Average overclocking for AMD X2 3800+?

[chinkgai]

Originally posted by: Gronich
bjc112 I ran the SuperPi 32M again at your settings of 263x10 with the your memory settings
2.5,3,3,8 2T and I got the initial value of about 33secs!

Could you give me a few more details about the settings you are running?

BIOS?
Aggressive Timing ON/Disable?
HTT?

Also if you have a chance could you do the SuperPi 16K, 32K etc. up to say 2M - This would he me rule out the IO speed differences of our HDD.

Many thanks

are u running two instances of superpi at the same time?

prime requires two instances as well simultaneously...

 

[Gronich]

Yes, I've tried it with two makes no difference - I think it's either the OS (I'm running W2K), BIOS rev, IO speed or something else.

I have compared my previous processor - Winchester 280x9=2520Mhz - using Everest/Sandra by using the same memory timings and clocking my X2 to the same speed and I appear to be losing a bit of bandwidth and have a 18% more latency. But testing using Quake III and 3dMark2001 give me near identical result slightly favouring the X2 which suggest the Everest/Sandra maybe reading wrong.

I need to make a comparison with another X2 system same speed and memory timings, etc. to help see what the issue is...

 

[chinkgai]

oh, i meant to test for stability...run two instances of 32mb and two instances of prime for at least 8 hours

 

[Duvie]

Originally posted by: aksnow7

Originally posted by: Markfw900

Originally posted by: aksnow7
I'm running stable at 2.4 Ghz but I haven't been able to go any higher. At 2.5 with a bump in voltage I was able to get Windows to load but it crashed shortly after. Normally it restarts at the loading screen.

My system:
Antec Truepower 500w
X2 3800+
Epox 9NPA+Ultra (flashed to latest bios)
2 Gigs OCZ Value

I've done the following:
Ram @ 2-3-3-7
HT @ 3
FSB: 240
Ratio: 10X
CPU V: +0.025v (Bounces around 1.45-1.49v)
Mem Freq.: DDR333(166mhz)

Any recommendations?

What HSF ? What temps ? what voltage for the core ? what memory timings and dividers ?

XP-90C w/ 92mm panaflo and AS5.

Temp sits around 32C. 45C under load.

Voltage is 1.45

Memory is currently at 2,3,3,7 2t using the DDR333(166mhz) divider.

NO real load!!!

I have a xp-90c with 92mm fan Zalmann pushing same cfm as the panaflo and I go from 29c idle to 50-51c load (2 intances of prime95) and my room temp is 72f...

That is with 1.456idle actual....

 

[aksnow7]

Originally posted by: Duvie


NO real load!!!

I have a xp-90c with 92mm fan Zalmann pushing same cfm as the panaflo and I go from 29c idle to 50-51c load (2 intances of prime95) and my room temp is 72f...

That is with 1.456idle actual....

The highest temps I've seen @ 2.4 Ghz after several hours of prime95 (2 instances, 100% CPU Usage) is 46C. I'd like to push it higher but anything over 2.4 is completely unstable. I'm fairly new when it comes to overclocking though, so I know there is something I'm not doing. Maybe memory timing or more voltage?

 

[Darkstar757]

Well i guess i should be nervous cause I have my manchester running at 2.6ghz ocz ram at 217mhz and the temp at full load are about 52C.

 

[Duvie]

Originally posted by: aksnow7

Originally posted by: Duvie


NO real load!!!

I have a xp-90c with 92mm fan Zalmann pushing same cfm as the panaflo and I go from 29c idle to 50-51c load (2 intances of prime95) and my room temp is 72f...

That is with 1.456idle actual....

The highest temps I've seen @ 2.4 Ghz after several hours of prime95 (2 instances, 100% CPU Usage) is 46C. I'd like to push it higher but anything over 2.4 is completely unstable. I'm fairly new when it comes to overclocking though, so I know there is something I'm not doing. Maybe memory timing or more voltage?

I am just amazed you can have such a small swing when I have a larger one with lower idle and obvious lower ambient temps.....

One thing ppl should check is what core fails...If you run 2 instances use affinity and set each to one of the ocres...Then watch as consistenetly one will faiul...my core 1 is the weakest by about 33mhz.....

 

[arcas]

Looks like an X2 3800+ is in my future.

2 questions:

1) Is there a straightforward way to estimate a CPU's power dissipation when you change the voltage and/or speed?

2) Has anybody tried undervolting their X2 and running at stock speed? If so, how low could you go? What were your temps?

 

[bjc112]

Originally posted by: n7
Damn, bjc112, you just seem to get lucky with all your A64s :Q

I must say I have been rather fortunate on my chips..

I did have one Venice that only did 2.4-2.5.. One that did 2.6 and one that 2.8 ( probably 2.9 but HTT was TOO high )

This X2 kicks ass.. Even though I truly didn't need dual core, it is nice to have the second core.. Things are P R E T T Y smooth!

 

[bjc112]

Originally posted by: Gronich
bjc112 I ran the SuperPi 32M again at your settings of 263x10 with the your memory settings
2.5,3,3,8 2T and I got the initial value of about 33secs!

Could you give me a few more details about the settings you are running?

BIOS?
Aggressive Timing ON/Disable?
HTT?

Also if you have a chance could you do the SuperPi 16K, 32K etc. up to say 2M - This would he me rule out the IO speed differences of our HDD.

Many thanks

I see you are running Win2k?

For some reason it kills your superPI time..

I was scoring 26min on my 2.8 and I believe some with Win2k would push you back in to the 30min range..

I run the small test for you and let you know what I get!

 

[bjc112]

Hey Gronich

Win2k is what is dragging you down..

But here are my times from 16k-2M

THESE SCORES are WITHOUT setting AFFINITY to 0 or 1.. Not sure if that has a small effect or not yet..

16k = 00.219
32k = 00.500
64k = 01.100
128K =02.516
256k = 06.485
512k = 15.391
1M = 39.797
2M = 1m 19.297

Ill check into affintiy to see if that affects my scores!

 

[Gronich]

To answer chinkgai question first I ran Prime95 to instances for 14 hours on Monday with no problems - I have had no speed related issues (i.e. Random lockups) at all at settings in my signature.

bjc112

My results at my signature speed:

Percentage Difference
16k= 0.219 0%
32k= 0.453 10.3%
64k= 1.031 6.7%
128k= 2.344 7.4%
256k= 5.969 8.6%
512k=14.031 9.7%
1M= 31.422 26.7%
2M=1min 13.063 8.5%

2800/2600Mhz= 7.7% speed difference

The speeds seem pretty much in line with what you'd expected with the speed difference and my memory timing with the exception of the 1M run.

Could you just re-confirm the memory timings and CPU speed your signature says 2.6Ghz but the 32M piccie says 2630Mhz and I will re-run for a direct comparison.

Many thanks

 

[aksnow7]

Originally posted by: Duvie

Originally posted by: aksnow7

Originally posted by: Duvie


NO real load!!!

I have a xp-90c with 92mm fan Zalmann pushing same cfm as the panaflo and I go from 29c idle to 50-51c load (2 intances of prime95) and my room temp is 72f...

That is with 1.456idle actual....

The highest temps I've seen @ 2.4 Ghz after several hours of prime95 (2 instances, 100% CPU Usage) is 46C. I'd like to push it higher but anything over 2.4 is completely unstable. I'm fairly new when it comes to overclocking though, so I know there is something I'm not doing. Maybe memory timing or more voltage?

I am just amazed you can have such a small swing when I have a larger one with lower idle and obvious lower ambient temps.....

One thing ppl should check is what core fails...If you run 2 instances use affinity and set each to one of the ocres...Then watch as consistenetly one will faiul...my core 1 is the weakest by about 33mhz.....

Yep same here, the second instance of prime95(affinity to core 1) fails first. I was actually able to boot into Windows @ 2.5 Ghz with a small vcore increase but prime95 failed after a couple of minutes. So I'm happy with a stable 2.4 ghz @ 46c load and will stick with this I suppose.

 

[bjc112]

Originally posted by: Gronich
To answer chinkgai question first I ran Prime95 to instances for 14 hours on Monday with no problems - I have had no speed related issues (i.e. Random lockups) at all at settings in my signature.

bjc112

My results at my signature speed:

Percentage Difference
16k= 0.219 0%
32k= 0.453 10.3%
64k= 1.031 6.7%
128k= 2.344 7.4%
256k= 5.969 8.6%
512k=14.031 9.7%
1M= 31.422 26.7%
2M=1min 13.063 8.5%

2800/2600Mhz= 7.7% speed difference

The speeds seem pretty much in line with what you'd expected with the speed difference and my memory timing with the exception of the 1M run.

Could you just re-confirm the memory timings and CPU speed your signature says 2.6Ghz but the 32M piccie says 2630Mhz and I will re-run for a direct comparison.

Many thanks

Those scores I just reported were @ 2600 mhz.. Memory @ 240mhz I believe, I will double check when I get home from work.

 

[amd4life]

wow is that a toledo or manchester core?

and for starters

(havent had to deal with dividers and HT for 1.5 yrs LOL)

I just want a stable 1:1 2.4 ghz O'clock

what is the best starting point??

HT 4, 240FSB? will that do it? or do I need to mess with the dividers also??

 

[silentvois]

Yeah, set Ht to 4 so wont reach over 1000. If you got ram that can run 240FSB then it's cool, you won't need to run any dividers. If you wanna stick to 200mhz for your memory then you can run a 166 divider on a 240fsb. You should be able to hit 2.4ghz easily. For mine I needed 1.4v vcore to hit 2.4ghz but it might be different for you.

 

[imported_Funkatation]

Got an X2 3800+ from ZZF last week. Have it running at 2.4ghz (prime95 stable) at default voltage! Haven't really tried to do more (stock HSF).

 

[LED]

2.4GHz here @ default volts {fluctiates from 1.3V-1.35v's}, Stock HS Fan...and ...I'm in 1 hellish storm with everything going including the UPS and have this thing loaded..running Tunes, Browser, Bunch of Progs, and still got some results with the RAM @ 2.5 on the 1st Gen 3800Duce and it'll take 1 Major Beating alright...here's what I got tinkering with it {CPU fluxing between 1.30-1.35}~ maybe due to UPS

Pics

 

[formulav8]

bjc112, your 1MB scores seem low for 2.6ghz. My 2.6ghz speeds with tweaked memory is 34.360. My 2.6ghz is more than: 13% faster than your 2.6ghz. Mine is not set to affinity either. Something doesn't seem quite right about yours if you are running at 2.6ghz. Anyways, just thought I would mention it.

Your 2MB score is faster than mine by about 5% though. Wierd. My memory is only running at 173mhz right now though @ 2225 timings.



Jason

 

[Superguy]

I hit 2560 on mine. Using a Thermalright SI-120.

I might be able to go a little higher if I could ever get a stable vCore. E6 stepping.

It'll post at 2.7GHz, but it's not stable.

 

[byoungvt]

Hi all,

I am selling a Toledo 3800+ on ebay brand new sealed in the box. If you email me after purchasing I will refund the shipping charge if you mention you saw this post. Here is a link:

Toledo X2 3800

 

[govtcheez75]

Originally posted by: Gronich
Thermalright XP90 with 90mm 74CFM 30dBa fan

:Q

...I don't know what impresses me more: The fact that he got 2.9ghz, or the fact that he has a 92mm fan that blows 74CFM while only creating 30dBA of noise. I want this magical fan!

 

[nombrecinq]

Is everyone who bought an X2 using a special 64-bit Windows or just the regular XP? Also is TankGuys the best and cheapest place to buy the Toledo core?

 

[Gronich]

Originally posted by: govtcheez75

Originally posted by: Gronich
Thermalright XP90 with 90mm 74CFM 30dBa fan

:Q

...I don't know what impresses me more: The fact that he got 2.9ghz, or the fact that he has a 92mm fan that blows 74CFM while only creating 30dBA of noise. I want this magical fan!

If you would like me to pop off the fan and tell you what it is let me know?

 

[govtcheez75]

Originally posted by: Gronich

Originally posted by: govtcheez75

Originally posted by: Gronich
Thermalright XP90 with 90mm 74CFM 30dBa fan

:Q

...I don't know what impresses me more: The fact that he got 2.9ghz, or the fact that he has a 92mm fan that blows 74CFM while only creating 30dBA of noise. I want this magical fan!

If you would like me to pop off the fan and tell you what it is let me know?

Yes. Please let me know the model of that fan. I want it!