Progress Update: P4 1.6A @ 2.56GHz

[w9design]

Here's my latest overclocking results, in case anyone's interested:

WAS: Intel P4 1.6A @ 2.608GHz (@ 1.7V -- is this safe?)
EDIT: Not quite. Got a Prime95 error at 2608. See my reply later on for details.
NOW: Intel P4 1.6A @ 2.56GHz (@ 1.66V)
Retail heatsink/fan, using Arctic Silver II on CPU and chipset
ASUS P4B266 @ 160MHz
256MB Samsung PC2100 DDR @ 160MHz CAS2
Right now, the HD is a lethargic old 10GB/5400RPM Western Digital while I wait for my IBM Deskstar to come back from RMA.
ATI All-in-Wonder Radeon (not overclocked)
Intel 10/100 PCI Nic (82559)
Onboard 6-channel sound (which, by the way is DAMN good. Don't need my SBLive now...)
ADS 1394 FireWire card.
Windows XP


Rock stable so far... wish me luck. This processor is AMAZING! Temps have peaked at 46C using entirely retail cooling.

Any of you in the Detroit area, I got the parts at Stone Computer in Royal Oak... PM me if you want the location.

UPDATE: I've been running Prime95 now for about four hours with no errors, peaking at 46C with only one case fan and a closed case. I'm going to go ahead and fire up SETI and get some results for you guys.

Nate

 

[Wolfsraider]

More importantly... is 1.7V safe for a Northwood? Am I going to shorten the life of my CPU noticeably?


truth be told nobody knows for sure what safe truly is but the general consensus is that 1.65 is the cut-off to be safe

do you really want to first to find their theories are correct and destroy a very good overclock just because you chose not to heed the warnings?that should be the question you should be asking yourself


and that is a very sweet overclock


good luck and let us know the answers to your/my question as we patiently await your findings

 

[kevman]

can anyone tell me a good place online to get the P4b266? Also would the p4b266-C be a good enough board? I know the other has usb 2.0, but is there any other advantage to it? Also does the p4b266 require additional standoffs on the case for the P4-Chip's heatsink?


EDIT: Is this an unlocked multiplier CPU???

 

[KenAF]

What do you mean by safe?

Is it going to catch on fire? No. Might running the processor at the voltage reduce its life to four years, instead of 8+ years? Maybe.

There are those running the P4 at 1.9 to 2.2+ volts with liquid or nitrogen cooling...by that standard, 1.7V isn't much.

 

[KenAF]

<< EDIT: Is this an unlocked multiplier CPU??? >>

No, but the new "Northwood" P4 is based on 100MHz (16x100=1600MHz in the case of the 1.6A), so he is simply increasing the FSB up to more Athlon-like levels. Increasing your memory speed from 100MHz to 150MHz gives you a 1600-->2400MHz overclock. At 2600MHz, he is using a base FSB of 162.5MHz, which is nothing special, when you consider some push their Athlon up beyond 200FSB. Of course, no Athlon is going to match the performance of a P4 2600.

Note there are better P4 coolers available than the stock heatsink...who knows, with something like the AVC Sunflower, he might reach 2900MHz.

 

[kevman]

<<

<< EDIT: Is this an unlocked multiplier CPU??? >>

No, but the new "Northwood" P4 is based on 100MHz (16x100=1600MHz in the case of the 1.6A), so he is simply increasing the FSB up to more Athlon-like levels. Increasing your memory speed from 100MHz to 150MHz gives you a 1600-->2400MHz overclock. At 2600MHz, he is using a base FSB of 162.5MHz, which is nothing special, when you consider some push their Athlon up beyond 200FSB. Of course, no Athlon is going to match the performance of a P4 2600.

Note there are better P4 coolers available than the stock heatsink...who knows, with something like the AVC Sunflower, he might reach 2900MHz. >>




Ok,

I'm a noob to overclocking, Whay would I need to increase the voltage if I can simply increment the FSB ?

 

[w9design]

Sometimes the processor can't run at a given overclocked speed on default voltage. By increasing the voltage, the processor is more likely to run at the overclocked speed.

Running at 1.7V (actually 1.68V) is only about 13% above default.

 

[KenAF]

kev,

Like w9design says, at significantly higher frequencies, the processor may require a bit more power. Through the bios, one increases the voltage slightly to provide the processor with sufficient power to run at the higher speed.

From most reports, it seems the $150 1.6A will run at 2.2GHz with default voltage, and 2.33GHz to 2.6GHz with slightly increased voltage. The $200 1.8A and $335 2.0A, in contrast, usually do 2.4GHz with default voltage or slightly increased voltage, and perhaps 2.6+GHz with increased voltage. The $500 2.2A reportedly does 2.6GHz with default voltage, and anywhere from 2.8GHz to 3.0GHz with slightly increased voltage (when using a better cooler like the AVC Sunflower).

Intel looks to release a 2.4A in April, and it will be interesting to see how high it can go. Personally, I am waiting for the new 533FSB boards (current boards will do 533FSB, but I want to overclock to 600+FSB) that are now due in May; they were to come in late March or early April, but the news today is that official availability was delayed until May 20. It also just became known that Intel intends to drop the price of the 2.2A in late May as well, such that it should list for around $220 on Pricewatch (see General forum). $220 for a processor that will run at 2.8-3.0GHz ain't too bad in my book...of course, $150 for a 2.6GHz processor ain't too bad either.

 

[w9design]

Here's the latest update:

-- After about an hour at 2608, I hit a ROUND OFF error in Prime95, so I decided to clock it down a bit, so now I'm at 16x160 (2560MHz). While I was at it I dropped the voltage to 1.675, which on my system is about 1.66V, just a little bit above what is recommended. Prime95 has been chugging for a couple hours like that with no problem.

-- I applied Arctic Silver to the CPU and the chipset heatsink (after removing the existing thermal compound of course). I have booted and used windows at 2755 (have sandra and wcpuid screenshots to prove it), but Prime95 hit a round off error almost immediately after starting. Oh well

-- Current temperature under full load @ 2560MHz with case closed and one exhaust fan is.... 42 degrees celsius!


Question about voltage :

I'm using the 300W power supply that came with my Evercase 4252 case (from newegg)... Right now my voltage is set to 1.675, but it seems to vary between 1.63 and 1.66. Are these fluctuations normal? I've got an Antec PP303X I can try as well, but I'd have to use the EZ-PLUG since it doesn't have ATX12V. Stability doesn't seem to have been affected (no lockups or reboots yet)

 

[jdurg]

I'll be interested in hearing how it goes for you with that case. I bought the exact same case and have an MSI 845 Ultra-ARU mo-bo in there. In about two weeks, I'll have the money to pick up the 2.2 GHz CPU. With the fan from the side of the case blowing right on the CPU socket, an exhaust fan in the back of the case and an intake on the front panel, I'm hoping to reach 2.6 GHz. Let me know how the power supply holds up. I'd like to avoid having to go and get a new one. Especially since I've installed all the hardware but the CPU and the video card. (I think for right now I'll just use my video card in my current computer to get my new one up and running. Might just wait until the GeForce4's are out for sale so that the price of the GeForce3 Ti-500's goes below 250.)

 

[bonkers325]

<< Just past Prime 95 now (on test 40 as i type this)... once i'm sure everything's stable I'll post a screenshot with Sandra and WCPUID.

Here's my setup, in case anyone's interested:

Intel P4 1.6A @ 2.608GHz (@ 1.7V -- is this safe?)
Retail heatsink/fan, using retail thermal interface (haven't even applied my arctic silver yet)
ASUS P4B266 @ 163MHz
256MB Samsung PC2100 DDR @ 163MHz CAS2
Right now, the HD is a lethargic old 10GB/5400RPM Western Digital while I wait for my IBM Deskstar to come back from RMA.
ATI All-in-Wonder Radeon (not overclocked)
Intel 10/100 PCI Nic (82559)
Onboard 6-channel sound
ADS 1394 FireWire card.
Windows XP


Rock stable so far... wish me luck. This processor is AMAZING! Temps have peaked at 44C using entirely retail cooling.

Any of you in the Detroit area, I got the parts at Stone Computer in Royal Oak... PM me if you want the location.

More importantly... is 1.7V safe for a Northwood? Am I going to shorten the life of my CPU noticeably? >>



some mighty fine overclock
damn my 1.6a is coming soon
wonder how high i can go

pray taht mine can come close to yours

 

[BadThad]

SWEET! I'm getting ready to order my AMD killer soon. This is the info I was looking for, thanks for sharing with us.

Personally, I wouldn't worry about the voltage. By the time eletromigration takes ahold, that cpu will be selling for $50 and you'll be looking to buy a 3.5 GHz@5 GHz. LOL

What I really want to know is how it runs SETI with the CLI interface and SETI driver. This is one of the main reasons I'm building one of these....to crank out some work units. How about loading up SETI on it and running for 24 hours? To me, that is the ultimate test. If it survives that, it will last for years.

 

[w9design]

Just finished a SETI work unit in 2.88 hrs. Working on another one now.

 

[Yourself]

I guess my 1.6a is the inverse of yours

barely do 133fsb w/1.6 volts using a Sunflower on an ABIT BD7


Self

 

[CATASTROPHY]

Nice.......im tempted to have you go buy me one..lol

 

[Yourself]

little update on my 1.6a

It appears as though I have a good chip(144X16)2304 mhz once I can get it booted. I ran Prime last night for 7 hours and all is good. The problem is in the board. I'm running an Abit BD7 and I am having all sorts of warm/cold reboot problems. It feels like I'm running my AMD setup w/all the reboot problems. Here are a few of my specs

1.6a@2304@1.625v
Abit BD7
512mb Crucial@Cas 2@2.7v
AVC Sunflower Heatsink

Idle temps-36c
Prime 95/Full load temps-~47c

I think I may try the ASUS P4B266 board....



Self

 

[kevman]

<< little update on my 1.6a

It appears as though I have a good chip(144X16)2304 mhz once I can get it booted. I ran Prime last night for 7 hours and all is good. The problem is in the board. I'm running an Abit BD7 and I am having all sorts of warm/cold reboot problems. It feels like I'm running my AMD setup w/all the reboot problems. Here are a few of my specs

1.6a@2304@1.625v
Abit BD7
512mb Crucial@Cas 2@2.7v
AVC Sunflower Heatsink

Idle temps-36c
Prime 95/Full load temps-~47c

I think I may try the ASUS P4B266 board....



Self >>




Let us know if things change for the better with the ASUS. I'm getting the same setup as w9design did and I'm hoping that I have the same luck as him. In addtion, what is prime? do I need to use it when I OC to ensure data integrity?

 

[gaidin123]

"Prime" is Prime95, a distributed processing client for finding Mersenne primes...ie really freaking large special prime numbers I believe.

It stresses a CPU more than almost anything else and it can detect when the CPU starts getting rounding errors and will flag them immediately. Basically if you can't run Prime95 for a while without getting an error, something is wrong with your system. It's free, they provide a great stress.txt readme file with a good amount of useful info for overclocking. Run the "Torture Test" for a while and see if you get errors. Some people consider their overclock "stable" even though Prime95 gets errors but I'm thinking the system is more stable if your CPU doesn't get rounding errors.

Gaidin

 

[JimmyJoe]

I just ordered two 1.6A's and two P4B266-C's from newegg. I will be using Crucial PC2100 and will post my results after I get it all up and running. I can't wait! I want it NOW!

 

[dbwillis]

I just got my 1.6a about 5 min ago, cant wait to get home and try it out, Ill be using it on a Shuttle AV40 motherboard, hopefully it can run 133 w/no problem (runs a 1.5 at 114x15 fine now)

 

[neovan]

<< Just finished a SETI work unit in 2.88 hrs. Working on another one now. >>


Holy crappers 2.88!!! Thats over 8/day...I gotta get me a couple of those! Now I just need the money

 

[Texmaster]

<< Here's my latest overclocking results, in case anyone's interested:

WAS: Intel P4 1.6A @ 2.608GHz (@ 1.7V -- is this safe?)
EDIT: Not quite. Got a Prime95 error at 2608. See my reply later on for details.
NOW: Intel P4 1.6A @ 2.56GHz (@ 1.66V)
Retail heatsink/fan, using Arctic Silver II on CPU and chipset
ASUS P4B266 @ 160MHz
256MB Samsung PC2100 DDR @ 160MHz CAS2
Right now, the HD is a lethargic old 10GB/5400RPM Western Digital while I wait for my IBM Deskstar to come back from RMA.
ATI All-in-Wonder Radeon (not overclocked)
Intel 10/100 PCI Nic (82559)
Onboard 6-channel sound (which, by the way is DAMN good. Don't need my SBLive now...)
ADS 1394 FireWire card.
Windows XP


Rock stable so far... wish me luck. This processor is AMAZING! Temps have peaked at 46C using entirely retail cooling.

Any of you in the Detroit area, I got the parts at Stone Computer in Royal Oak... PM me if you want the location.

UPDATE: I've been running Prime95 now for about four hours with no errors, peaking at 46C with only one case fan and a closed case. I'm going to go ahead and fire up SETI and get some results for you guys.

Nate >>




Dude get some screenies!

3dmark, sisoft

 

[Amused]

Run 3DMark, and publish your scores on Madonion. I gotta see this.

 

[w9design]

Voila...

Benchmarks at 2560MHz available here: http://pentium4.laferle.com/

My memory scores are weird, now, though.
I was getting about ~2450MBps before at 2-3-3-6 / 160MHz using the 1002 BIOS.
I'm now getting 2380MBps at 2.5-3-3-6 and 2-2-2-5 / 160MHz using the 1005 BIOS set either to Normal, Turbo 1, or to Turbo 2. Something seems weird. There should be SOME variation in scores, right?

And yes, I'm aware that rambus would be giving me ~4GBps of bandwidth right about now at a 3x multiplier. I don't care

My 3DMark2001 scores SUCK. Royally. Why? Because I have an All-in-Wonder Radeon. Not Radeon 8500. Radeon. The Original. But if you're interested, I got 5217 3Dmarks at 640x480x16.

 

[dbwillis]

Well my Shuttle AV40 must not be a great overclocker, cant seem to get the 1.6a above 114 FSB (didnt mess with core voltage)
My 3dmark2001 scores did go up a good amount, from 4397 (1.5@1.5) to 5015 (1.6a@1.77)
Maybe Ill get an Asus and try 133fsb