presler 920 @ 4.1ghz stock voltage & STABLE

[BOLt]

Originally posted by: RichUK
I want to see 5Ghz now

Pretty sweet OC you have there, especially from 2.8, how did the previous 90nm DC's OC anywhere near, i guess they were a lot worse?

Ohh, is that stable, i don?t personally consider 1M stable ? Also what temps, is it throttling (i think you can use an app to check this)?


EDIT: typo

anandtech's ces show article (i forget which day) featured affordable phase-change units. ocz had a $200 offering where they demo'd a presler running at 5.125GHz at -25C load... (yes negative)

 

[stevty2889]

Originally posted by: qoou
is there a good motherboard for this cpu where I could use dual channel ddr1 pc3200 sticks insteas d of having to buy all new ram?

Asus P5P800 supports it, and it uses the 865 chipset with DDR, but also uses AGP instead of PCIe. The Asrock 775dual 880pro board supports it as well, has support for both DDR and DDR2, AGP and PCIe, but doesn't have any vcore options, which could limit your OC. I may even try it on my Abit AS8 even though it has no official support for it, seens how it is able to run my unknown Engineering Sample, that none of my other boards will post with.

 

[NoStateofMind]

OP, nope not enough!!! Download THE ENTIRE INTERNET, PLAY EVERY GAME YOU OWN , and WATCH A MOVIE @720p , While ripping a DVD at the same time. Then come back to me, and I'll tell you what else you need to be "certified stable"

 

[Technonut]

Originally posted by: PC Surgeon
OP, nope not enough!!! Download THE ENTIRE INTERNET, PLAY EVERY GAME YOU OWN , and WATCH A MOVIE @720p , While ripping a DVD at the same time. Then come back to me, and I'll tell you what else you need to be "certified stable"

When I read stuff like the above quote, it seems like I am on the Hard OCP Forums.... Not AnandTech.

 

[ND40oz]

Originally posted by: stevty2889

Originally posted by: Duvie
Is the 955 chipset (all mobos) gauranteed to work with this????

I've even seen 865 chipset boards with Presler/Cedar Mill suport, such as the Asus P5P800..definatly doesn't require the 955 chipset, all 945/955/975 chipset boards should support them, as long as the Mfg's update the BIOS. 915/925 should work as well if even the 865/875 will support them..

Does the same go for the nForce SLI chipset?

This preslers got me thinking, I was going to buy a barebones (Shuttle SD11) for my pentium m 750, but I might just ebay it and buy a presler and a board. This 7800GT thats sitting in the box needs a home.

 

[Markfw]

Originally posted by: lopri
Well, I have to make two points.

First, the OP's OC is just awesome anyway you look at it. Whether it's the chip or his skill, hell, being able to POST @4.0GHz+ with a 2.8GHz/$280 chip is already miraculous to me.

Thats the same as OC'ing a Athlon64 from 2.0 to 2.8, and you can do that with a $160 Opteron today on 90nm process. So now they are equal ?

Can you imagine FX57 @4.0GHz? Even just a POST?

OK, comparing a P4 speed to an Athlon64 is rediculous and we all know that.
[/quote]

I understand that OP claimed "Being stable" first, but we should all first congrat his (or Intel's?) achievement. Awesome chip/job, D94!

Again, many have done this equivalent on Athlon64, so again, they are up to the same platau ? And until its stable thats not really a valid OC.

Second, I think by now we can somehow agree on "Stability". Yes, stability can be relative but then again so is everything else. If we decide to be "relative" on everything, discussion becomes impossible. That's why there are common sense and statistics. Say, if you're given a choice to pick a chip that does 12hrs Prime95 and a chip that does Super PI 32M, which one would you pick? With Prime95 unstable setup (especially LargeFFT) I have had quite a few mysterious lockups, reboots, etc. while playing games. Yes, Prime95 is not perfect but it's the best tool we can agree on for stability of a CPU. So at least for me, (and hopefully for many others) it's like this.

He has done any of the above. One hour isn't even close to proof of stability

.
1. Prime95 SmallFFT: CPU Stability
2. Super PI: Memory Configuration
3. Prime95 LargeFFT: Overall System Stablity

Bleh.. this is not what I was going to say.

Anyway, congrats again on great OCs. It's not just the OP's achievement. But it could be a sign of not-so-distant future. Think about it. Socket 939 Opteron? Do you think AMD all of a sudden decided to lose money on their chips by selling those overpowering chips at fraction of the prices? I don't think so. It means AMD's manufacturing has leaped to a different level. It's also obvious from the fact that these chips are showing a very uniform performance pattern. On the same line, this Pentium D 9xx series might just mean that Intel has finally mastered the Netburst architecture and the manufacturing. Unfortunately for them, it's like the last minute spurt since a new architecture is coming after this. Still very impressive. Just like Socket 939 Opterons.

Now there I agree, and you just said what I said above. They have finally come close to AMD. But with 65nm, and AMD did it on 90. So in that light, its not that impressive.

After all, we're observing a huge leap in CPU manufacturing that has been stagnant for such a long time. From both AMD and Intel. Arguing over stability is somewhat off the point, IMHO. If someone can POST @4.5GHz, the chip could be stabilized @4.0GHz by someone else. What's more important to us is the progress.

OK, thats just wrong. So what if I could post my X2 @ 4 ghz. Without stability its like a car without steering or brakes, worthless.

P.S. Mark, your posts are getting worse and worse.

Well, I didn't know you followed my posts so closely, but with the stuff in this post, if thats the way you believe, I can see how you could say that.

 

[d94]

now, iv been running dual foldin all night (bout 8hrs) and im going to rip TWO dvd's w/dvd shrink!

my folding is running as services but i will certainly post both logs and the dvd shrink times (Scrennies)

 

[Damn Dirty Ape]

those temps are from a 65nm? Wow -- those remind me of my P4 650 3.4ghz prescott.. I thought someone here posted an idle temp of 28c idle and 40c load from a 920

Clarification?

Originally posted by: Sentential
Pictures of the chip:
Physical Box w/t Manual
CPU w/t heatsink on manual
Size comparison against an 1800XP

This should give you an idea of temps:
Idle in Windows
Load In windows

Here are some Sisandra scores for you to view:
CPU Arithmetic
Aprox RAM Bandwith

____

Overall I am very pleased with these chips, while the clocks on them so far are kinda mediocre in time they will do alot better.

As for stability I prefer using a combination of CPU burn and looped 3d01SE to see how stable my system is. For this box I looped 3d01SE for 8 hours without any crashed and 6 hrs with CPU burn. Neither showed any abnormalities

 

[d94]

Originally posted by: Hellspawn
those temps are from a 65nm? Wow -- those remind me of my P4 650 3.4ghz prescott.. I thought someone here posted an idle temp of 28c idle and 40c load from a 920

Clarification?
he's running a chip on air that doesnt have a flat blase, intel's issued rma ## for him if he wishes to do so, that's why his temps are high

 

[remagavon]

Prime does test harder than superpi does. I can run 32m runs at 2.9ghz on an opteron 175 but my prime stable speed is 60mhz slower.. 2.84ghz.

 

[dexvx]

Originally posted by: Markfw900
Well, thats getting closer. Try 60 days, and something that stress the system more than F@H, like 2xF@H and sql and telnet and ftp and Lotus notes and sametime and Firefox and encoding all at once for those 60 days. Then talk to me. and you still spendt way more to do it (see previous post) for the same performance.

Are you saying when a new CPU comes out, and one wants to overclock it, that they should stress test it for 60 days and THEN post on the forums? Wtf kind of logic is that? AFAIK, prime95 small FTT torture test (1 instance per core) is better adept at torturing the CPU than F@H or the general applications you use.

Also, software does cause hardware lockups.

 

[Hacp]

How about the single cores? I have always considered the single core P4s a better overall chip in terms of function than the A64s.

 

[allies]

:thumbsup: That overclock is sweet!

2006 is looking like a real sweet year for Intel. I think it's sad that they were able to keep their pronounced market share for the P4 lifetime when AMD's Athlon 64 smacked it around. Oh well, I guess that's the insurance you get when you have made a huge presence for yourself.

 

[Sentential]

Originally posted by: d94

Originally posted by: Hellspawn
those temps are from a 65nm? Wow -- those remind me of my P4 650 3.4ghz prescott.. I thought someone here posted an idle temp of 28c idle and 40c load from a 920

Clarification?
he's running a chip on air that doesnt have a flat blase, intel's issued rma ## for him if he wishes to do so, that's why his temps are high

^^ What he said, but if you desire a picture take a look at this:

This is what my CPU looks like
This is what my friend's 651 looks like

Note how poor the IHSs are in terms of flatness

 

[RichUK]

Originally posted by: Sentential

Originally posted by: d94

Originally posted by: Hellspawn
those temps are from a 65nm? Wow -- those remind me of my P4 650 3.4ghz prescott.. I thought someone here posted an idle temp of 28c idle and 40c load from a 920

Clarification?
he's running a chip on air that doesnt have a flat blase, intel's issued rma ## for him if he wishes to do so, that's why his temps are high

^^ What he said, but if you desire a picture take a look at this:

This is what my CPU looks like
This is what my friend's 651 looks like

Note how poor the IHSs are in terms of flatness

I suppose with a screw down HS like a TT-BT you could force good contact, but i guess that would put too much strain and flex on the motherboard.

 

[Markfw]

Originally posted by: dexvx

Originally posted by: Markfw900
Well, thats getting closer. Try 60 days, and something that stress the system more than F@H, like 2xF@H and sql and telnet and ftp and Lotus notes and sametime and Firefox and encoding all at once for those 60 days. Then talk to me. and you still spendt way more to do it (see previous post) for the same performance.

Are you saying when a new CPU comes out, and one wants to overclock it, that they should stress test it for 60 days and THEN post on the forums? Wtf kind of logic is that? AFAIK, prime95 small FTT torture test (1 instance per core) is better adept at torturing the CPU than F@H or the general applications you use.

Also, software does cause hardware lockups.

10 minutes to an hour is nothing. At the time I posted this, that is all he had done. 12 hours is much better. (he did that) What I really want to know, is what are the full loads temps on air ! I might even dump my 820D if I could find a sucker to buy it ! (probably not now that the 9xx series hit) And can you hit this OC on a $80 motherboard like I have, or do you have to have a $200 one like he does ! As I said befre, it it takes water cooling and a $200 motherboard to get that OC, you get way better bang/buck with a neo4-f and a n X2 3800 retail.

 

[fatty4ksu]

Intel is ba...ckkkk!!!

 

[allies]

Originally posted by: fatty4ksu
Intel is ba...ckkkk!!!

:disgust:

Intel never left. I didn't hear you commending AMD on their superior processors during 2001, 2002, 2003, 2004, and 2005. Let's keep fanboyism to a minimum please.

 

[gxsaurav]

Hmm, 40 C at idle & 57C after 55 hrs at load is quite high,

although i m not sure, what AMD athlon64 X2 CPU this is comparable to, i guess at stock speed of 2.8 GHz this compares to Athlon64 X2 3800+ which has a lower overall temprature

 

[Hard Ball]

Let's see the setup of your cooling!!!!

we want pics!

 

[Hard Ball]

Originally posted by: fatty4ksu
Intel is ba...ckkkk!!!

Wow, fatty again doing what he does best; what a surprise, eh.

 

[GreatPiggo]

Originally posted by: allies

Originally posted by: fatty4ksu
Intel is ba...ckkkk!!!

:disgust:

Intel never left. I didn't hear you commending AMD on their superior processors during 2001, 2002, 2003, 2004, and 2005. Let's keep fanboyism to a minimum please.

Heh, I'm no fanboy, and I'll give you 2001, 2004 and 2005, but the 2.4 to 3.4 Northwood P4s had a clear lead there for awhile.

That's a nice overclock. My first dual-core build will probably be either a 920 or 3800+; if I see more overclocks like this, the price difference will tip me over to Intel.

And please, 30 day stability tests? That's kind of venturing into asinine levels; I don't think even the biggest of WOWaddicts would push a system that much, even while folding and encoding.

 

[dexvx]

Originally posted by: allies

Originally posted by: fatty4ksu
Intel is ba...ckkkk!!!

:disgust:

Intel never left. I didn't hear you commending AMD on their superior processors during 2001, 2002, 2003, 2004, and 2005. Let's keep fanboyism to a minimum please.

Uh no.

Willamettes were more or less on par with the T-Bird, early Athlon XP's.
Early Northwoods were better than Athlon XP's. With the C-revision and Canterwoods, they dominated the Athlon XP's.

http://www.anandtech.com/cpuchipsets/showdoc.aspx?i=1834

A64 didn't come out till late 2003. But even the first FX-51 and the A64's were only comparable to the Northwood-C/EE, performance wise. Basically, performance wise, there is little performance difference between a 3.4 Northwood-C/Canterwood (available more than 2 years ago) to a Prescott-2M 3.4/955X, available now.

http://www.anandtech.com/cpuchipsets/showdoc.aspx?i=1884

Lets keep your fanboism to a minimum please.

 

[qoou]

Originally posted by: Markfw900

Originally posted by: dexvx

Originally posted by: Markfw900
Well, thats getting closer. Try 60 days, and something that stress the system more than F@H, like 2xF@H and sql and telnet and ftp and Lotus notes and sametime and Firefox and encoding all at once for those 60 days. Then talk to me. and you still spendt way more to do it (see previous post) for the same performance.

Are you saying when a new CPU comes out, and one wants to overclock it, that they should stress test it for 60 days and THEN post on the forums? Wtf kind of logic is that? AFAIK, prime95 small FTT torture test (1 instance per core) is better adept at torturing the CPU than F@H or the general applications you use.

Also, software does cause hardware lockups.

10 minutes to an hour is nothing. At the time I posted this, that is all he had done. 12 hours is much better. (he did that) What I really want to know, is what are the full loads temps on air ! I might even dump my 820D if I could find a sucker to buy it ! (probably not now that the 9xx series hit) And can you hit this OC on a $80 motherboard like I have, or do you have to have a $200 one like he does ! As I said befre, it it takes water cooling and a $200 motherboard to get that OC, you get way better bang/buck with a neo4-f and a n X2 3800 retail.

If you can run prime 95 on small ffts for ten minutes and ten minutes on blend then that is stable. even if your computer crashes once a day, that is stable. THeres no reason to test for more than ten or twenty minutes unless it is for mission critical systems like the space shuttle.

 

[d94]

Originally posted by: Hard Ball
Let's see the setup of your cooling!!!!

we want pics!

my cam is broken @ the moment, but i do have a pic chillin on my HD:
http://img15.imageshack.us/img15/1428/dscf00304rf.jpg
http://img15.imageshack.us/img15/6099/dscf00316sy.jpg