The Official X2 Overclocking Discussion Thread

[alpha0ne]

A 3800+ cannot be a toledo... Toledo means 2x1MB cache... The 3800+ are *ALL* manchester with 2x512KB cache...

ALL E6 revision: ADA3800DAA5CD are Toledo with half the cache disabled

E6 revision: ADA3800DAA5BV are Manchester

[Jupiler]

Nice chip

Can you help me? Please post screenshot of CBid detected your Manchester(download here).

Because CBid detected my Manchester version 4.2

here http://xtremesystems.org/forums/showthread.php?t=90422

Thanks alot.

Here you go.

[[cTx]Philosophy]

SO far this thing is finally starting to clock nice..
I might just have to sell my opty and run with this if she keeps scaling

Do u guys think this is good prime dual stable on 1.55?
http://valid.x86-secret.com/show_oc?id=74775
oooh BTW im not done its just 5:21 am and ima let her burn overnight..
this is CCBWE 0538 MPMW toledo 3800

[iMAX]

Here you go.

Thanks guy,

But what happen, with my Manchester rev. E4 CBid 0205 detected version 4.2

[Dagalidis]

I love my cpu when running Cool....(XP120)

In a few days will be watercooled...

[ManagHead]

Tomorrow I'll get a 3800+, think I get LDBHE 0601, hope it will handle the cold..

[rui]

Hi to all.
This is my first post here, but been reading for some time. My name is Rui and i am from Portugal.

I am posting here in absolute desperation.
First, let me post the facts.

I have a MSI K8N Neo2 Platinum with the 1.C3 Sideeffect bios and an AMD Athlon X2 3800+ CPU (10x200=2000), with stock cooling. According to CPU-Z, the CPU is a Toledo Core, not the Manchester core, so I am assuming that it has 512KB cache disabled in each core.
My PSU is an Enermax EG365P-VE with the following specs: Max continuous output 350W, 32A on the +3.3V line, 32A on the +5V line and 26A on the +12V line.
I also have a NVIDIA Geforce 6800GT, 2GB of RAM (1 dimm Kingston 1GB and 1 dimm Corsair 1 GB, on dimm slot’s 1 and 2), both dimm’s are at 3-3-3 timings, and always running at speeds at or below specs (200Mhz).
I also have 2 hard drives on each the IDE channels, and a DVD-RW +/- on the second IDE channel.

I am trying to overclock the CPU, and first, to discover the max FSB (HTT), I choose the ram divider of 166 Mhz, then choose the HT multiplier of 3, then choose the CPU multiplier of 7.
The max FSB I got stable (passing Prime95) was 235Mhz.
This was a big disappointment to me, because I read all over the net people getting bus speeds of +250.

Also, when sticking with that 235 Mhz FSB, and then upping the CPU Multiplier, o only got to 8X multiplier (this gives 1880Mhz). Because if I choose 9X multiplier (2115), the system cannot pass Prime95. All this with the core voltage of +1.5, by choosing the CPU VID of 1.55 and then +8,3%.
(Speaking of core voltage, shouldn’t I get a much higher core voltage, since 1,55+8,3 % should give me 1,67, but in CPU-Z I can only see values of 1,55, more or less. It seems that the setting CPU VID isn’t doing anything, the only thing that changes the core voltage are the % increments that one can choose.)

BUT if I low the FSB to 220, and then choose the CPU multiplier of 10X (2200), the system is stable and passes Prime95 test.
So it seems that my CPU can do at least 2200, but not 2115, when the FSB is at 235. But on the other hand, if I only choose FSB 235 and CPU speed is low, the system is stable. Remember that the HT multiplier is at 3X, so I am no where near the HT limit of 1000Mhz.
If i leave the CPU multiplier at 10X, and start increasing FSB, i can only go to 225, with a stable system, but i have to choose a VID of 1,55 and a 8,3% increase. The core voltage is resported by CPU-Z to be around 1,57, more or less. The temps at full load go up to 58º celsius.
Amazing isn’t it?

Can it be that I got extremely unlucky and have both a motherboard and cpu that are overclocking unfriendly, or is something wrong with my system?
All opinions will be greatly appreciated.

[smids]

Could be power supply as that 365W PSU won't be friendly on a dual core which you are trying to overclock, especially with a 6800GT which is a power demanding card.

You do not need worry about the HTT frequency of 1000, so long as you are above something like 700mhz as it makes very little difference.

It might be the memory controller on the CPU holding you back, especially given the weird results you have been getting. Have you set the AGP/PCI lock in BIOS also, by setting the AGP/PCI frequency to 34?

[StoneRhino]

Hey guy good news and bad news.

I got a new 4400+ X2 CCBWE 0604VPMW. Major problem is the thing is rock solid stable at even stock voltages 1.335 volts at 11 * 250. 2750mhz but it won't let me go to 251. It freezes at windows load in screen. Is this the cold bug problem everyone is talking about I didn't have this problem with my last 4400+. I know this chip has way more overclocking potential though I just need to get the FSB past 250. My last chip was stable to 315 with the settings I have now.

I already have weakest drive and data strengths.

I tried all the voltages up to 1.75 to see it it makes a difference nope.

I run dual prime95 for stability testing I run for 8 hours.

Suggestions?

PS my last chip was a CCBWE 0527RPMW. It was a bad overclocker only getting 2850 mhx stable at full load at -15 core temp. I have a custom phase change SS unit. Really powerfull. It is in jewlers shop getting pin put back on .

[rui]

Could be power supply as that 365W PSU won't be friendly on a dual core which you are trying to overclock, especially with a 6800GT which is a power demanding card.

You do not need worry about the HTT frequency of 1000, so long as you are above something like 700mhz as it makes very little difference.

It might be the memory controller on the CPU holding you back, especially given the weird results you have been getting. Have you set the AGP/PCI lock in BIOS also, by setting the AGP/PCI frequency to 34?

Thank you for your post.

The power supply isn't very strong, but since it's is a Enermax, and it has on the +12V rail about 26A, i thought it would be enough. But, like you said, it's a dual core, and the Geforce IS demanding.

I have the AGP speed at about 33, because i thought that the 34 for the AGP/PCI lock was only for the first bios of this motherboard. But i will try that when at home.

On the other hand, it's strange that the system can't stand Prime95 when CPU is about 2115 (235*9), but can do it at about 2200 (220*10). Like you said, something is holding it back.
I read at Anandtech that almost all X2 3800+ where of the Manchester core. But since AMD had some Toledo core's that had bad caches, they disabled half the cahce and sold them as X2 3800+ CPU's. Mine, like i said, is a Toledo core.
Could it be that these X2 3800+ toledo core's are low quality CPU's that couldn't pass tests for being sold as X2 4200+? That would explain the very low overclocking capacities.
I forgot to mention on my first post that if i don't choose sucha a high voltage (+1,5) when overclocking, i can't even boot to windows.

Again, all opinions will be greatly apreciated.

[ManagHead]

Argh, got CCBWE istead of LDBHE

Dual prime at 2500mhz 1,4V atm..

[don_vercetti]

got a very nice looking LDBHE 0601 UPMW

AMD64 TCaseMax - v1.18
----------------------

CPU Information (CPU #1):
Standard CPUID: Family: F, Model: B, Stepping: 1
Extended CPUID: Family: F, Model: 2B, Stepping: 1
CPUID String: 20FB1
Processor APIC: 0
Processor: AMD Athlon 64 X2 3800+ (Manchester)
CPU Speed: 2009.32 (200.93 x 10.0)
Revision: BH9-E4
Platform: Socket 939
Startup VCore: 1.350v
Maximum Case Temperature = 69C
TDP: 84.4 Watts

Check out the max tcase! The TDP is very nice indeed.


(p.s. sorry for the bad photo - is from my phone)
Just running it at stock to check everythings working nicely. Currently loading at 27c with dual sp2004, on my WC. All in all, looks good.

[Absolute_0]

Tell us how that 0601 clocks. I'm seriously considering getting an LDBHE 3800+ x2 so i can run on my phase...

[stone_cold_Jimi]

Tell us how that 0601 clocks. I'm seriously considering getting an LDBHE 3800+ x2 so i can run on my phase...

Me too. The E6 I have is ok up to that coldbug moment but... maybe if we leave it another month, 2006 batches will be a certainty?

[dogsx2]

got a very nice looking LDBHE 0601 UPMW

Just running it at stock to check everythings working nicely. Currently loading at 27c with dual sp2004, on my WC. All in all, looks good.

That's the same stepping I have coming Saturday. Was that a retail also? Get that bugger cranked up so I can feel good about buying that one of mine.

[don_vercetti]

Sadly my X1800XL just died, so i'll have to wait until i can get a replacement. I'm trying to persuade ebuyer to send me a x1800xt for saturday delivery, but i dunno how succesful that will be.

Anyway hopefully it'll be good...looks nice on paper, and has VERY nice tcase stats.

[smids]

Sadly my X1800XL just died, so i'll have to wait until i can get a replacement. I'm trying to persuade ebuyer to send me a x1800xt for saturday delivery, but i dunno how succesful that will be.

Anyway hopefully it'll be good...looks nice on paper, and has VERY nice tcase stats.

Should have waited till their stock of both ran out and got a X1900XT from them. Could have run on a PCI card until then

[don_vercetti]

well the x1800xt was £240, so i'm not complaining.

And the only PCI card i'v got has trouble running above 640x480...it's suitable for blind flashing, but that's about it.

Besides, i just want to get a card for my computer, and overclock this cpu.

[Leverpospade]

Just resived my x2 3800+, and it's a LDBHE 0601 UPMW, so I'm hopeing that it likes cold.

I'm dual priming at 2700 @ 1,3 V rigth now. Haven't test it higher yet Looks like those 0601 UPMW are some wild chips

[don_vercetti]

SWEET
i can't wait to get mine going now. What's your batch number?

[K.I.T.T.]

lol, go don

I cant believe you killed THREE videocards in the past few months though

[don_vercetti]

not 3...only one was killed, other two died of unknown natural causes. Anyway that is OT

[Leverpospade]

Not sure what a batch number is, but expect it to be this: 169975AA60792

[Leverpospade]

Okay, I have been testing my CPU for a couple og hours, and it seems to be cind of strange to handle. At 2700 MHz @ 1,3 v, core 0 fails, but core 1 just keeps rolling along. When a then go up to 1,4 v core 0 still fails at 2700 MHz. But If I proced to 1,45 v both cores fails

I'll be testing some more next week...

[don_vercetti]

to hot?

What cooling are you under? Not that vapochill i presume?