Read First Before Increasing Voltages!

No_Style

In the midst of high overclocking P4s, I've noticed a lot of individuals stepping over the max tolerables for their CPUs.

[COLOR=Orange]Intel [/COLOR]

Intel Celeron(R) Processor up to 1.10 GHz
Section 2.9 - Table 4 states:
Max Voltage: [Typical Operating voltage from Table 5] V + 0.1 V = [MaxVCore] V

Intel Celeron Processor for the PGA370 Socket up to 1.40 GHz on 0.13 Micron Process
Section 2.10 - Table 6 states:
Max Voltage: 1.75 V

Intel(R) Celeron Processor in the 478-Pin Package at 1.80 GHz
Section 2.9 - Table 5 states:
Max Voltage: 2.10 V

Intel Celeron Processor on 0.13 Micron Process in the 478-Pin Package
Section 2.10 - Table 6 states:
Max Voltage: 1.75 V

Intel Celeron D Processors 335, 330, 325, and 320
Section 2.10 - Table 2-7 states:
Max Voltage: 1.55 V

Coppermine Documentation (500 MHz to 1.13 GHz)
Section 2.9 - Table 6 states:
Max Voltage: 2.10 V

Intel Pentium 4 Processor in the 423-pin Package at 1.30, 1.40, 1.50, 1.60, 1.70, 1.80, 1.90 and 2 GHz (Willamette Socket 423)
Section 2.9 - Table 4 states:
Max Voltage: 2.10 V

Intel Pentium 4 Processor in the 478-pin Package at 1.40 GHz, 1.50 GHz, 1.60 GHz, 1.70 GHz, 1.80 GHz, 1.90 GHz, and 2 GHz (Willamette Socket 478)
Section 2.10 - Table 5 states:
Max Voltage: 2.10 V

Intel Pentium 4 Processor with 512-KB L2 Cache on 0.13 Micron Process and Intel Pentium 4 Processor Extreme Edition Supporting Hyper-Threading Technology Documentation (Northwood)
Section 2.10 - Table 2-5 states:
Max Voltage: 1.75 V

Intel Pentium 4 Processor on 90 nm Process Documentation (Prescott)
Section 2.9 - Table 8 states:
Max Voltage: 1.55 V

Mobile Intel Pentium 4 Documentation
Section 2.10 - Table 6 states:
Max Voltage: 1.75 V

Mobile Intel Pentium M Documentation (Centrino)
Section 3.9 - Table 4 states:
Max Voltage: 1.75 V

Intel Pentium M Processor on 90nm Process with 2-MB L2 Cache Documentation
Section 3.8 - Table 4 states:
Max Voltage: 1.55 V

Intel Pentium 4 Processor 560, 550, 540, 530 and 520
Section 2.10 - Table 2-7 states:
Max Voltage: 1.55 V

Intel Pentium 4 Processor Extreme Edition on 0.13 Micron Process in the 775-Land Package
Section 2.10 - Table 2-4 states:
Max Voltage: 1.55 V

[COLOR=Green] AMD [/COLOR]

AMD Duron 600 MHz to 950 MHz
Section 7.9 - Table 8 States:
Max Voltage: 1.7 V

AMD Duron 900 MHz to 1300 MHz
Section 7.8 - Table 7 States:
Max Voltage: 1.75 V + 0.5 V = 2.25 V

AMD Duron 1400 MHz to 1800 Mhz
Section 7.8 - Table 7 States:
Max Voltage: 1.5 V + 0.5 V = 2.0 V

AMD Athlon XP 1500+ to 2100+
Section 6 - Table 1 States:
Max Voltage: 1.75 V + 0.5 V = 2.25 V

AMD Athlon XP With 266 FSB (1700+ to 2200+)
Section 8.8 - Table 13 States
Max Voltage: 1.65 V + 0.5 V = 2.15 V

AMD Athlon XP With 333 FSB (2500+ to 3000+)
Section 8.8 - Table 14 States:
Max Voltage: 1.65 V + 0.5 V = 2.15 V

AMD Athlon XP With 400 FSB (3000+ to 3200+)
Section 8.8 - Table 14 States:
Max Voltage: 1.65 V + 0.5 V = 2.15 V

I hope this will save a lot of angst, money loss and wasted time to all overclockers out there. It does not matter if you have low core temperatures, the fact that the processor is receiving a higher voltage than it is capable of handling, deteriorates it. Remember, just because the voltage is set at 1.65 V, it does not necessarily mean the actual voltage rating is 1.65 V (Asus boards are notorious for not obeying their set VCore). I suggest everybody who's overclocking give the documentaion a quick read.

Keep that voltage low

**Note: There is certainly more than just mere voltages that will regulate the health of your processor. AMD processors have high max voltages, but within their documentation there is also thermal power to consider.

UPDATE: I've updated the real absolute max. Thanks to the inquiry from Jazzyjazz2004. I've must have misread somewhere. The update should make more sense now. Sorry for confusion.

UPDATE 2 (9/3/04): Added Mobile Pentium 4 and Pentium M max voltage info.

UPDATE 3 (6/4/04): Added Prescott Max Voltage Info

UPDATE 4 (3/9/04): Added Duron Max Voltage Info and Celeron, LGA775 max voltage info. Fixed some minor typoes.

UPDATE 5 (30/10/04): Fixed addition errors.

__________________

HardWired

Couldn't agree more about keeping the volts low...that's some good info for others.

I've read from many overclockers not to even go over 1.70 for any great length of time w/ Northwood chips, regardless of Intel's stated 1.75 max. I use a P4PE also, and I'm able to get 3.3GHz out of my 2.53b multivid at 1.525, even though the board overvolts it to 1.632.

I could get 3.4GHz or higher stable(ran it that way for a short while...even 3.5GHz) at 1.65+, but I don't want to flirt with that 1.70 threshold. I've seen it referred to as "Sudden Northwood Death Syndrome" or SNDS. It a slow but sure failure over a couple of days or weeks and has been attributed to running at or over 1.70 vcore.

I can only afford an upgrade every 6mo. to a year, so I'm keeping this one at a low vcore and am very happy with my 800 MHz o'c.

No_Style

This information is not apparent. The "manual" for the Intel Chips that come in boxes or whatnot, do not show you any of this stuff. This info seems to be purely online.

__________________

No_Style

God Damn. This is sticky worthy isn't?

__________________

No_Style

I'm kissing the 1.7 V threshold right now. It fluxes to that mark every so often. Living life on the edge.

__________________

Necrosis

Indeed it is true, and I'm not sure about your 50% tolerance though. Usually for components it's around 10%. What this is called is dielectric breakdown. Basically every diode, transistor has a certain voltage where it starts to break down. It is sometimes used to limit flow, so intentionally done on purpose in the normal operation of a circuit. I've noticed with my board there is indeed a .5 variant in Vcore voltage. I had it set at 1.7 (1.74 - 1.75 in MBM) and noticed my full load temps were awfully high. Now it's set at 1.65, and hits around 1.71 in MBM with a full load temp for 2 hours of 43C. I'm taking these temps. from the diode not underneath the socket.

JCLW

1.75 is the accepted max voltage for Tullies, but I've yet to see a Tualatin core die from anything except a pin mod gone bad. Mine barely hits 40C @ 1.825v, which is the max vcore a VRM 8.5 spec board will supply.

- JW

Warpy

Hi

From looking at the spec docos my cpu isnt capable of overclocking!!

I heard some bad news about the 2200+ - Also noticed that not many ppl have them.

I wonder why they released the chip in so far as being so different to the others?

B

__________________

NucleuZ

I gotta get myself one of those P4 2,4C or 2,6C...they seem to overclock nicely.

I read that once you reach top overclocking levels, you should go down with the voltagelevels. Also higher voltage doesn't always mean higher overclocking levels.

tastyweat

Some decent info there mate - luckily - I don't have to adjust my volts atm

2800barton @ 200x11 on 1.65v

__________________

Sharkims

On the newer P4 boxes it's rather apparent. 2.4c - "MAX VOLTAGES: 1.525v" 3.0c "MAX VOLTAGES: 1.55v"

My personal experience with this 2.4c is I am more stable at 1.525v than I am at 1.55v+. So, this is very, VERY true.

__________________
System-1:
MSI K8N Neo Platinum || A64 3200+(200x11) - Zalman 7000A-AlCu || 1x1gb OCZ PC3200@400(3-3-3-7) || 9800XT || WD Raptors 2x36 RAID0
System-2:
MSI 875P Neo-FIS2R || P4 2.4c(FSB:200) - Intel Cooler || 2x256 GEIL PC3200@400(2-3-3-6) || AIW9000 || WD SE 120x2 RAID0
System-3:
DELL Inspiron 9100 || P4 3.0c(mobile) || 2x256mb PC3200(3-3-3-8) || Mobility Radeon 9700 - 128mb || WSXGA+ 15.4" || 60gb 5400RPM 8mb

Jazzyjazz2004

"AMD Athlon XP With 333 FSB (2500+ to 3000+)
Section 6.4 - Table 4 states:
Max Voltage: 1.5 x (50%) 1.65 V = 2.475 V"

I downloaded the PDF file and cannot decipher where you derived this formula yielding the max voltage is 2.475V. Yes, I went to the correct section and table.
Are you saying that it is SAFE to run my Athlon XP 2500 at 2.475 V?

Jazz

__________________
nVidia 5900 XT @ 351\303,Abit NF7-2 Rev. 2, Athlon XP 2500 MOBILE @ 210*12@1.825V = 2520 Mhz, SLK-947-U HS, 512 MB Corsair 3000 ddr @ 210 Mhz, 200 GB Western Digital SATA, Lite On Cd-Rw/DvD Combo Drive, Antec TruPower 480W PS

No_Style

Sorry about the confusion, I looked at the docs again and I realized I was looking at the wrong place. Sorry for confusion. It's updated now. Thanks Jazz~!

__________________

Jazzyjazz2004

Here I come 1.95 (safety margin). I currently run Athlon XP 2500 at 12.5x185@1.75...smokin'!

__________________
nVidia 5900 XT @ 351\303,Abit NF7-2 Rev. 2, Athlon XP 2500 MOBILE @ 210*12@1.825V = 2520 Mhz, SLK-947-U HS, 512 MB Corsair 3000 ddr @ 210 Mhz, 200 GB Western Digital SATA, Lite On Cd-Rw/DvD Combo Drive, Antec TruPower 480W PS

Zardon

running my barton @ 1.93 v 2500mhz, probably shorten its lifespan but the performance is smokin

Jazzyjazz2004

LOL... I can imagine that it's smokin'!

I stopped at 185x12.5 not only because my memory is maxed at at 2-2-2-2-6 (is that too many 2's?) but because any subsequent raising of the overclock using just the multiplier required such a large voltage increase that I didn't feel the additional 100 to 200 Mhz was worth it.

For instance;
I got 479 Mhz increase (from default 1833 Mhz to 2312 Mhz) for a 0.10 voltage increase (1.65V to 1.75V)...but to go ANOTHER 100 Mhz needs another +0.1 V, for a total of 1.85V... an additional 200 Mhz requires a grand total of 1.95V. The extra 100 to 200 Mhz didn't seem worht it, especially since my 3d Benchmarks stayed the same. I suspect my graphics card is my current bottleneck and I need an upgrade.

Jazz...and PEACE to You!

__________________
nVidia 5900 XT @ 351\303,Abit NF7-2 Rev. 2, Athlon XP 2500 MOBILE @ 210*12@1.825V = 2520 Mhz, SLK-947-U HS, 512 MB Corsair 3000 ddr @ 210 Mhz, 200 GB Western Digital SATA, Lite On Cd-Rw/DvD Combo Drive, Antec TruPower 480W PS

G@MeBoY

running my XP2000 at 10x230@2.0 for 4 months, not smokin' ...not diein' to jazzyjazz : cant you lower multiplier and go for FSB speed ? FSB will give you more increase than 100mhz on CPU. Better to have 230FSB with loosen memory latency settings and 2ghz CPU, than 2.3Ghz CPU with 190FSB...

__________________
-= G@MeBoY =-
::Full color with 3DNOW! tech::
XP2000+ thoroughbred @ 10x230FSB - watter cooled by home made sistem - EpoX RDA3I - ASUS V9950SE/Td @470/820 - 2x256 DDR400 V-Data - !Live Player - 80GB Maxtor DiamondMax9+7200rpm - Tagan TG-380 black PSU - chieftek Matrix case -
WinXP SP1

No_Style

Anybody want to help me with finding the documentation for the Athlon XP-Ms or determining Athlon 64 (Opteron, FX, Regular) max voltages? I can't seem to determine what the max is for 64 bit chips, and finding XP-M documentation was taking me in circles over at AMD's homepage.

__________________

Jazzyjazz2004

Yes, I have tried upping the FSB and lowering the multi, but system becomes unstable...I need faster clocking memory. Jazz

__________________
nVidia 5900 XT @ 351\303,Abit NF7-2 Rev. 2, Athlon XP 2500 MOBILE @ 210*12@1.825V = 2520 Mhz, SLK-947-U HS, 512 MB Corsair 3000 ddr @ 210 Mhz, 200 GB Western Digital SATA, Lite On Cd-Rw/DvD Combo Drive, Antec TruPower 480W PS

G@MeBoY

To jazz: set vdd voltage to 2.0V and you'll do fine, if not set vdimm voltage to 2.9V...Corsair MUST do it

__________________
-= G@MeBoY =-
::Full color with 3DNOW! tech::
XP2000+ thoroughbred @ 10x230FSB - watter cooled by home made sistem - EpoX RDA3I - ASUS V9950SE/Td @470/820 - 2x256 DDR400 V-Data - !Live Player - 80GB Maxtor DiamondMax9+7200rpm - Tagan TG-380 black PSU - chieftek Matrix case -
WinXP SP1

Roadee

Not only will the higher voltages decrease the life span of the CPU, but it will also shorten the lifespan of the MB. High FSB and voltages will add undo heat to many of the circuits of the MB to include the Mosfets. Many of the NF2 boards are well known for the Mosfets overheating, and this in turn causes unstable voltages, in which can cause instability and occasionally damage to computer components.

I would NEVER suggest to anyone that they run at such a high voltage unless they are using extreme cooling solutions on ALL components involved. Which would include high end water cooling, passive cooling on the south bridge, increased cooling on the north bridge, passive cooling on the mosfets, as well as the other components involved to include increased case cooling.

Overclocking is fun for the experienced, EXPENSIVE for the non-experienced.

__________________


We are born naked, wet, and hungry. Then things get worse.

No_Style

Excellent line. Hence, the reason I posted the info here.

__________________

G@MeBoY

This is OC and modding section of forum, is it ? Jazzyjazz is not unexperienced user, so I suggested higher voltages to him, thus, only for testing purposes. Nothing's gonna get friend in 15 mins, and the life span of CPU will get reduced from 100 years to 10 years (guessing) . In 10 years you an I and everyone will change at least 1 CPU. But, what Roadee said is true and bare that in mind when OC-ing, ALLWAYS CHECK temps in regular bases, if you notice that something is too hot, either cool it more or lower the voltage/clock And of course, there is allways an option to higher the latency's of memory instead of raising the Vdimm voltage, on nForce2 boards not much is lost with higher latency. I would personally rather use 230FSB with 3-3-3-11 than 200FSB with 2-2-2-11 (talking about nForce2 chipset).

__________________
-= G@MeBoY =-
::Full color with 3DNOW! tech::
XP2000+ thoroughbred @ 10x230FSB - watter cooled by home made sistem - EpoX RDA3I - ASUS V9950SE/Td @470/820 - 2x256 DDR400 V-Data - !Live Player - 80GB Maxtor DiamondMax9+7200rpm - Tagan TG-380 black PSU - chieftek Matrix case -
WinXP SP1

Zardon

This is far from being a "new" article and im sure alot of you have read it already (some of it is outdated I know but its interesting to look back), but for those who havent this makes interesting reading here and here as well

Jazzyjazz2004

Sorry for this double post, but when I hit the EDIT button, the site goes to the DELETE post window...

Finally, here is the spec for my ram (taken from the CORSAIR site)...

"FOR IMMEDIATE RELEASE
Corsair Announces Availability of 512Megabyte
XMS3000 and XMS3200 DDR DIMMs

512 Max Speed
Corsair 3000
Max Rated 187.5
cas latency-2
ras to cas delay-3
ras precharge-3
Bank Cycle time-6"

Please note the "Max Rated" speed for my ram is 187.5 and that I am running at 186...hmmm, not much to go,huh?

My timings are a little better yielding 2.0\2\2\5 vs2.0\3\3\6 from above list.

Jazz

__________________
nVidia 5900 XT @ 351\303,Abit NF7-2 Rev. 2, Athlon XP 2500 MOBILE @ 210*12@1.825V = 2520 Mhz, SLK-947-U HS, 512 MB Corsair 3000 ddr @ 210 Mhz, 200 GB Western Digital SATA, Lite On Cd-Rw/DvD Combo Drive, Antec TruPower 480W PS

Zardon

you arent letting it redraw, edit will show a box below the delete option. its working fine.

Zardon

ram generally has some headroom, and you will probably get more from it, if you drop to 2.5CAS or lower to 8/4/4/2 8/4/4/2.5 settings like that. always bench with sandra or CTlaw something like that to see the effects of your changes. takes alot of time, worth it. you can also up voltages, just need to watch you dont fry it if you go too far, especially if you dont have heatspreaders.

G@MeBoY

As far as I know, no one fryed memory on 2.77V, also never single sided modules are heat friendly, but using heat spreaders is always recommended. My ram is specified to 200MHz 3-3-3-8 yet it runs fine at 230 with 3-3-3-11 (11 because of NF2 chipset) and at 224 it runs 2.5-3-3-11. All this @ 2.9V :O but I have heatspreaders, what amases me is how cool it is, I mean the modules are cold, they are at same temperature 22.9V and at 2.55V... [?]

__________________
-= G@MeBoY =-
::Full color with 3DNOW! tech::
XP2000+ thoroughbred @ 10x230FSB - watter cooled by home made sistem - EpoX RDA3I - ASUS V9950SE/Td @470/820 - 2x256 DDR400 V-Data - !Live Player - 80GB Maxtor DiamondMax9+7200rpm - Tagan TG-380 black PSU - chieftek Matrix case -
WinXP SP1

Jazzyjazz2004

G@MeBoY;

Well, you just had to ask, could I overclock the fsb some more??? I could tell you weren't going to leave it alone...but after all "This is OC and modding section of forum, is it ?"

It's been 3 hours and a couple of cups of coffee later, but the answer IS...Yes!

I can run at the following speeds and achieve the following Sandra 2003 Pro Memory Bandwidth (SMBM) and CPU Multimedia (CPUM) scores;


1) *185x12.5@1.775v = 2312 Mhz SMBM:2603\2424 (MB/s) CPUM:12741\13512
(* My original starting value)

2) 200x11.0 @ 1.775v = 2200 Mhz Missed this 'un.

3) 210x11.0 @ 1.775v = 2310 Mhz SMBM:2925\2686 (MB/s) CPUM:12678\13464

4) 220x11.0 @ 1.900v = 2420 Mhz SMBM:2982\2728 (MB/s) CPUM:13287\14106


Temps did NOT change from 42 C at any time or voltage.

My 3dMark2003 benchmark did not change much, 4920 to 5063. This again tells me that my graphics card is is the bottleneck.

Please note the huge voltage increase (0.125v) for the additional 110 Mhz going from 210 to 220 fsb. That's a LOT of volts for an approximate 0.02% increase in SMBM, huh?
I got the biggest jump going from MY default of 185x12.5 to 210x11.0...approximately 12% increase in SMBM.
I think I will try to settle in at about 215x11.0 since the scores won't be that different and I can also "lose" some of that 1.9v...shucks, the scores are not that much different at 210x11.0 either!
Keep on PUSHING folks with your challenges and encouragement.

Thank You,
Jazz

Zardon;

Your memory timings helped me beat the "overclock bug". For whatever reason, I could not overclock with ANY other multi except 11.0 above my original 185 fsb. I used your timings and "Ooops, there it is." Don't ask me "WHY" I didn't relax the timings to the extent that you suggested...BRAIN FREEZE, DUH!
Your reminder was what I needed!
I have even been able to back the timings down to 7-3-3-2 vs. 8-4-4-2.5...the improvement was insignificant in 3dMark 2003 Pro...5035 to 5063.

Thanks again,
Jazz

Now, if I can just get a NEW graphics card that can take advantage of ALL this extra memory bandwidth you folks found for me... Uhmm, let's see, I COULD donate blood every day this month...
J

__________________
nVidia 5900 XT @ 351\303,Abit NF7-2 Rev. 2, Athlon XP 2500 MOBILE @ 210*12@1.825V = 2520 Mhz, SLK-947-U HS, 512 MB Corsair 3000 ddr @ 210 Mhz, 200 GB Western Digital SATA, Lite On Cd-Rw/DvD Combo Drive, Antec TruPower 480W PS

G@MeBoY

Hmmm. at 220FSB (CPU clock doesent matter) you should be getting around 3100 in sandra. Here is mine @ 220:
memory bandwith bench. 3377/3138
latencys are 11-3-3-2.5.

Do you have 2x256 modules of corsair or only one ?! (if you have 2, check if they are in Dual channel mode aligment)
and please, in memory bench. sandra gives "bandwith efficency" info in window bellow, please post what is the value that says. As I said you should be getting around 3100 in mem bench with that RAM and abit nf7 (which is sligtly better than my board)

try lowering the latencys one by one, see which ones work and which ones dont.

__________________
-= G@MeBoY =-
::Full color with 3DNOW! tech::
XP2000+ thoroughbred @ 10x230FSB - watter cooled by home made sistem - EpoX RDA3I - ASUS V9950SE/Td @470/820 - 2x256 DDR400 V-Data - !Live Player - 80GB Maxtor DiamondMax9+7200rpm - Tagan TG-380 black PSU - chieftek Matrix case -
WinXP SP1