Small update.
FSB: 420 (erryday)
CPU 4.2ghz with a 10x multi
1.34 Vcore
8gb kit at 1680, linked 1:1 with the FSB, 8/8/8/20, stock timings are 9/9/9/24 at 1600, so I'm pretty happy.
I managed to squeeze some more juice from the cpu, whilst lowering the temps by lapping the cpu, it was concave like mad, and I managed to get over 10c decrease by lapping c:
Doing 470 in cinebench R15, I'm content, dis gud.
Damn fine work.
A few things to consider here.
CPU and motherboard isn't the same socket.
Bios is custom.
Ram modules are much newer than the motherboard.
Said motherboard doesn't like using all ram slots.
CPU has been running years in a server.
People struggled to hit 4ghz with the Q9650 at release.
Motherboard is ancient, and has been running 24/7 since I bought it back in 2008, so caps and mosfets are likely degraded.
To compare release dates, its like running games released in 2008 on the highest settings etc, with a late pentium 3, or a very early pentium 4, as it was released in 2002.
I'm very happy with 4.2, with such a old cpu, with such a old motherboard, and with such low voltages and temps.
Hey there, going to be doing this'd myself soon hopefully. Running Asus P5Q premium with modded BIOS. Just cut the board down, haven't applied the little sticker to the chip yet. Will aim to report back!
Nice, those xeon chips are really nice overclockers, especially the x5470 with its x10 multi.
Don't forget to put the sticker in the right place :D
Nice one buddy! Swift reply too - thanks.
Well, it boots to post / bios OK! Now for OS I guess. Lol. Maybe enough for tonight, at least the really funky stuff is done, on with the tinkering.
Obviously I'm gonna lap it mind, just wanted to test it worked first...
But all in good time.
If you're going to lap it (like I did) Try taping the pads and stuff so you can handle it easier, electrical tape works fine.
When peeling the tape off though do it slowly, since removing tape fast actually creates static discharges.
Sounds like good advice. Please can you explain what you mean by taping the pads? I guess that they are the contacts and jazz in the the back?
Yeah, when lapping its a good idea to tape over the contact pads/landing pads, those gold circles more or less.
possible reasons for instability is the cpu supports up to 1333mhz ram up to 1.35v's nice job on the 771 mod
High end Xeons at 45nm should all do 1600mhz more or less, what you get depends more on your motherboard's north bridge.
Mine doesn't want to go over 1900 on the FSB.
Realtemp is reporting 58~65 degrees c idle... Funny thing is the hardware monitor in the bios (of the says p5q premium that I'm running the X5470 on) reports the cup 25-30. Could real temp be reporting wrongly?
Don't use realtemp, stay with something like speedfan.
Haven't tried speed fan yet, but hwmon and occt also both report the core temps high (which is what I've always (conservatively) gone off as opposed to tin).
It could be a dodgy seating of my cooler I guess.
Am running 3.8 ghz at 1.15 (not having tried very hard). Cant run intel burn test though, temp goes above 70, which is out of my comfort zone.
What parameters do you oc to, in terms of Max temp etc?
Don't run intel burn test on core 2 quads, they really don't like it.
3.8ghz with 1.15 is very good, like very very good. I maxed around 65c or so on the cores with a dual 120 AIO.
4.2 with 1.34v.
Hi,
just bought an already modded X5470 on Aliexpress, an Asus P5Q SE2 mobo, downloaded the latest modded bios from some russian site and everything went fine until i fired up Prime95.
There were 2 options. Failing prime almost instantly and/or Windows error message then autorestart. The second option came at upper vcores: stable prime but immediate system shutdown after a while. The higher the vcore the faster shutdown (under full load by prime ofc).
Then I had a hard time finding the sweet spot even for the stock 3.33ghz. It is 1.25 vcore and 1.3v FSB. No shutdown and no error in prime.
However it is very strange that core0 is always at way higher temp than the others. Currently running prime and it's 63/57/57/57 C an sometimes the difference between core 0 and others is 8-9 C. The shutdown is independent from temps because sometimes it shuts down at 65, sometimes 74 for example.
So the situation is as follows: the cpu is stable even at 4.ghz (fsb 400, ram 800) on higher voltages and temperatures but the PC shuts down even at normal temps after a while under full load (after about 10 mins if settings are well balanced). Prime fails at lower vcores but no shutdown, "only" error message+restarts.
Ram is 4GB Corsair XMS2 DHX C4. Cooler is brand new Cooler Master Hyper TX3 EVO.
I have tried playing with various other setting in bios but no success.
Did anyone run into the same? What could be wrong?
Any suggestion welcome, thanks
Looked into the power supply? reseated the CPU?
Core temps varying isn't unusual, most of the time its just the thermal sensors being off more than anything else, so your CPU is running fine.
Checked the RAM as well, 1 stick, 2 and so on.