P4 EE OC...AI7 might be the answer

[mdzcpa]

So far I've gotten to 3.925mhz with my EE. That's with a basic water/pelt system.

The strange thing is how the system reacts when I hit the "wall". Anything closer to 4ghz results in spontaneous reboots. Prime can run for 45 minutes, than it just auto reboots. Same with 3DMark...one time it's rock solid, the next time it auto reboots. With my P4 3.0, as I neared the "wall" Prime would error out, 3DMark would freeze or CTD. But not with this EE though. It's just reboot city.

As I increase the voltage to try to stable things, the reboots get worse. Temps are still sane...under 18c at full load...-3c at idle with 1.75v set in the BIOS.

I'm thinking of giving the new Abit AI7 a try. I'm having terrible core voltage dips that I never had with my P4 3.0 on this IC7-G. I think the EE is overwhelming the CPU power regualtion. Not in terms of sheer voltage, but in it's ability to deliver raw amps. I beleive the AI7's Prescott compatible FMB 1.5 voltage specs may help.

Thought's?

BTW, in the following pic I have the voltage set to 1.775v. It's dipping badly and I am not even under load.

http://home.comcast.net/~mdzcpa/photos/EE3925.JPG

[Kanavit]

sweet, you got a nice system there.

[ZhaoYun]

Hmmm, which psu are you using?

[eva2000]

i have an AI7 http://www.xtremesystems.org/forums/showthread.php?s=&threadid=23342 :)

send your 3.2EE my way and i'll let you know :D

[Creative]

mdzcpa with my P4C800-E I am experiencing pretty much word for word what yours is doing. 4ghz and its stable for 2 hours of 3dmark benching and even 1 hour of prime but then it will just reboot without warning and wont post - or if i do reboot with windows it wont post

Heres what I reported at OCAU lastnight after having a stressful night lastnight. This is just a straight copy n paste from the post I made so sorry for offtopic stuff :)

Ive been having a few headaches tonight and I dont know if it is the end of the line for the cpu or if its another problem. I want to try an abit board to see how it goes so might see if SC will let me borrow his for a day this weekend :thumbup:

The problem is that I can prime at 4ghz for roughly an hour and have other stuff running at same time so pretty rock solid. I then stop the testing and reboot the pc as normal and then it wont post. I have to then sit here pressing reset for about 20 mins and then it finally works. I get into bios and then change some stuff and 9/10 it wont post again and we go thru another 20 mins :sick:

I have been doing some trial and error and it seems the problem is more prone when using 5:4 mode or 3:2 mode. It always seems to disappear straight away and post first go when using 1:1 mode. I have even when got into the bios finally lowered to as low as 230fsb and still it wont post. Change that to 245 and 1:1 upon entering next time and bang, works first go post and into windows :)

The temps are getting pretty hot I must admit also - 4Ghz @ 1.75vcore is reporting -11 in MBM5 and the evap is dropping to -32. This is using 1.85 in the Asus bios :rolleyes: I have never seen it below -38 nor have I had a cpu run below -18 since having the prommy and Asus combo. This is the reason I have hit my wall at 250 I think. I was talking with FUGGER and he agrees and also says that I can use up to 1.80vcore with r134a :)

So there is the current state of my rig atm. As promised serious benchies will be coming in the next few days for those who asked now that I have my max speed. The exciting news is the prommy will probably be modded to r404a in the weeks before christmas. This may lead to higher OCs as the evap temps will be ~15dc lower :thumbup:

Thanks for having patience guys :Pirate:

Does that sound similar to what your experiencing bud? Pretty frustarting isnt it and Im like you, looking for another board to try. Your on the MAX3 yer? Cos thats what I wanted to try....but maybe AI7 then? :confused: :)

[eva2000]

The Stilt in my AI7 thread here confirmed he saw a computer retailer show casing an AI7 in Mach2.. :D

[Creative]

Well plug it up then and tell how us how she goes pls :thumbsup:
hehehehe

[DaGooch]

It may be an amperage problem with the present board when you are pushing that EE that high. Amperage requirements from the power supply seem to on a drastic rise as of late.
http://developer.intel.com/design/Pentium4/guides/254310.htm

[Iridium192_217]

Instead of getting a new mobo, why not get a Prometeia & hit 4 Ghz easy?

[dpa]

could you give us some benchmarks at 3.9Ghz?

Super Pi @ 1M Kalk
PCMark 2002..
3DMark 2001..

weird problem.. Why do EE have the same stepping as my 2.4 C? "M0"..

[Iridium192_217]

They have the same stepping because the 2.4c m0's failed the 3.2 ee production process. There are a lot of the 2.4c m0 going around these days.

[Creative]

Originally posted by dpa
could you give us some benchmarks at 3.9Ghz?

Super Pi @ 1M Kalk
PCMark 2002..
3DMark 2001..

weird problem.. Why do EE have the same stepping as my 2.4 C? "M0"..
26360 (http://service.futuremark.com/compare?2k1=7273822) 3DMark2001

32sec Pifast 1Meg

98xx cpu
17xxx mem
for PCmark

Will have screenies soon, just need to upload :0

This is at 250fsb 1:1 2-6-2-2 and P4 EE on P4C800-E

[The Stilt]

Does anybody know, how to make Vdimm mod for this mobo?
I need more than 3.2V to make my KHX PC3000 sticks run properly.

Thanks

[dpa]

Originally posted by Creative
26360 (http://service.futuremark.com/compare?2k1=7273822) 3DMark2001

32sec Pifast 1Meg

98xx cpu
17xxx mem
for PCmark

Will have screenies soon, just need to upload :0

This is at 250fsb 1:1 2-6-2-2 and P4 EE on P4C800-E

thanks.. looks good..

[mdzcpa]

Originally posted by Creative
Does that sound similar to what your experiencing bud? Pretty frustarting isnt it and Im like you, looking for another board to try. Your on the MAX3 yer? Cos thats what I wanted to try....but maybe AI7 then? :confused: :)

Yep....that sounds like me. I can usually get it to post though. It's more about how it oddly just reboots itself instead of throwing any rounding errors in prime or CTD during 3D.

I'm using the IC7-G MAX II though. It showed rock solid voltages with my P4C 3.0 in there.

The PSU is an Enermax 650w that I've never had trouble with. In fact, this PSU and board (IC7-G Max II) was able to take my 3.0 to 4.1 without trouble. No voltage drops at all.

I know the EE runs hotter for sure, but it isn't showing signs of heat trouble. Once prime has run for 45 minutes, temps are pretty much at their max. Than 30 minutes after that, she reboots. Meanwhile I can watch my core v fluctuating and dipping badly.


Originally posted by Iridium192_217
Instead of getting a new mobo, why not get a Prometeia & hit 4 Ghz easy?

That's not the point. I don't want to go phase change for a number of reasons.

Besides, it's a power problem, not a heat issue. The reboots happen loooong after the CPU has reached max temps.


Originally posted by DaGooch
It may be an amperage problem with the present board when you are pushing that EE that high. Amperage requirements from the power supply seem to on a drastic rise as of late.
http://developer.intel.com/design/Pentium4/guides/254310.htm

Thanks for the link:)

Yep, I'm pretty sure it's amps. The whitepapers call for 91A!!! Holy cow. I didn't think the current boards supplied more than 70A or so, and the new Prescott ready FMB 1.5 supplies 78A if I recall.

I think I'm gonna give a FMB 1.5 Prescott board a whirl. Does anyone know if Abit is coming out with Max version with Prescott support?

[Soulburner]

BUT....how do you know the CPU has reached max temp? Where are you measuring this from? Unless you have some kind of magical probe inside the CPU die, you will never know. There are components inside the CPU that you can't measure the temperature of, and even though you may report a stable temp something inside may be warmer.

[Iridium192_217]

Well it seems to me that your simply unstable at the speeds your at. Your reboots seem to prove that. If I oc to a slightly unstable speed, it might work for a while but after of seti or something it will reboot.

What are your reasons for not getting prometeia? Sell all your water cooling stuff & it won't cost to much. Are you just trying to prove that you can hit the big four without phase change?:p:

[Tedinde]

Are you just trying to prove that you can hit the big four without phase change?

That's easy, i can hit 4ghz with a Zalman. What stucks is it costs me $600+ for another 300mhz??

Rig up and Icewater Res. and see if Phase change cooling will help you @ all and if it's worth it first. And if the Extra expense is worth it to you.

[Iridium192_217]

Originally posted by Tedinde
What stucks is it costs me $600+ for another 300mhz??


Well in his case he probably spent around $300+ plus for his water so its is not much less than a Prommy.

[mdzcpa]

Originally posted by Soulburner
BUT....how do you know the CPU has reached max temp? Where are you measuring this from? Unless you have some kind of magical probe inside the CPU die, you will never know. There are components inside the CPU that you can't measure the temperature of, and even though you may report a stable temp something inside may be warmer.

Temperatures are relative. The CPU diode gives a good idea of what's happening under the IHS. It's very doubtful certain "areas" of the silicon continue to rise is temp while the diode temps remains completely unchanged for more than 45 minutes.


Originally posted by Iridium192_217
Well it seems to me that your simply unstable at the speeds your at. Your reboots seem to prove that. If I oc to a slightly unstable speed, it might work for a while but after of seti or something it will reboot.

What are your reasons for not getting prometeia? Sell all your water cooling stuff & it won't cost to much. Are you just trying to prove that you can hit the big four without phase change?:p: [/QUOTE]


I've already hit 4.1ghz with my P4C 3.0 without phase change.....that's not the issue.

My main reason for not going to phase change is portability and practicality. In fact, I just took my rig to a LAN party last night. I played for 12 hours straight without a hiccup. It's very light and portable, totally self contained, and it's 100% turn-key. It all fires up and shuts down at the tough of a single button. If I run a Prommy, I'll loose the easy portability and still have to water/TEC cool the GPU anyway.

Bottom line, it's primarily a portable gaming box....not an ORB stomping benching rig that consumes the entire work bench and half of the room. I find it much more challenging to build a rig for top performance while keeping everything self contained within a traditional case. That's my thing.

As far as just "being unstable" when it reboots.....well that's kinda obvious. The question is why. Once the system reaches max temps, heat is no longer a variable. And the reboots are random...sometimes it happens in just a few minutes when the system is very cold, and sometimes after an hour or more....all the while the core v is fluctuating and sagging at more than .088v from what's set in the BIOS....that's a lot. And it never did that with my 3.0 in there.

For even more interesting fun, I hooked up my half completed chiller system to the rig this morning and dropped my loaded temps by 10c....and didn't make a lick of difference. I could only run it for a few minutes because of condensation issues, but it didn't help anyway. It would still do the reboot thing even when temps wer sub zero.

I dunno guys....it sounds like VRM to me. Perhaps super cooling does allow the CPU to remain a bit more stable under less than optimal voltage regualtion. But, I'd rather tackle proper VRM and FMB standards than go phase change and loose the purpose of the rig. (no offense to those who love phase change : ) )

[mdzcpa]

Originally posted by Tedinde
That's easy, i can hit 4ghz with a Zalman. What stucks is it costs me $600+ for another 300mhz??

Rig up and Icewater Res. and see if Phase change cooling will help you @ all and if it's worth it first. And if the Extra expense is worth it to you. [/QUOTE]

I was posting before I saw this. Yep...I got my chiller going this morning. Didn't help:( See above.


Originally posted by Iridium192_217
Well in his case he probably spent around $300+ plus for his water so its is not much less than a Prommy.

Yep, but all that was invested over a year and a half ago, and long before the Prommy was even around.

As I said above, phase change doesn't fit the concept of what I'm building. The vapo is a bit closer to being a fit, but isn't as strong as the prommy. In addtion, you still have to mod either to water/tec the GPU.

[Iridium192_217]

doesnt your "gamer" rig already have h20 cooled parts?

IC7-G; H20 w/ 226w TEC

[mdzcpa]

Originally posted by Iridium192_217
doesnt your "gamer" rig already have h20 cooled parts?

IC7-G; H20 w/ 226w TEC


Yep...sure does. It has the Swiftech H202-B kit upgraded with the MCW462-UHT (226w), MCW50-T (80w), high flow 120mm rad fans, and an eheim 1250. The 226w TEC runs off a Meanwell S320. Everything is totally self contained within the case:)

Anyway...this is getting pretty OT.

I'll be trying an AI7 next weekend and will report if it does better with VRM issues. My buddy picked one up this weekend for a rebuild before the LAN party, and his voltages are absolutely rock solid on that board. The VRM design is totally new.

[Schneider]

Yep, I'm pretty sure it's amps. The whitepapers call for 91A!!! Holy cow. I didn't think the current boards supplied more than 70A or so, and the new Prescott ready FMB 1.5 supplies 78A if I recall.
The Gigabyte 875p boards are said to have FMB 2.0 power standards which should easily cover that sorta Ampage. Maybe you should give the 8KNXP/8IK1100 a go

[Ragnarok]

8KNXP = slowest canterwood board...

[Schneider]

Originally posted by Ragnarok
8KNXP = slowest canterwood board...
It is possible to speed it up :p: just tweak a few settings and its back up to all the other canterwood boards

[Creative]

ALso mdzcpa u forgot to mention also that the reason why we dont think the cpu is at its maximum is that it is solid as a rock for both of us for ~45 mins and then bang - reboots :(

If it wasnt stable, dont you think it could and would crash within the first 5 mins...at the most 15?

I dunno, im just not convinced these cpus are maxed :banana:

[crotale]

I've also been wondering why Prime reboots the computer.
My brother got a watercooled P4 which can prime for 30+ minutes at 290FSB, but to get rid of the random reboots when running prime he has to back down to 284...
Sure, the things do get hot, but usually prime reports an error before reboot.
When I think about it, my comp has also started to reboot when running prime...

I've never had this problem before, and the only thing I've changed recently is graphics card, and I see we all got the Radeon 9800Pro.

Creative and mdzcpa: Which catalyst driver version do you use.
I know it sounds strange, but worth looking into...
I've also asked my brother to try a GF4 card and see if it helps...

[crotale]

ok, my brother swithed to a GF3 card, now prime:ing for about 2 hours, which before just rebooted after 40min.

He used cat3.9 with the ATI card. We will try diffrent cats tomorrow to test this theory...

[eva2000]

Originally posted by Creative
ALso mdzcpa u forgot to mention also that the reason why we dont think the cpu is at its maximum is that it is solid as a rock for both of us for ~45 mins and then bang - reboots :(

If it wasnt stable, dont you think it could and would crash within the first 5 mins...at the most 15?

I dunno, im just not convinced these cpus are maxed :banana: don't know about that i've had prime fail in the 28th hr before

[Creative]

yer but who besides you runs prime for 28 hours :D
heheheheh

Im considering even not even using it from now on....Im sick of this prime debate :)

[Soulburner]

I only prime through test 1 (all 6 passes). If anything is unstable enough to potentially crash the system (during gaming, etc), it will be caught by then.

[BBThumper]

.....sometimes I think people just run benches and prime with there rigs. Whatever happen to gaming and porn?........

[Soulburner]

I do all of the above with my system...:D

[ZhaoYun]

Same here, because that's what multi-tasking is all about. :p:

[st0nedpenguin]

Originally posted by BBThumper
.....sometimes I think people just run benches and prime with there rigs. Whatever happen to gaming and porn?........

Surely the finest reason for HT known to man, pr0n on the 14 inch monitor, BF1942 on the 21 incher. :D

[althes]

Originally posted by st0nedpenguin
Surely the finest reason for HT known to man, pr0n on the 14 inch monitor, BF1942 on the 21 incher. :D
You got that rite.

[Mr Gadget]

mdzcpa

Since the 3.2C EE is based on the 3.2 Xeon, I have been wondering if the 3.2C EE CPU multiplier can be lowered like the multiplier on the Xeon. Have you tried to lower the multiplier on your EE? I have read several reviews from the major hardware sites, but to my knowledge none of them have tried to do this.
If the multiplier can be lowered, it just might provide me the incentive to drop a grand on one of these EEs!

[mdzcpa]

Sorry for the delay, been out of town on business since Sunday.

Anyway, unfortunately the multipliers on the EE are locked. A real bummer for sure. I just got back into town and I'm gonna be looking around for a way to unlock them...but I think it's a long shot.

[Soulburner]

You have no chance at unlocking it. Intels way of doing it makes it impossible.

[mdzcpa]

I know...but there have been a few rumors circulating around in regards to the EE multipliers.

[Iridium192_217]

Originally posted by BBThumper
.....sometimes I think people just run benches and prime with there rigs. Whatever happen to gaming and porn?........

amen, benching can't touch any of those:D

[Soulburner]

There is nothing different about the chip, its a Pentium 4 Northwood 100%.

If someone proves me wrong by finding a way, great, but I don't see that happening.

[mdzcpa]

Ummm, that's why I said "it's a long shot" ;)

[Iridium192_217]

do you have your AI7 yet?

[Mr Gadget]

Originally posted by mdzcpa
Sorry for the delay, been out of town on business since Sunday.

Anyway, unfortunately the multipliers on the EE are locked. A real bummer for sure. I just got back into town and I'm gonna be looking around for a way to unlock them...but I think it's a long shot.

Thank you for the confirmation. I was hoping that Intel might have taken the same approach that they did with the Xeon, but looks like they decided not to. Well I suppose I could look at a 3.06 Xeon system... humm might actually be a little cheaper than the EE :D

[mdzcpa]

Originally posted by Iridium192_217
do you have your AI7 yet?

It's on order:) Should be here by next weekend.

BTW, I'm also looking into a used Prommy just for benching. It won't be my permanent gaming rig, but it would be fun for testing purposes:)

[Iridium192_217]

http://www.xtremesystems.org/forums/showthread.php?s=&threadid=23731

ahh so you are getting a prommy after all;)

keep us posted once you get the AI7