Is sandy-bridge far more difficult to overclock than Core 2?

[TBSN]

There used to be a sticky for overclocking C2Duo and Quad processors, but not for the newer processors (nehalem, sandy bridge).

I looked around online for a guide like the one we had here for C2D, and it seems like overclocking is WAY more difficult. A ton more voltages to consider, making multiplier adjustments for normal AND turbo modes, for each core, etc.

How much more difficult is it than C2D? I'm thinking of upgrading but if it is too difficult than I'll go with a non- "K" processor.

 

[smakme7757]

There used to be a sticky for overclocking C2Duo and Quad processors, but not for the newer processors (nehalem, sandy bridge).

I looked around online for a guide like the one we had here for C2D, and it seems like overclocking is WAY more difficult. A ton more voltages to consider, making multiplier adjustments for normal AND turbo modes, for each core, etc.

How much more difficult is it than C2D? I'm thinking of upgrading but if it is too difficult than I'll go with a non- "K" processor.

Overclocking SB isn't hard. You just up the multiplier and you're good to go. Usually the AUTO voltage can compensate, although fine tuning always give a better result.

Overclocking on any platform isn't hard, but SB has made it slightly easier.

 

[exar333]

SB overclocking is actually easier IMHO than C2D/C2Q. You just need a decent MB and a 'K' CPU and you are good to go. Up the multiplier (and likely the voltage a bit) and test it out. It's VERY easy.

You don't even usually need fast RAM either, because you can adjust the frequency to run at 2/3, 3/4, 4/5, etc. of the adjusted speed. SB doesn't really benefit from anything higher than DDR3 1600 anyway, so there is little point to go higher other than benchmarks.

 

[RussianSensation]

Overclocking SB is easier than any processor in history outside of XP2500+ (or at least off the top of my mind). I think you are way over-complicating things. All you need to do is adjust load line calibration to enabled, set VCore to 1.35-1.38V, set a lower RAM:CPU ratio and crank Multiplier to 44-47x and you are done. You may need to increase CPU PLL voltage for multipliers 48-50x.

The other part is using VCore offsets at load, but you should do that AFTER you figure out what voltage you need to be stable at load. That's the "tuning" stage.

It absolutely is NOT difficult to overclock SB CPUs. Overclocking has never been easier since you no longer have to worry about locking PCIe, playing with northbridge or southbridge voltages, playing with RAM ratios (since memory bandwidth is no longer important), tuning QPI/FSB frequency (those are not important either).

Honestly, if you are that worried, just get a MSI P67/Z68 board. They have excellent 5 second overclocking button (and also Gigabyte and Asus boards have automatic overclocking). Automatic overclocking works pretty well but tends to be on the conservative side, both in regard to upper frequency ranges and applying too much voltage.

Just look at this excerpt from HardOCP review:

Overclocking

"I’ll keep this relatively brief. Overclocking the P8P67 is a lot like overclocking any other Sandy Bridge setup and certainly just like overclocking the rest of the P8P67 family. I’ve done that a lot now so I pretty much just go right for what works. Even with that said, the P8P67 was very nice to work with. I started off by setting the DIGI+VRM control for "Ultra High" and the CPU current control to 130%. With this alone I was able to hit a multiplier of 47x without doing anything other than raising the multiplier. The voltages came up automatically. I hit the wall there.

This was the limit I had seen on a few other boards. After enabling CPU PLL overvoltage I was able to go higher. I eventually hit a speed of 4.916.2MHz with a 49x multiplier. I didn’t even have to adjust the CPU voltage manually. It automatically came up to 1.416v which is where I ultimately needed to be anyway. I tried to hit 50x on this CPU to no avail. No matter what settings I tried, manual and otherwise it wouldn’t do it. I had never seen this chip do that, so I might have simply been at its limits." ~ Full review

 

[toyota]

or you can just set the turbo to 4.2 for all four cores and leave basically everything else alone or on auto. then you can idle at 1.6 and go to 4.2 only when under load.

 

[VirtualLarry]

Aww, shucks. You guys are tempting me to upgrade to SB. Must resist! (No money!)

 

[toyota]

Aww, shucks. You guys are tempting me to upgrade to SB. Must resist! (No money!)

just wait for Ivy Bridge.

 

[Idontcare]

Overclocking SB is easier than any processor in history outside of XP2500+ (or at least off the top of my mind). I think you are way over-complicating things. All you need to do is adjust load line calibration to enabled, set VCore to 1.35-1.38V, set a lower RAM:CPU ratio and crank Multiplier to 44-47x and you are done. You may need to increase CPU PLL voltage for multipliers 48-50x.

The other part is using VCore offsets at load, but you should do that AFTER you figure out what voltage you need to be stable at load. That's the "tuning" stage.

It absolutely is NOT difficult to overclock SB CPUs. Overclocking has never been easier since you no longer have to worry about locking PCIe, playing with northbridge or southbridge voltages, playing with RAM ratios (since memory bandwidth is no longer important), tuning QPI/FSB frequency (those are not important either).

Honestly, if you are that worried, just get a MSI P67/Z68 board. They have excellent 5 second overclocking button (and also Gigabyte and Asus boards have automatic overclocking). Automatic overclocking works pretty well but tends to be on the conservative side, both in regard to upper frequency ranges and applying too much voltage.

Just look at this excerpt from HardOCP review:

Overclocking

"I’ll keep this relatively brief. Overclocking the P8P67 is a lot like overclocking any other Sandy Bridge setup and certainly just like overclocking the rest of the P8P67 family. I’ve done that a lot now so I pretty much just go right for what works. Even with that said, the P8P67 was very nice to work with. I started off by setting the DIGI+VRM control for "Ultra High" and the CPU current control to 130%. With this alone I was able to hit a multiplier of 47x without doing anything other than raising the multiplier. The voltages came up automatically. I hit the wall there.

This was the limit I had seen on a few other boards. After enabling CPU PLL overvoltage I was able to go higher. I eventually hit a speed of 4.916.2MHz with a 49x multiplier. I didn’t even have to adjust the CPU voltage manually. It automatically came up to 1.416v which is where I ultimately needed to be anyway. I tried to hit 50x on this CPU to no avail. No matter what settings I tried, manual and otherwise it wouldn’t do it. I had never seen this chip do that, so I might have simply been at its limits." ~ Full review

^ this

I just upgraded from a Q6600 setup to a 2600K setup.

The OC'ing aspects of the 2600K just couldn't be any easier. No really! That's not hyperbole, I can't fathom how they could make it any easier.

From within windows I can adjust my CPU overclock dynamically, on the fly, as I see fit. And its all just multiplier ratios. No more futzing around with straps and FSBRAM rations and so on.

It just all happens in the background automagically.

I take that back, it could be on step easier...it could be as easy as overclocking a video card with that MSI Afterburner utility in which you just drag a slider from within windows and you dynamically dial in any MHz and Voltage you desire (within the programmed limits of course).

But still, OC'ing today's leading edge CPU's is nothing like OC'ing chips from the C2D era.

 

[pantsaregood]

just wait for Ivy Bridge.

die shrink

 

[toyota]

die shrink

so? if you think Ivy Bridge is not going to be faster then you are wrong. they will likely keep the same TDP envelopes so that will allow for a large frequency and/or cache increase. it may not be worth upgrading to for Sandy Bridge owners but its a better option for Core 2 owners to upgrade to if they feel they can wait.

 

[blackened23]

so? if you think Ivy Bridge is not going to be faster then you are wrong. they will likely keep the same TDP envelopes so that will allow for a large frequency and/or cache increase. it may not be worth upgrading to for Sandy Bridge owners but its a better option for Core 2 owners to upgrade to if they feel they can wait.

Isn't IB a ways off? Last I heard on overclock.net , it was delayed until spring '12. I could be wrong though.

 

[toyota]

Isn't IB a ways off? Last I heard on overclock.net , it was delayed until spring '12. I could be wrong though.

yes probably spring 2012. that's not that far away IF he is pretty happy with his current system though. if he really needs something now then of course Sandy Bridge is the way to go.

 

[VirtualLarry]

Yeah, I am satisfied enough with my current rigs to wait for Ivy. I'll have to save money until then to afford it, too, unless I can sell some of my custom brand-new rigs I built last year. Trying to sell them at electronics flea markets proved somewhat futile. Got over $1000 invested in custom rigs that are slowly becoming obsolete.

Edit: You think that the ASRock Fata1ity P67 Professional board (P67 with IDE and floppy), will support IB, and PCI-E 3.0? I really, really want a board with IDE and floppy. If it's not compatible with IB, though, then I guess I'll have to go 2600K. (Maybe 2500K. Thought I would like the option for 8 threads for DC.)

 

[RussianSensation]

All P67 and Z68 boards should be compatible with IB with a simple BIOS update. I was under the impression that Intel would be launching P7x series chipset to coincide with IB launch. P7x chipsets would have native PCIe 3 support and at least 4x SATA 3 ports off the Intel chipset. But given how much you have invested in your systems, I wouldn't advise you to keep upgrading just for the sake of throwing $ away.

If you are having trouble selling your Q9300/6600 rigs, my suggestion is to either sell the CPUs alone Or try bundling them as Mobo + CPU + Ram as a package. I am sure there are people who would be interested in upgrading their E6400/6600/6700, etc. dual core systems on socket 775 and would buy your Q6600 in a heartbeat.

 

[VirtualLarry]

If you are having trouble selling your Q9300/6600 rigs, my suggestion is to either sell the CPUs alone Or try bundling them as Mobo + CPU + Ram as a package. I am sure there are people who would be interested in upgrading their E6400/6600/6700, etc. dual core systems on socket 775 and would buy your Q6600 in a heartbeat.

No, I'm not selling those. I built three rigs with Athlon II X4 630 chips (2.8Ghz quad-core), 4GB DDR2 (which I paid a pretty penny for, it was still $100 for 4GB of DDR2 when I built them, sadly), and I have some HD4850 512MB cards to throw in with them to make quad-core gaming rigs.

I spent $340 on each of them, without counting the Win7 license ($50), and the HD4850 (which I valued at $80, I paid $166 with tax when they first came out.)

Edit: (changed 2.8Ghz dual-core to 2.8Ghz quad-core)

 

[Meghan54]

The OC'ing aspects of the 2600K just couldn't be any easier. No really! That's not hyperbole, I can't fathom how they could make it any easier.

Maybe send an Intel employee or an employee from your chosen motherboard manufacturer to set it up for you? About the only way it'd be easier.

 

[BonzaiDuck]

There used to be a sticky for overclocking C2Duo and Quad processors, but not for the newer processors (nehalem, sandy bridge).

I looked around online for a guide like the one we had here for C2D, and it seems like overclocking is WAY more difficult. A ton more voltages to consider, making multiplier adjustments for normal AND turbo modes, for each core, etc.

How much more difficult is it than C2D? I'm thinking of upgrading but if it is too difficult than I'll go with a non- "K" processor.

That overclocking guide has migrated verbatim to another forum. What's worse, I was juggling too many items when I came across it, and forgot to remember the URL. It had become a "sticky" there . . .

 

[qliveur]

Overclocking Nehalem was more complicated than overclocking Core2, yes, but Sandy Bridge should be far simpler than both.

I enjoyed overclocking my 930 to the limits of my cooling. It was an adventure.

 

[WhoBeDaPlaya]

Multi-overclocking is dead easy. Nehalem overclocking could be a little daunting, since you had to ensure that so many things (uncore, BCLK, maybe RAM, etc.) were all holding up.

 

[StrangerGuy]

It's so much easier with BCLK and RAM out of the equation. LGA775 you need to find out which board did the most FSB, how far your DDR2 can hold up 1:1 to said FSB, DDR2/NB voltages, NB internal latency over FSB ranges/FSB holes and whether your chipset or mobo flaks out with 4 DIMMs instead of 2. Besides SB speestep and C1E works perfectly even when overclocked, same can't be said for 775.

 

[scooterlibby]

I almost felt guilty with the ease in which I OC'd my 2500k compared to the Q9550 and the Athlonx2 before that. I have a great deal of respect for the people that had to solder and really understand electronics to OC with earlier CPU's. Anyone remember that Tom's article about the ethics of taking a Pentium from 100 to 200 mhz? That was actually a controversy!

 

[TBSN]

I almost felt guilty with the ease in which I OC'd my 2500k compared to the Q9550 and the Athlonx2 before that. I have a great deal of respect for the people that had to solder and really understand electronics to OC with earlier CPU's. Anyone remember that Tom's article about the ethics of taking a Pentium from 100 to 200 mhz? That was actually a controversy!

I've been trying to dig up that article, if you find it post it here, it sounds interesting!

 

[BathroomFeeling]

You mean this one?
http://www.tomshardware.com/reviews/overclocking-guide,15.html

 

[Majic 7]

For someone like me that doesn't have years of experience fiddling with voltages and parameters SB is perfect. On my 9550 I kept it at 3.6 because I didn't have the patience or experience to keep it stable above that, not long term anyway. Spent hours and hours at XtremeSystems researching, all for naught. Now I can get 4.2 at stock volts with minimal effort. Did have it at 4.3 but as soon as I added 8GBs more ram it got unstable. Lowered to 4.2 abd passed 12 hrs of memtest and 13 hrs of Prime. Just too easy.

 

[grkM3]

sandy is made to overclock easily,intel even made a commercial with a grandmother showing how easy it is to overclock.

Heck on my asus mive board If I didnt even need to tweak any sub timing or anything else for a crazzy overclock all I have to di is raise the v core to 1.45 and set multy for 50x and I have a stable 5ghz oc,its that simple.

The xmp ram profiles run the memory at there set timings and there isnt much more you need to do.

Here is the intel commercial

http://www.youtube.com/watch?v=_GR5_X1CfUA

this is a real add from intel lol

it really is that easy since most boards will up the vcore on there own so all you have to do is change the multy.