will fix. chip is 1.8a sl68q, older packdate, and chip does 150 on default, but the board and mem are not stable. i crunch seti so i look for stability.

[Clevor]

"You say the 1.002 bios is the best for O/C'ing? But that bios doesn't support C1 P4's. I think it's the 1004a that does that.

How much of difference on the sticks did the 1002 bios make compared to others?"

Techtones, I only tried the 1.004 and found I had to drop the ram maybe 2-4 mhz. Where Memtest ran at 156/4x on the 1.002 BIOS it would lock up on the 1.004. Cyuet on the Overclockers.com forum and I think someone on the Asusboards forum confirm this.

I haven't run the board lately but I have since gotten some C1s I will test. As I mentioned, it's entirely possible the latest, 'fixed' boards have been neutered by Asus. People are saying it is benching slower, or that they can't run their RIMM4200 much beyond 152/4x. Maybe I should have kept the 27Z board I sent back to the vendor, as it did do 156, 160, and 162/4x, but was unstable at 154, 158, and most of the 140s. I had to disable Turbo for stability.

I also believe there are now -CT9 RIMM4200 out, starting from week 0246. I plan to get some to replace my 156/4x sticks but figure the later the batch the better, so have been procrastinating. But according to Chainbolt on another post, he tried dozens of sticks and says they aren't necessarily getting better. Heck, I also got Kingston 16-bit RIMMS (presumably Samsung -CT9, 32-ns) that are doing do 156/4x now on a P4T-E. Next step is to see how they run on a TH7II.

Jdmcnugent, if you can't even run 133/4x, sounds like you have one of the bad boards. Is it a 27Z?

[TechTones]

I can tell you one thing,

I was getting like 3800 in Sandra mem at 152/4x on the Iwill P4R533. On the Asus I get that at 155/4x. I thought 32 bit was supposed to be a little faster then 16 bit?

What's new about these -CT9 sticks?

And yes, I'm kind of stuck at 152-153/4x for real stability.
I also noticed some sticks will boot at 160/3x and some won't which doesn't make sense.

[Clevor]

Techtones, the -CT9s I believe are 13 (or 0.13?) um, true 32-ns.

3800 at 155/4x seems a bit low. I get 3950 at 156/4x on my P4T-E. Is that a 'fixed' P4T533 you got?

As I recall, I got about the same mem scores on my P4T533 as the P4T-E, but the latter benched better in 3DMark2001. It's because the RIMMS are arranged in two banks of two equidistant from the CPU. With two RIMMs on the P4T533, one gets in the way of the other. Some people feel the electrons move so fast should it make a diff? But if you look at the Matrox Parhelia card, they went through all the trouble to angle the ram so they are equidistant from the CPU.

I really didn't see any performance increase between 16 and 32-bit rimms at all, as far as Sandra scores, but didn't check unbuffered scores.

[TechTones]

Yes mine is a fixed board. Ricktek VRM instead of Intersil HIP.

I also find I can't run Prime stable at say 150/4x with 2 sticks as opposed to one. One OC's higher too. Now the Epox 4T4a+ wouldn't even boot into windows for 15 sec before rebooting. it's going back for sure.

I wonder if I really need 512mb. I'd like it but I can't OC as high.

I might try the 1002 bios. That's the official release version right?

I haven't seen any CT9's yet.

[kiko]

my P4T533 has serial 27ZG1 (July)
My 256MB samsung CN9 is week 0229
bios 1005 final

this is my best w/o vrimm MOD.

http://kikoguru.supereva.it/img/4132.png

sandra only

w/superPi 160*4x turbo max
w/3dmark/pcmark 156*4x turbo max w/1.6A/2.53B/2.8C1.

Ciao

[Dark_Angel]

Originally posted by TechTones
I haven't seen any CT9's yet.

I have 2 RIMM sticks to put in a P4T533 and I think it says -CN1.They are 1200mhz...

[mkleingeld]

Just got my Prometeia yesterday and after spending the whole afternoon putting it on the board I'm pretty happy. Before I was able to do about 155 - 156 x 4 turbo max stable. Now I'm running at 173 x 3 turbo but it's limited by the chipset I think because when I go further it freezes after running seti for some minutes. If it was the cpu or the mem it would give bsod. I think I'll go buy a more substantial chipset cooler next weekend. Does anybody have experience with other coolers on the chipset of the p4t533? And Dark_Angel, if you really have rimm4800 mem, go see how far they go, now! please :)

[TechTones]

Originally posted by mkleingeld
Just got my Prometeia yesterday and after spending the whole afternoon putting it on the board I'm pretty happy. Before I was able to do about 155 - 156 x 4 turbo max stable. Now I'm running at 173 x 3 turbo but it's limited by the chipset I think because when I go further it freezes after running seti for some minutes. If it was the cpu or the mem it would give bsod. I think I'll go buy a more substantial chipset cooler next weekend. Does anybody have experience with other coolers on the chipset of the p4t533? And Dark_Angel, if you really have rimm4800 mem, go see how far they go, now! please :)

Hi :)

My ex wife was Dutch. She was born in Amoortes Forte. I was there like 10 years ago. Very nice place Holland ;)

I'm getting a Prommy too. Did you have any problems attaching the CPU kit to the CPU considering the sicket is upside down and in middle of the board? I have the same board too.

[TPR- Darwin]

Originally posted by mkleingeld
Just got my Prometeia yesterday and after spending the whole afternoon putting it on the board I'm pretty happy. Before I was able to do about 155 - 156 x 4 turbo max stable. Now I'm running at 173 x 3 turbo but it's limited by the chipset I think because when I go further it freezes after running seti for some minutes. If it was the cpu or the mem it would give bsod. I think I'll go buy a more substantial chipset cooler next weekend. Does anybody have experience with other coolers on the chipset of the p4t533? And Dark_Angel, if you really have rimm4800 mem, go see how far they go, now! please :)

you are doing slightly better than me, highest I can run seti is 170fsb any higher locks up like yours :-)

Are you running motherboard monitor ?, what are your settings as mine doesnt work. :stick:

[mkleingeld]

Originally posted by TPR- Darwin
you are doing slightly better than me, highest I can run seti is 170fsb any higher locks up like yours :-)

Are you running motherboard monitor ?, what are your settings as mine doesnt work. :stick:

Nope, I downloaded CPUcool. This one does give the negative temps ok. You can also change the fsb in windows with it but I don't know what PLL the p4t533 uses. It's an ICS I think but which one... You have a Prommie as well now? How's your contact patch on the cpu? I had to modify that attachment thing a bit to get it correctly on my cpu. But it's not to my liking yet so I'll go home a bit earlier today and spend some more hours trying to get a better contact patch. Btw, I found out that it really is the heat of the cpu that makes it block. It's not the northbridge. Right now I'm getting about -15 degrees on idle. That should be a lot less I think. Hope to get a lot better performance this evening.
Btw, the best I have done for some pi benches are 176 fsb :) But I want more!!!

[kiko]

Originally posted by mkleingeld
Nope, I downloaded CPUcool. This one does give the negative temps ok. You can also change the fsb in windows with it but I don't know what PLL the p4t533 uses. It's an ICS I think but which one... You have a Prommie as well now? How's your contact patch on the cpu? I had to modify that attachment thing a bit to get it correctly on my cpu. But it's not to my liking yet so I'll go home a bit earlier today and spend some more hours trying to get a better contact patch. Btw, I found out that it really is the heat of the cpu that makes it block. It's not the northbridge. Right now I'm getting about -15 degrees on idle. That should be a lot less I think. Hope to get a lot better performance this evening.
Btw, the best I have done for some pi benches are 176 fsb :) But I want more!!!

the PLL that uses is ICS 950703 and is not supported by cpucool.
Datasheet here http://www.icst.com/products/pdf/ics950703.pdf :)

Ciao

[mkleingeld]

Thanx Kiko.
Hey, if you read through that datasheet it seems to be possible to set the fsb at 226mhz. If there is a program that supports this, it should prove usefull when the Prescot comes out.

[kiko]

Originally posted by mkleingeld
Thanx Kiko.
Hey, if you read through that datasheet it seems to be possible to set the fsb at 226mhz. If there is a program that supports this, it should prove usefull when the Prescot comes out.

unfortunately there are no prgs (at the moment) that support the PLL.

Ciao