Ok for my project i have been working on i got a 2500+ barton for the cpu. Well at first bios and windows said it was a 2500 barton intill today. Now it says its a 3200+ barton. Both bios and windows said it. So i ran cpuz and it said the same thing I checked the specs and stuff and the cpu is running at the same fsb and multiplyer as a 3200 barton when yesterday it was running at 2500 barton speeds? i got the cpu 3 months ago so it should be one of the super locked ones? I am not complaining but what the hell happened?
I've seen this happen a few times with the Barton core XP's. From what I've heard, the explination is basically that while overclocked, the core basically just gives, and it's read as whatever speed its running at the time. If you're worried about it, I would clock it back at the normal 2500+ speeds and see what it's read as.
ok ill do that right now. Im not worried im excited edit your right when i dropped the fsb down itsaid it was a 2500 barton but when i set it to auto is sets it to 3200 speeds?
its at 85degrees F as im writting this (watercooled) I tried a higher fsb but my ram didnt like anything highier than 200. 1.75 what would that do? It is allready stable can play video games all day on it.
^What he said is right, I have a 2500+ Barton, it doesn't even run stable at 200x11 1.775v I have it 1.8v and it runs stable under 60°C stock fan/hsf on load.
I have my 2500 oc'd to 3200 speeds, only thing is i have to run it at 1.85volts.. however it has been running just about 24/7 like this since feb My Thermalright SP-97, with a 90mm fan silent fan, keeps it nice and cool at 37-39C idle and 45 ish load. Although, I was building an identical setup for a friend a while back and his barton could do the same speeds at default core voltage and that idled at ~30C (hsf - aero 7 lite). I shoulda just nabbed it
yep mine is a default voltage 1.65. Wierd. Im going to pull the waterblock off this weekend and copy the code on it and see if i didnt get a 3200 by mistake.
Pretty sweet mate . From what i rem, there are a few decent 2500+ that do 3200 at stock, but the XP-M are just great - i know someone that could do 3200+ speeds at 1.55v As for codes, there are a few things you want to look out for: - the part on the second line that says AQXCA etc. Generally speking we have AQXCA < AQXDA < AQXEA < AQUCA (this was the code on most 3200+) < AQYHA < IQYFA < IQYHA etc etc... or something like that. This isnt also the be all end all for the quality of CPU you have. The last 4 digits of the second line also play a part in the quality - the lower the better (something with 0003 on the last digits of the first line would be better than a 4582, even if the 0003 is a AQYHA and the 4582 a IQYHA). Ive also heard the ones on a brown PCB are better than the green but dont know for sure. If ive got anything wrong, please let me know.
I'm building a new system with a Barton 2600+, does this mean that I can oc it much higher? Now I'm getting pumped.
Different chips OC to different speeds. The only speed they have to be stable at is their default clock speed. Some XP 2400+'s that get off the line can sell as 3200+'s, some can just do their speed and not a Mhz more stable. Just all depends on your chip.
Sure all chips are different, but it sounds like he has had no problem running 3200 at stock vcore. Now that I know its watercooled, I'd say bump it to 1.85v and 210FSB, unless your memory wont follow (put its voltage to the max ~2.9v). You may have to slightly losen the timings, but the clock increase will overshadow it. Chances are that heat wont hold you back (esp if you are at 85F now!) so see what it is capable of.
The original rating of the CPU makes little difference TBH. You might get a 2600 that can only do 2600 speeds at stock volts, or you might get one that does 3200 (or more!) at stock - you just have to wait untill it arrives. Even if you dont get 3200 speed at stock, you could always pump a few more volts (1.75-1.8 is fine imo) and see how far you can puch it. The 2600+ bartons are rare arent they? Most were thoroughbreds if im correct (256k cache + higher clock speed instead of barton 512k cache and lower clock speed). The 2600XP-M were all barton tho. Id also agree with user0001 - if youve got water, you HAVE to pass at least 1.8 volts through it. What would be great is if you got your 2500+ to run at 2.5Ghz )
How do u get a fairly accurate of Vcore then? AFAIK, different boards and different bios versions all behave differently. Some undervolts, some overvolts. I think NF7 series quite overvolt their Vcore, eg. if it shows 1.575 in the Bios, probaly the actual figure is just about 1.5 . And Monitor programs like CPU-Z, WCPUID, or MBM, all take data from the Bios
I think the way this works is kind of as you've said. You set the CPU core voltage (1.65 say), and on a mobo like most NF7s the bios will read a voltage which fluctuated between 1.61-1.64 (for example - biggish drop, but small fluctuations). This is undervolting (the mobo providing less volts than what you have set it to). By increasing the Vcore setting to 1.675 it'll probably go up by to 1.635-1.665. The Vcore setting will never always always be same as the actual voltage going through the chip, but the bios will pick up the fluctuations in this. Programs like MBM use the winbond monitoring chip (i think this is the same chip that the bios uses for temp/volt monitoring) so it will show you the same voltages that you would expect in the Health Check menu in the bios. The only way you can really get an accurate reading of you Vcore is a voltmeter, but tbh just checking it in the Health menu in the bios should be ok for most people. Also note, that some of the NF7 undervolts more when under load (say from my example 1.65v running at 1.61-1.64 v, under load this might drop even more to 1.6-1.63).
i had a 2600+ mobile that ran at 1.65v 200x12 prime stable for over 24hrs. i could have gone higher but i was only using a CoolerMaster HHC-001 at the time. XP-m's rock! now i have a 2.0ghz P4-m at 3.0ghz. that chip was free.