PDA

View Full Version : New A64 3400+ Ver.06 @ Superb Overclocking


Dagalidis
10-29-2004, 11:35 AM
I just got my new CPU A64 3200+ on my hands and starting to play with my new toy...

I use watercooling by EXOS and see what this litle baby can do right out of the box....

http://www.thelab.gr/attachment.php?s=&postid=116905

One picture = 1000 words.... :D

Watch the Generation - Version and model info...

On IHS i can read 0432 means week 32 of 2004 and some rummors said that probably new sSOI was used ..... ;)

Now i try to make stable 2830+ MHZ using low Vcore (1.62 - 1.7) and i think will get soon a stable 2900+ MHZ using ONLY 1.72 Vcore....

Until now i have rock stable
1) 2650 @ 1.47 Vcore
2) 2775 @ 1.52 Vcore

and raising as Cpu Burn Proccess keep going .... :toast:

Soon i will Inform you about this project....

spaceman
10-29-2004, 11:42 AM
Sweet. Looks like they refined the process a bit on that later week. :cool:

saaya
10-29-2004, 11:47 AM
p rating 3700+? :stick: LOL :D

nice results there dagalidis! :toast:

Shroomalistic
10-29-2004, 11:47 AM
what do you use to burn in your cpu?

Dagalidis
10-29-2004, 12:03 PM
what do you use to burn in your cpu?

SANDRA burn or Prime95 Turture ;)

EDIT : this night i will use 3DMARK01 loops...

Zeus
10-29-2004, 01:17 PM
Nice CPU Dag! :up:
Looks like i have to get me one of these soon. ;)

Zejtan
10-29-2004, 01:28 PM
i dont acctually know if we have the same cpu, but it looks like it to me,
i just ran a loop in 3dmark-01 at 270x10.
highest 1M superpi so far is 234x12=2808 Mhz, on low voltage.

the diffrence is, im "on air" =)

eva2000
10-29-2004, 02:08 PM
nice!.. can you do superpi 16M for us to compare :)

Tek
10-29-2004, 03:42 PM
On IHS i can read 0432 means week 32 of 2004 and some rummors said that probably new sSOI was used ..... ;)

If that is true then it would explain why my 3700+ does 2880mhz at 1.52v, as it is also a 0432:

http://server5.uploadit.org/files/nexus2525-cpuinfo.JPG

My announce date is one day later than yours - May 31. Dunno if that means anything.

I hope it is true, but it looks like an very excellent CPU you got there :toast:

EDIT: OT - Nice program for info. First time I have actually seen my 1ghz HTL. I have 4600+ P-rating :D

STEvil
10-29-2004, 03:43 PM
wow.. wonder if a 754 version will be out? I'd dump my CG2800+ in a heartbeat! ;)

Mariachi
10-29-2004, 03:49 PM
A 754 version of what :confused:
This cpu is already 754pin, that or i'm high right now :stick:

PanaBob
10-29-2004, 04:59 PM
Dagalidis, where did ya get your CPU from?

Hallowed
10-29-2004, 05:52 PM
So 2.8 GHz with watercooling? Winchesters are doing 2.8 on air.

Still though, nice to know even the 754's are getting some help. :toast:

Can she do 3.0 Dag?

eva2000
10-29-2004, 05:54 PM
If that is true then it would explain why my 3700+ does 2880mhz at 1.52v, as it is also a 0432:

My announce date is one day later than yours - May 31. Dunno if that means anything.

I hope it is true, but it looks like an very excellent CPU you got there :toast:

EDIT: OT - Nice program for info. First time I have actually seen my 1ghz HTL. I have 4600+ P-rating :D
yeah but that's with phase change IIRC heh

IIRC my 3700+ is May 31 too

enduracell
10-29-2004, 06:20 PM
version 0.6 (from CBID) probably processed with strained silicon (http://www.penstarsys.com/editor/tech/cpu/amd/str_sil/index.html) ;)
Check ur version if ur CPU rules :D (Mine unfortunately is 0.5 and it doesn't clock well)
Version 0.6 is for Newcastles only, Clawhammer and Winchester use other ones for sure...

Mrki
10-29-2004, 06:43 PM
my NC 3400+

30 May :D

only version 0.5 ...

http://mitglied.lycos.de/siedl/3400+_cbi.jpg

Crow
10-29-2004, 06:53 PM
where can I get central brain identifier?

saaya
10-29-2004, 08:54 PM
If that is true then it would explain why my 3700+ does 2880mhz at 1.52v, as it is also a 0432:

http://server5.uploadit.org/files/nexus2525-cpuinfo.JPG

My announce date is one day later than yours - May 31. Dunno if that means anything.

I hope it is true, but it looks like an very excellent CPU you got there :toast:

EDIT: OT - Nice program for info. First time I have actually seen my 1ghz HTL. I have 4600+ P-rating :D

http://xtremesystems.org/forums/attachment.php?attachmentid=17818

Gogeta
10-29-2004, 09:08 PM
I'd like to see more than one or two Winchesters that do 2.8 on air...I haven't seen too many around here, but may not be around at the right times.

Dag and Tek - These results are amazing!

GenTarkin
10-29-2004, 10:45 PM
what version number am I lookin for on CBiD with a 3400+ CG Clawhammer? it says 2.5 would 2.6 be newer one you speak of?

CTKP
10-29-2004, 11:14 PM
where can I get central brain identifier?
http://cbid.amdclub.ru/

=[PULSAR]=
10-29-2004, 11:30 PM
Anyone else have a .06 version chip. OMG a chip with that clockability and 1MB of cache would rock, are these only being found on NC cores?

Dagalidis
10-30-2004, 02:42 AM
Yes my CPU is a piece of ART but my ANTEC TruePower PSU starting to be a piece of crap... :(

Its not possible to get 2800+ Mhz stable for heavy bench...
My +12V rail droped down to 11,48 Volts when Vcore is higher than 1.68 - 1.72 V.. :mad:
This also comfirmed by Digital Multimeter and right now the MAX stable is around 2770 MHZ @ 1.62 Vcore @ 11.58 +12V ...

When uping Vcore up to 1.7Vcore i get ONLY 2730MHZ stable.....

NEED probably to go with OCZ 520W PSU... :confused:

My cpu arrived from PC-CORNER store here in PATRAS - GREECE ... :toast:

Shroomalistic
10-30-2004, 08:30 AM
im having the same problem with my 12v rail. but for me its the mhz thats making it drop. anything over 2.6 and my 12 drops to 11.62 and prime instantly errors. I can up the voltage to the cpu but that dont help. I think im gonna mod my 12v rail soon.

Tek
10-30-2004, 08:40 AM
=']Anyone else have a .06 version chip. OMG a chip with that clockability and 1MB of cache would rock, are these only being found on NC cores?

No, they are not or at least my 3700+ is a 0432 and it is an excellent overclocker as well scaled a bit(3ghz 32M stable, can boot at 3.1ghz, and 3.2ghz screenshots). But as Eva2000 said, I have phase change and that kinda screws the comparison to theirs.

Dagalidis - Go with the OCZ 520w. I have it myself and it comes highly recommended, even though my DFI screws up the readings, it nonetheless delievers in spades.

Again, excellent CPU you got there :toast:

eva2000
10-30-2004, 08:55 AM
No, they are not or at least my 3700+ is a 0432 and it is an excellent overclocker as well scaled a bit(3ghz 32M stable, can boot at 3.1ghz, and 3.2ghz screenshots). But as Eva2000 said, I have phase change and that kinda screws the comparison to theirs.

Dagalidis - Go with the OCZ 520w. I have it myself and it comes highly recommended, even though my DFI screws up the readings, it nonetheless delievers in spades.

Again, excellent CPU you got there :toast:
btw, what camera you used to take pic of the cpu's IHS stepping/code.. it's so clear etc :)

Tek
10-30-2004, 08:58 AM
It is a Sony DVD cam. DCR-DVD201E. I love those little 1.7gb discs. They destroy mem cards :D

NoStra
10-30-2004, 03:56 PM
Measure the 12V at your molex connector...you will see that it is not your PSU that is crapping out. But the board that is sucking....

Pejsen
10-30-2004, 05:07 PM
Well it looks crazy :cool:

I have a 0430, wich is offcourse only v 0,5, that also does nearly 3Ghz stable in my prom(2987Mhz)

But then mine could just be a good clocker and not neceserely use SS?

STEvil
10-30-2004, 06:29 PM
Measure the 12V at your molex connector...you will see that it is not your PSU that is crapping out. But the board that is sucking....

They already said they measured with a DMM.

The problem is the PSU's cannot supply enough amperage on the +12v rail, not voltage. Modding the rail for more volts will only get you a very minimal increase.

ugp
10-30-2004, 07:12 PM
Here is my CPU...I am going to pull my Heatsink off and take a picture of my CPU stepping.

GenTarkin
10-30-2004, 07:34 PM
No, they are not or at least my 3700+ is a 0432 and it is an excellent overclocker as well scaled a bit(3ghz 32M stable, can boot at 3.1ghz, and 3.2ghz screenshots). But as Eva2000 said, I have phase change and that kinda screws the comparison to theirs.

So...ur sayin the Clawhammers that are not 3700+ wont be seeing the new version(strainedsilicon) anytime soon and just the lowend newcastles?....hrm time to get rid of my CG 3400+ clawhammer lol

Tek
10-30-2004, 08:39 PM
So...ur sayin the Clawhammers that are not 3700+ wont be seeing the new version(strainedsilicon) anytime soon and just the lowend newcastles?....hrm time to get rid of my CG 3400+ clawhammer lol

What I am saying is, I think that all 0432 - be it Clawhammers or Newcastles, whatever the P Rating, probably has strained silicon :)

If that is true, then my 3700+ would be a s754 limited FX-55.

eva2000
10-30-2004, 08:41 PM
What I am saying is, I think that all 0432 - be it Clawhammers or Newcastles, whatever the P Rating, probably has strained silicon :)

If that is true, then my 3700+ would be a s754 limited FX-55.
nah more likely it's just newer 3700+ are more mature like any later/improved cpus

Tek
10-30-2004, 08:53 PM
You're probably right :)

Geforce4ti4200
10-30-2004, 09:09 PM
I have not heard amd using strained silcon on anything but the fx55 and the occasional fx53 or 4000+ that didnt pass as a fx55. I think those high 2GHz clocks with water is normal and just evidence of amd's new improved stepping. strained silcon is even better!

"Winchesters are doing 2.8 on air."

probably a 3500+ right? what can I expect from a good 3000+ winchester

WeStSiDePLaYa
10-30-2004, 09:38 PM
hmmm, i wonder if the mobile newcastles will be recieving SS? 1.8ghz with 1.2vcore at stock, add some SS to that and your looking at some really high clocks.

Zebo
10-30-2004, 10:24 PM
Just get an enermax 465... no need to spend high dollar on OCZ since it has just as high a massive 33A 12V rail. I burned my antec "true" 480 then subsituted for enermax (thanks newegg you so kind) and have been rock stable ever since.

Here is a screen while playing farcry for 15 min (http://www.csc.calpoly.edu/~smanning/enermax.jpg)..all my stuff is crazy OCed plus water pump plus a bunch of 120mm fans plus 4HD's etc

Zeus
10-31-2004, 01:04 AM
Here is my CPU...I am going to pull my Heatsink off and take a picture of my CPU stepping.

Oi, week 45 is a pretty young one, how does it OC?

Edit: oops week 45-2003.

Just get an enermax 465... no need to spend high dollar on OCZ since it has just as high a massive 33A 12V rail. I burned my antec "true" 480 then subsituted for enermax (thanks newegg you so kind) and have been rock stable ever since.

Here is a screen while playing farcry for 15 min (http://www.csc.calpoly.edu/~smanning/enermax.jpg)..all my stuff is crazy OCed plus water pump plus a bunch of 120mm fans plus 4HD's etc

I have that same PSU and the 12V rail seems to be ok compared to other PSU's although 12V seems to be a little low even at idle according to SmartGuardian.
It says 11.83V idle and under load it drops to 11.71V.
Measured by DMM it's alittle above 12V
But...this is only at 1.55Vcore.

I will measure 12V with DMM under load with some 1.7Vcore and see how much it drops.
I think the higher the Vcore the more the 12V rail is taking a dive, no?

STEvil
10-31-2004, 01:48 AM
if the PSU cant handle it, it will.

ugp
10-31-2004, 06:56 AM
Add higher clock speeds I get memory errors in MemTest show it shows me to believe it is my CPU.

bldegle2
10-31-2004, 07:45 AM
clockage.........incredible.

wandering off now, making a note to myself to get the prommie back up and running again................................

where does one get an a64 attachment setup for the Prometia's???

baldy :D

Tek
10-31-2004, 08:38 AM
Just get an enermax 465... no need to spend high dollar on OCZ since it has just as high a massive 33A 12V rail.

Isn't the 520W 12V rail 38A? :confused:

Dagalidis
10-31-2004, 04:00 PM
Still can't get it stable higher than 2755 - 2775 even with 1.92 Vcore....
Please note that at 2775 @ 1.55 Vcore is rock stable.....
If add 1 more MHZ then i get random Restarts.... :(

WTF... what is going on here ... what i'm doing wrong ... :confused:

Tek
10-31-2004, 04:50 PM
Still can't get it stable higher than 2755 - 2775 even with 1.92 Vcore....
Please note that at 2775 @ 1.55 Vcore is rock stable.....
If add 1 more MHZ then i get random Restarts.... :(

WTF... what is going on here ... what i'm doing wrong ... :confused:

Don't know if this can be of any help, as I am new to AMD64 and OC in general.

I have problem getting same excellent results(3ghz/1ghz) if I lower the multiplier to x11 or x10. The lower it gets, the more instability I have experienced regardless of HTT and RAM multiplier/divider etc. I can do 32M stable at 11x273mhz, but then one mhz more and it fails. Never seen above 10x27?, yet.

Also, my CPU doesn't like high voltage. Anything above 1.76v and I get instability, and I must lower the HTT to x3. Also, I have the chipset set to 1.8v and the AGP to 1.6v. Chipset voltage at least helps my x4 HTT, as 1.6v is enough for x3.

Try 12x230 or with 11x240, HTT x3 and clockgen FSB from there. You'll get 2.8ghz - 2.9ghz, I think and hope :)

Zebo
10-31-2004, 06:12 PM
Zeus, never belive on board monitors. Get yourself a multimeter for $20 measure 12 while gaming then see.

Tek, at a price;) At twice the price. OCZ/Zippy/PCP&C are all the best but enermax with it's high 12's is right there at Antec priceing.

Zeus
11-01-2004, 12:01 PM
Zeus, never belive on board monitors. Get yourself a multimeter for $20 measure 12 while gaming then see.

Tek, at a price;) At twice the price. OCZ/Zippy/PCP&C are all the best but enermax with it's high 12's is right there at Antec priceing.

Did the test:

PSU: Enermax 465w 33A @12V

Idle @ 1.76Vcore SmartGuardian-11.83V, DMM -12.16V (molex)
Load @ 1.76Vcore SmartGuardian-11.67V, DMM - 12.10V (molex)

Conclusion: Don't trust SmartGuardian! :)

macci
11-01-2004, 12:41 PM
Did the test:

PSU: Enermax 465w 33A @12V

Idle @ 1.76Vcore SmartGuardian-11.83V, DMM -12.16V (molex)
Load @ 1.76Vcore SmartGuardian-11.67V, DMM - 12.10V (molex)

Conclusion: Don't trust SmartGuardian! :)
Molex reading and mobo reading are two different things! Molex reading tells you how much volts your giving and mobo reading tells you how much its actually getting. There is always a voltage drop there (because of connector resistance and mobo internal resistance). The lower the voltage drop the better.

Rabbi_NZ
11-01-2004, 01:17 PM
I am about to buy a 3000+ DTR, are Mobiles/DTRs usually younger steppings (more recent) than Desktop versions or vice versa?
It's a 754 CG stepping if that matters... and I think it has 1MB but not confirmed yet...

feesh
11-01-2004, 01:29 PM
Very nice results. Where did you guys get your 0432 cpu's?

Zeus
11-01-2004, 02:15 PM
Molex reading and mobo reading are two different things! Molex reading tells you how much volts your giving and mobo reading tells you how much its actually getting. There is always a voltage drop there (because of connector resistance and mobo internal resistance). The lower the voltage drop the better.

Yeah, there's definately a difference between the voltage through the mainboard's circuit and on the molex but this test was more to show that my 12V rail of my PSU seems to handle it nicely.

Going from the figures that SmartGuardian shows one could doubt it's PSU.

I think a 0,06V drop on the molex is quite ok, don't you think?

STEvil
11-01-2004, 06:07 PM
Yeah.

Whats bothersome is the larger drop on the mobo.. that indicates a large current resistance which would make for horrendous overclocking possibly.. or a burned up trace.

Dagalidis
11-02-2004, 12:07 AM
Still on Burning Process here..... :D
http://www.thelab.gr/attachment.php?s=&postid=119405

http://www.thelab.gr/attachment.php?s=&postid=119410

BTW my CPU is week 34 of 04 :D and not 32 as mentioned before by mistake .....

=STE=BloodAngel
11-03-2004, 01:58 AM
ADA3400AEP4AX
CBAZC 0434MPMW
1061822I40513

Version 0.6

=STE=BloodAngel
11-03-2004, 02:08 AM
And another one :banana: :banana4: :banana:

Crankster
11-03-2004, 02:54 AM
Ya passed the 2.9 gig wall?
Grats!

=STE=BloodAngel
11-03-2004, 03:10 AM
This is another cpu as the one in my signature.

I ordered a 1Mb cpu and got this one :mad: . The wall is still there I think. I can do a couple of good Pifast runs. But sometimes it just gives an error in computation. Temps are ok, increasing Vcore doesn't work. Already tried 1.72V.

Mrki
11-03-2004, 09:55 AM
my 3400+ NC is arrived today ;) ...

i will now testing!

0433 VPMW

http://mitglied.lycos.de/siedl/CBI_NC_0433_VPMW.jpg

TEDY
11-03-2004, 10:28 AM
what is so special about this 0.6 and where did you order it + price ? 10x

jinu117
11-03-2004, 10:33 PM
Yeah where are you gusy getting 0.6? I got burnt out by winchesters not booting up on phase cooling so looking back at some 754 procs.

Dagalidis
11-04-2004, 01:09 AM
Those NC 0.6 Version are the best cpus ever had... ;)
Also no problem with high FSB - HTT ...
But i need a new PSU to get it stable there... :( http://www.xtremesystems.org/forums/attachment.php?attachmentid=18903

pompeyboy
11-04-2004, 01:11 AM
My version 0.5 seems to be as good as the version 0.6, i managed to get SS of 3ghz on mine but cant get it stable at that speed , at 2880mhz i dont have any probs with stability, i have run this speed 24/7 for a week no problems(3400+ newcastle btw).I will look into this 0.6 version a bit more me thinks :)

Dagalidis
11-04-2004, 01:18 AM
My version 0.5 seems to be as good as the version 0.6, i managed to get SS of 3ghz on mine but cant get it stable at that speed , at 2880mhz i dont have any probs with stability, i have run this speed 24/7 for a week no problems(3400+ newcastle btw).I will look into this 0.6 version a bit more me thinks :)

2880 stable Overclock is just amazing my friend... ;)

BTW.. Welcome to the XTREMESYSTEMS :toast:

bias_hjorth
11-05-2004, 04:45 PM
24/7 stable for now + Cpu shot.
http://www.oc-central.dk/tobias/cpiden.gif
http://www.oc-central.dk/tobias/cpushot.gif
benchshot - memory crippled cause this cpu wants more :D (http://www.oc-central.dk/tobias/29.jpg

eva2000
11-05-2004, 05:14 PM
24/7 stable for now + Cpu shot.
benchshot - memory crippled cause this cpu wants more :D (http://www.oc-central.dk/tobias/29.jpg
sure you have enough vcore there lol

bias_hjorth
11-05-2004, 05:17 PM
the cpuz readout is wrong.. its around 1.87 and yes :D Scales through 3ghz with 1.75 and I was playing around.. Heck I got the cooling to support it :)

jinu117
11-05-2004, 06:58 PM
Good stuff... I just got 3500+ v0.6 and it got to prime 95 stable with stock v-core to 2.75ghz :) Time to start upping volt and climbing. (Mach II R404 cooled) Even with dividers etc, I might need to start looking at TCCD memory if it scales to what I think it might be able to :)

[XC] serlv
11-05-2004, 07:56 PM
So, I must be the 4th or 5th person to ask. Where, what resellers are these chips coming from? :confused:

Comeon, out with it. We want some, too. :)

Riverna
11-05-2004, 08:12 PM
nice results

jinu117
11-05-2004, 08:20 PM
So, I must be the 4th or 5th person to ask. Where, what resellers are these chips coming from? :confused:

Comeon, out with it. We want some, too. :)

woops... sorry... got it locally at pcclub.
They even have X800XT PE @ 550 or so. Now price was steep but I had to RMA motherboards which they would have taken 15% restocking fee which I talked them out of ... so pretty much average price I paid for :P 2.89ghz 1.584v priming now. I see 3ghz priming soon... Once I hit the limit I will start posting bunch of results.

Riverna
11-05-2004, 08:40 PM
It is not stable, my Board run only with 270FSB stable :(

http://www.forumdeluxx.de/gallery/data/500/8886maximum1.JPG

Zeus
11-06-2004, 01:40 AM
Good stuff... I just got 3500+ v0.6 and it got to prime 95 stable with stock v-core to 2.75ghz :) Time to start upping volt and climbing. (Mach II R404 cooled) Even with dividers etc, I might need to start looking at TCCD memory if it scales to what I think it might be able to :)

Well, if your s 939 is ver0.6 then it's very likely every ver0.6 has got the stained silicon.
Good news!
I see for NC it's v0.6 and for CH it's v2.6

Bias_hjorth,
I thought these ss CPU's were supposed to be low voltage? :D

bias_hjorth
11-06-2004, 04:04 AM
Bias_hjorth,
I thought these ss CPU's were supposed to be low voltage? :D


Not anymore :D


edit: 3ghz 24/7 stable :slobber: :D
http://www.oc-central.dk/tobias/3ghzst.gif

Zeus
11-06-2004, 07:07 AM
3.599v??? :confused:

bias_hjorth
11-06-2004, 07:10 AM
readout is beeing shrewed up past 1.8v

=[PULSAR]=
11-06-2004, 07:31 AM
bias hjnorth wow thats pretty impressive what kind of cooling?

jinu117
11-06-2004, 10:38 AM
Just a little short of 3ghz ATM with 3500+ 0.6 newcastle. Funny thing is, 1.65 will do the job till 2941mhz. Anything after that (up to 1.85v) won't do squat for stability.
In fact after trying few different LDT and FID, it seems like my board isn't too stable around 270FSB. Using bios 1.36b. I have feeling this CPU has a lot more to it than what it is doing now considering I am not in that high of volt. Now, how do I step over this 270FSB instability? (Basically prime 95 might run for few seconds to 30 mins around 270+ FSB range and system crash around 275-280FSB even when I set my FID so I know I am in safe zone for CPU speed). I would like to rule out memory (OCZ 3700EB 512x2) but not quite sure what can I do :P (It is on 166mhz setting ATM since I know my EB can't scale that high at timing). How I wish I had upward FID unlocked on this case... :)

bias_hjorth
11-06-2004, 10:58 AM
=']bias hjnorth wow thats pretty impressive what kind of cooling?
Prommy @ r404a

jinu117
11-06-2004, 11:15 AM
Prommy @ r404a

Rest of system spec please? :P
Motherboard, memory, PSU especially :)
Just wondering why am I hitting this 270ish FSB wall when my CPU is already doing 100% stable 2941mhz @ 1.65v (bios... actual readout during 100% load is 1.68 to 1.696v)
Could it be 2 sticks of EB? Could it be just not as good MSI as others? (I dunno... I've heard way north of 300mhz too often to think 270ish is the end of mobo...)
Please let me know :)

bias_hjorth
11-06-2004, 11:43 AM
sure: ballistics pc3200 (2.5x2x2x0), crappy 8kda3, q-tec 550w psu -
270htt is your max? tried lowering htt multi?

jinu117
11-06-2004, 11:50 AM
sure: ballistics pc3200 (2.5x2x2x0), crappy 8kda3, q-tec 550w psu -
270htt is your max? tried lowering htt multi?

Sure did. LDT (HTT multi x2, x2.5, x3 x4 has been used)... It seems to be issue with FSB itself for some unknown reason after playing with FID as well. IE) 270x11 = prime failure, 270x10 = prime failure, 270x9 = prime failure while 267x11 267x10 267x9 passes prime with flying colors.

Here is screenie of my overclock for time being (100% stable from what I've tried so far...)

http://home.comcast.net/~jinwpark/2941.gif

SoundWave
11-06-2004, 11:55 AM
sure: ballistics pc3200 (2.5x2x2x0), crappy 8kda3, q-tec 550w psu -
270htt is your max? tried lowering htt multi?

What's your Vdimm at that speed, and you're not running 2,5-2-2-0 at 272 are you?

bias_hjorth
11-06-2004, 11:56 AM
Actually I had the exact same issue on the old k8nnxp. I solved the problem by booting at 270 and upping the htt from clockgen. Not the perfect solution so essencially I bought a new board.

bias_hjorth
11-06-2004, 11:57 AM
What's your Vdimm at that speed, and you're not running 2,5-2-2-0 at 272 are you?


3.01vdimm . No I am running with a 183mhz ratio (around 242 as far as I can calculate)

jinu117
11-06-2004, 11:58 AM
Hmmm @ 3.2v this stick has done me only 3-2-2-8 on 255mhz on my socket 754. Haven't fully tested it on this rig yet. From initial look of it, it only seems to be 100% stable @ 250FSB on dual channel on MSI for 3.2v. I am running on 166mhz right now to test CPU. But I think I am stuck since I think I am FSB throttled to use different multi to get the ram up. Thus the reason wondering why my FSB is limited. (memory or board itself... hmmm)
Maybe I will use single dimm or some other setting to figure this out.
How high of dual channel 1gb did people able to achieve on MSI K8N Neo2 as far as FSB is involved? (I could care less about memory divider).

bias_hjorth
11-06-2004, 12:06 PM
Hi bias_hjorth!

As you see, CBId incorrectly shows a core voltage for your CPU. This occurs because the Brain doesn’t support W83627THF I/O chip well. A label that is located above the Processor button shows an I/O chip ID. It helps me to find the bug with determining of the core voltage and CPU temperature. I hope I have fixed a bug this time. Download the newest build: http://cbid.amdclub.ru/files/cbid73t.zip

http://www.oc-central.dk/tobias/247.gif
Newest build. Although the voltage is def. a board issue since the board cant registrer over 1.93 vcore. (1.86mosfet)
Btw. keep up the superb job on this program :)

bias_hjorth
11-06-2004, 12:09 PM
Hmmm @ 3.2v this stick has done me only 3-2-2-8 on 255mhz on my socket 754. Haven't fully tested it on this rig yet. From initial look of it, it only seems to be 100% stable @ 250FSB on dual channel on MSI for 3.2v. I am running on 166mhz right now to test CPU. But I think I am stuck since I think I am FSB throttled to use different multi to get the ram up. Thus the reason wondering why my FSB is limited. (memory or board itself... hmmm)
Maybe I will use single dimm or some other setting to figure this out.
How high of dual channel 1gb did people able to achieve on MSI K8N Neo2 as far as FSB is involved? (I could care less about memory divider).

What superPI are you pulling at that speed?

[XC] serlv
11-06-2004, 01:28 PM
Bahh, humbug! My local PCClub only has up to about week 27, 28. The CompUSA next door doesn't even have any, at all.

Where to get these post week 32, rev 0.6 3400+'s? :(

Anyone know of any online resellers where they are coming out of, for sure?

jinu117
11-06-2004, 06:33 PM
What superPI are you pulling at that speed?

Was checking few things... since that is my 24x7 setup parameter, once I am done with it I will test. I bet it sucks for that mhz due to memory hold up... don't expect much :) (BTW, I got over HTT bottle neck and got it prime 95 stable to 3.05ghz... however, that could only be done with single stick which is not what I was looking for... so... I think my daily setup is going to be the final setup as is now unless some memory or somethig else I didn't think of changes dual channel HTT blockage around 270FSB). Another interesting thing is I think DDR booster is negatively impacting my memory's stability... wierd. Worked just fine on my ASUS K8NE Dlx.... it shows voltage fine, etc but I guess I get less mhz out of memory @ 3.2, 3.3v etc thann when I am @ 2.7v... strange.. strange... About time the much waited product i mod section is available... :)

Here is the screenie... only pulling 30 seconds... Now let me put my lame excuses :) This is daily rig with whole bunch of stuff installed, windows xp pro, etc, etc. I still haven't created my benching HDD yet. (anyone got image I can just steal? :P) Also, memory is running @ 210.6mhz 3-2-2-10 :)
I tried single channel with 253mhzish memory (3-2-2-10) and 3.05ghz CPU and got 32 seconds... guess dual channel helps on super pi a LOT.
I know I got room for improvements and probably can knock down to 29 maybe even 28 by just optimizing for benching... but I am proud to get into 30 with 24x7 prime 95 stable, 3dmark stable, etc, etc machine @ this point :)

http://home.comcast.net/~jinwpark/mainrig/2.94ghzSuperPi.jpg

Dagalidis
11-07-2004, 07:15 AM
Note: To be sure you will get an A64 CPU 754 Socket Version 0.6 Just make sure you will get From week >= 33.... ;)

WeakSauce
11-08-2004, 10:17 PM
Here is what I got out of my 3400+ NC ver.6. It's a 0432 stepping. Pretty good for stock voltage. I am having trouble with my 2x512's even though they are bh5. I can't get much higher than 230 with cpc enabled and I don't want to disable it because it just kills 3d and I play tons of games. Anyway just thought I would share with you guys. I'll post some super pi 1m and 3dmark2001 at this speed too.

jlccarv
11-08-2004, 10:28 PM
Hot dang! ^^ :eleph:

Very impressive clock!

WeakSauce
11-08-2004, 11:21 PM
here's a 3dmark run at same speed. I'm happy with it.

jinu117
11-08-2004, 11:40 PM
here's a 3dmark run at same speed. I'm happy with it.

I need to learn how to get good score on 3dmark... only getting around 25k with 3ghz and GT overclocked to Ultra speed... sigh. Must be quite a few points missing somewhere.

WeakSauce
11-09-2004, 12:50 AM
I would think that you should get a much higher score than that with what you got going. I just did a 3dmark2001 and aquamark run at 250x11 (2.75ghz) with cpc enabled at 1.6vcore and 3.4vdimm and got 24,989 on 3dmark and 57,200 on aquamark. That's with my 9800xt@447/399. You should get much higher than me I would think.



Oh btw i'm idling at about 39C and prime load and about 53C, that seem high with my cooling? I'm on 10/15 bios with 4vdimm.

kakaroto
11-23-2004, 06:15 AM
what about 0.6 CPU @stock winchester voltage :p:

boxed cooler.

idle = 33C!!

unstoppable
11-24-2004, 07:47 AM
Note: To be sure you will get an A64 CPU 754 Socket Version 0.6 Just make sure you will get From week >= 33.... ;)

Not all of them are 0.6, according to CBI my 3000+ 0440 is ver. 0.5 .

shimmishim
11-24-2004, 09:55 AM
Bahh, humbug! My local PCClub only has up to about week 27, 28. The CompUSA next door doesn't even have any, at all.

Where to get these post week 32, rev 0.6 3400+'s? :(

Anyone know of any online resellers where they are coming out of, for sure?

i got a week 33 3400+ (754) from newegg last week!

it does almost 2.8 ghz at 1.55 volts on my prommy (unmodded).

takes almost 1.75 to hit 2.95 stable


also.... how do you determine if it's 0.5 or 0.6? thanks!

Tek
11-24-2004, 10:17 AM
Not anymore :D


edit: 3ghz 24/7 stable :slobber: :D
http://www.oc-central.dk/tobias/3ghzst.gif

Hey Bias_Hjorth, very nice OC you have there :toast:

Can it do 12x250 / 1ghz HTL stable? Also, why is your voltage so high(not the +3v reading, but the actual +1.8v), compared to my own 3700+? Is it because you use 11x273, where I use 12x250?

Anyways, very nice. I think, you'll break my own 3.2ghz screenie very soon :p:

perkam
11-24-2004, 10:41 AM
I believe they're bringing the 3400+s over to the 939 pretty soon. At least I think. and its OCs like that which make those chips all the more attractive to invest in...

Way to Go Hjarth !! :D

Perkam

Zeus
11-24-2004, 10:47 AM
These ver.0.6 are some awesome overclockers.
Here's mine watercooled at only 1.7V:

http://www.xtremesystems.org/forums/attachment.php?attachmentid=19840

Just plain water, nothing exotic. :cool:

jinu117
11-24-2004, 10:48 AM
Ahhh just checked this and found out my posted 30second is pathetically slow compared to what I get now. 3.05ghz and changed memory to TCCD and I am in 28 seconds. (Check the mbot thread for screenie:P)
Still haven't done tweaks.
Can't go faster than that even on suicidal run as I do believe I need about 3.15ghz for additional 1 second where 287 mhz on dual channel on my pair of memory is stressing it unfortunately. (memory unstability seems to effect me more than cpu instability as this cpu can bench all the way till 3.2-3.3ghz on what I am benching on but memory just can't keep up)
Honestly guys, how much does tweaking your system helps you on superpi? I've even ran it with steam up (okay I forgot) and get same score. Do you get anything for tweaking at all? (I do know 3dbenches get impacted)

kakaroto
11-24-2004, 11:50 AM
nice z3us :D 43sec for water is sweet!!

Here some of my CPU :D

http://www.xs4all.nl/~famchow/A64_3400/3400+%200.6/Pifast_3.1GHz_11.JPG

3.1GHz is completely stable

perkam
11-24-2004, 11:52 AM
Hey kakaroto and Zeus, are those 512 KB 2.4 Ghz 3400+s or the 1 MB 2.2 Ghz 3400+s ??

kakaroto
11-24-2004, 11:54 AM
both 3400+ MPMW 0434 :p:

perkam
11-24-2004, 11:56 AM
both 3400+ MPMW 0434

"Cant process data, please re-input" ???

Perkam

Zeus
11-24-2004, 12:04 PM
both 3400+ MPMW 0434 :p:

He meant both NewCastle's.
My old 3400+ could not even come close to this one, hit a brick wall at 2,55GHz.

K, you didn't do to bad yourself mate. :toast: Any left?

Rabbi_NZ
11-24-2004, 12:11 PM
Is it possible to get a Version 0.6 ClawHammer?

Zeus
11-24-2004, 12:15 PM
Is it possible to get a Version 0.6 ClawHammer?

Yes but then they are version 2.6
Your only chance will most likely be the overpriced 3700+.

I would love to have a CPU that does the same clocks like the one i have with 1mb cache. :slobber:

Rabbi_NZ
11-24-2004, 12:19 PM
Yes but then they are version 2.6
Your only chance will most likely be the overpriced 3700+.

I would love to have a CPU that does the same clocks like the one i have with 1mb cache. :slobber:
Thanks Zeus, what are the chances of a 0437XPMW 62w Mobile 3400+ 2200MHz 1MB being a 2.6?
I have one coming my way real soon :)

mbm
11-25-2004, 01:34 AM
where do I check my ver. number?
I have a 3400+ NC -I can get it stable over 2700mhz.
CPUZ dont read my Vcore. Has this something to do when I havent installed the AMD64 driver?
What does this do anyway -Isn´t it just for Cool`n`quiet?

bias_hjorth
11-25-2004, 02:24 AM
Hey Bias_Hjorth, very nice OC you have there :toast:

Can it do 12x250 / 1ghz HTL stable? Also, why is your voltage so high(not the +3v reading, but the actual +1.8v), compared to my own 3700+? Is it because you use 11x273, where I use 12x250?

Anyways, very nice. I think, you'll break my own 3.2ghz screenie very soon :p:

Ahh sorry I didnt see your question before now. Well I havent really tried anything big besides the one in my signature. I truely can run the system for everyday above 3ghz. The reason the vcore was way off were due to a wrong sensor detection. I´ve been PM´ing back and forth with VVJ (owner of the awesome CPU identifier) and we figured the problem out which should be solved in the newest revision :)
I can highly recommend the program any day :)

http://cbid.amdclub.ru/

edit:

Here´s one showing the right voltage:
http://www.oc-central.dk/tobias/24-7.gif

mbm
11-25-2004, 01:47 PM
OK fund the data on the CPU:
ADA3400AEP4AX
CBAXC 0433RPMV
Newcastle ver. 06

but what is holding me back then.
Can only do 2700Mhz stabil.
My temp are below 50C on load.

Rabbi_NZ
11-25-2004, 01:52 PM
Mobile 3400+ 1MB 62w @ NewEgg (http://www.newegg.com/app/ViewProductDesc.asp?description=19-103-477&depa=1)
from the picture it is Week 0437XPDW... I know someone else that just got a 0437XPMW from ExcaliberPC and it is running 2700MHz + with less than 1.8v... on air cooling. They are also 130nm SOI Version 2.6 :)

kakaroto
11-25-2004, 01:57 PM
He meant both NewCastle's.
My old 3400+ could not even come close to this one, hit a brick wall at 2,55GHz.

K, you didn't do to bad yourself mate. :toast: Any left?

yes there is some left :)

http://www.xs4all.nl/~famchow/A64_3400/3400+%200.6/superPI8M_3.1GHz.JPG

But the problem why I hardly can bench 310x10 is some bug on the board.
310MHz mem is stable, but with 310x10 not... BSODS ...

VVJ
02-20-2005, 10:18 AM
Hey guys, it's interesting to know which Version has the newest CBBID revision of Athlon 64!

kakaroto
02-20-2005, 10:33 AM
Hey guys, it's interesting to know which Version has the newest CBBID revision of Athlon 64!

version 3.1

shuRe
02-20-2005, 12:37 PM
weaksause how come ur bandwidth is so low? im running the following setup on air and its 500mbs better off..

http://img80.exs.cx/img80/5981/24k3dmarks20016sr.th.jpg (http://img80.exs.cx/my.php?loc=img80&image=24k3dmarks20016sr.jpg)

VVJ
02-20-2005, 10:59 PM
kakaroto, the same. Perhaps, there is no significant difference between the newest CBBID revision and previous revisions of the Winchester core.