Cpu Voltage

[t33lo]

What are some symptoms of the chip not having enough voltage? I have the 3700+ san diego core and it runs at 1.36V. I plan on Overclocking it from the base 2.2ghz anywhere up to 2.8. When would it be good to increase the voltage to say 1.5V?

<i>I pretend to know what I'm talking about.</i>

 

[Coyote]

What are some symptoms of the chip not having enough voltage?

Instability when running Prime95 or other stress inducing program.

Don't know if anybody can accurrately say when your CPU will need to go up to 1.5v. All are different YMMV.

When OCing I only bump the vcore, and by smallest amount possible, when it fails P95. I run memtest86 first to ensure ram is not the problem.

Mobile XP 2600+ (11X215)
Abit NF7-S v 2.0
Maxtor 60GB ATA 133 7200RPM
1 gig Corsair XMS PC3200
eVGA 6800GT
Enermax Noisetaker 420 watts
Win2K sp4

 

[t33lo]

Well I set the DIMM voltage to 2.8v for my 2x512 crucial ballistix pc3200. Whenever I increase the fSB past 215mhz, the comp will not load into windows but rather restart after getting past mobo splash screen. What do I do so that I can increase the FSB and not have this problem? What exactly is the problem?

My ram is set at 2-3-2-8 and is in dual channel.

<i>I pretend to know what I'm talking about.</i>
<P ID="edit"><FONT SIZE=-1><EM>Edited by t33lo on 08/23/05 11:33 AM.</EM></FONT></P>

 

[t33lo]

I now have it at 235mhz @ 2.5-2-2-5 1T running at 2.8 v. When should or would I want to use a divider?

<i>I pretend to know what I'm talking about.</i>

 

[fishmahn]

It's a balancing act - you raise the FSB until it's unstable, then you get to decide if it's the CPU, RAM, chipset, or what that's reached its limit, and you compensate for that limit.

I would put your ram on a divider and turn up the FSB to see how fast the CPU can go, then back everything down to stock, lower the CPU multiplier and turn it up again to see how fast the RAM can go. Once you have that, you can decide what the best combination is.

Use a divider if your ram can't go faster, but your CPU can, and you want to ramp up the CPU further.

Mike.

<font color=blue>Outside of a dog, a book is man's best friend. Inside the dog its too dark to read.
-- Groucho Marx</font color=blue>

 

[Ampz]

I have the same chip exactally, I'm running 261fsb and lowered the mutiplier to 10, and my ram is running at about 216mhz its ddr400 my voltage is around 1.535 or so atm (untill my water cooling parts come in its running fine atm. Don't hit 1.6v untill you get water if you dont plan to then don't go that high it's not a good idea.

 

[t33lo]

I'm going to reformat and start fresh now. I'm just curious, is it generally better to have a lowfsb but a higher multiplier or a high fsb and lower multiplier. When I mean lower i mean basically 10 not 11 (11 is my chips max).

<i>I pretend to know what I'm talking about.</i>

 

[fishmahn]

What will the new school say?

Mike.

<font color=blue>Outside of a dog, a book is man's best friend. Inside the dog its too dark to read.
-- Groucho Marx</font color=blue>

 

[t33lo]

Well I had my computer set at 240mhz x 11 @ 2.6ghz. My ram was set at 2.5-2-2-5 and at 2.8v (specified voltage). I did not increase my vcore either. When playing games, It would just restart after say, 20 minutes. The gfx card wasn't getting too hot, the chip wasn't, and the nb was at about 54*C. I went back into cmos and adjusted the vcore to 1.40 from 1.375. Windows booted up in under one green load bar. I'm curious as to what would cause a shut down like that w/o temps getting too hot. Here's the error report:

BCCode : 1000007f BCP1 : 0000000D BCP2 : 00000000 BCP3 : 00000000
BCP4 : 00000000 OSVer : 5_1_2600 SP : 2_0 Product : 768_1
Technical information:

C:\DOCUME~1\T-Lo\LOCALS~1\Temp\WER2f92.dir00\Mini082405-01.dmp
C:\DOCUME~1\T-Lo\LOCALS~1\Temp\WER2f92.dir00\sysdata.xml

At the new settings just mentioned above, I got a not exact in round error when running superpi to 8M iterations. It seems to be a memory problem. Whats a good way to troubleshoot and tweak the memory to its optimal settings?

<i>I pretend to know what I'm talking about.</i>
<P ID="edit"><FONT SIZE=-1><EM>Edited by t33lo on 08/24/05 07:52 AM.</EM></FONT></P>

 

[t33lo]

Well i rean memtest which is integrated into my cmos and didn't pass one test. I thought this was kind of odd so I set my cmos settings to the default values and still didn't pass one test. I also changed just the dimm voltage to 2.8v from 2.6 and still didn't pass one test. Any reasons why memtest isn't passing me?

<i>I pretend to know what I'm talking about.</i>

 

[t33lo]

Currently I have it running at 240mhz x 11 @ 2.64GHZ. I have the voltage set at 1.525 but cpuz says its at 1.504. I ran superpi at 1 million iterations and got 32s. I then ran it at 8million and got 5min 8s. I didn't get any errors in the proccess which is good news becuase whenever I have my settings off I get non convergent square errors and exact round errors as well. I guess its just a matter of temperature while gaming that will keep this thing going good? Are there any other programs besides memtest (its built into my cmos and fails every test without a single pass even when comp is set to normal settings) that will stress test it like super pi does? How about any other good RAM programs for stresstesting?

<i>I pretend to know what I'm talking about.</i>

 

[t33lo]

Why do you suppose the memtest built into my bios fails every test without passing anything? Is it possible my bios is out of date? I have DFI Lanparty SLI-DR

<i>I pretend to know what I'm talking about.</i>

 

[t33lo]

I ran 3dmark05 and it crashed during the last test. It crashed during the CPU test of the space station battle. I set the vcore to 1.535 i think but cpuz says its 1.504. Which one is more accurate? I think i might want to up the vcore. Is higher than 1.6 dangerous even though the chip may be somewhat cool?

<i>I pretend to know what I'm talking about.</i>

 

[t33lo]

Do you think my memory timings are a bit tight at 2.5-2-2-5?

<i>I pretend to know what I'm talking about.</i>

 

[t33lo]

Crucial Ballistix PC3200 2x512mb @ 2.8V. It said it runs at 2-2-2-6 @ 2.8v at their website. I hear the higher you increase the fsb, the lower you must bring the latencies. Is this true? And if so, does that mean a lower fsb but fast latencies provide better for gaming performancE?

<i>I pretend to know what I'm talking about.</i>

 

[t33lo]

I plan on having the ram for a little bit longer than 3.2v will allow What is the difference between 1T and 2T?

<i>I pretend to know what I'm talking about.</i>
<P ID="edit"><FONT SIZE=-1><EM>Edited by t33lo on 08/24/05 06:37 PM.</EM></FONT></P>

 

[t33lo]

To go to 3.2v i need to adjust a jumper on my mobo i think. You said the chip shouldn't get above 50*C. If under load it does to say 55, is that such a bad thing? I know that prescotts can get higher under load. Are we just comparing apples to oranges or can the temps stay above 50*C?


<i>I pretend to know what I'm talking about.</i>

 

[t33lo]

For some strange reason, I am running at 2.75ghz (250.1x11) with timings of 3-3-3-6 (2.8V) with a vcore of 1.456v. I have not experienced any instablillity whatsoever. The temps stay below 48*C under load. How can this run w/o increasing the vcore to at least 1.55v?

<i>I pretend to know what I'm talking about.</i>

 

[t33lo]

I see. Is it neccesary to have a computer stable enough to handle prime 95? Will I run accross anything demanding such as prime95 where I would want that stability?

<i>I pretend to know what I'm talking about.</i>

 

[t33lo]

Yeah it runs through 3dmark05 and pcmark05 without a hitch. I get 8900 in 3dmark05 but pcmark05 is low at 4800. Strange.

<i>I pretend to know what I'm talking about.</i>

 

[ChipDeath]

At 200Mhz you can run them at 2-2-2-5 1T 2.8V.
While at 250Mhz you'll have to use 3-3-3-6 2T 2.8V

I seem to recall someone (Crashman?) saying the Ballistix PC3200 and PC4000 are exactly the same, just at 200Mhz the SPD chip has tighter timings.

If that's the case, then he should be able to use CL2.5 & 1T at 250Mhz as I can with my Ballistix PC4000. I'm currently at 265Mhz 2.5-6-6-10-1T or so (relaxed due to pushing the OC a bit )

---
<font color=red>"Life is <i>not</i> like a box of chocolates. It's more like a jar of jalapeńos - what you do today might burn your a<b></b>ss tommorrow."

 

[t33lo]

Unfortunatley I can't run it at 265 at those timings. I even lowered the multiplier and even adjusted the vcore and dimm voltage. I can't wait to overclock more but need a mounting piece for my zalman cnps 7000 cuz I lost it

<i>I pretend to know what I'm talking about.</i>

 

[ChipDeath]

Must be downbinned as Wusy says then. Ah well.

All the same, getting PC3200 to run at 265Mhz ain't half bad, even if the timings aren't amazing.

It's just a case of playing with it. You'll not find the 'sweet spot' for some weeks probably. That's the way it always is with me.

Just remember that there <i>is</i> a limit which you simply won't be able to get past with your hardware. I don't know where it is, but it is there somewhere. It's an easy thing to forget for an overclocking obsessive

---
<font color=red>"Life is <i>not</i> like a box of chocolates. It's more like a jar of jalapeńos - what you do today might burn your a<b></b>ss tommorrow."

 

[t33lo]

After overclocking, my nf4 chipset under heavy load gets to 55*C. How is this for a temp? I hear this chipset runs a bit hot anyways.

<i>I pretend to know what I'm talking about.</i>

 

[ChipDeath]

They do get fairly hot. for Overclocking you may want to improve the chipset cooling a little - either by adding a fan to the chipset hs(assuming yours doesn't have one) or by buying/making a better hs.

I glued an old SocketA Aluminium cooler to the chipset on my old Nforce2 board. Actually gave me an extra ~10Mhz FSB too.

Generally though, it's better to not epoxy stuff on. It's a bit permanent and can cause issues later.

---
<font color=red>"Life is <i>not</i> like a box of chocolates. It's more like a jar of jalapeńos - what you do today might burn your a<b></b>ss tommorrow."