I got the worst 2500k of all time!

[repoman0]

Looks like it needs about 1.35V to get 4.2GHz, and hits temps of nearly 75C under a Hyper 212+ :hmm: Granted that is with LinX AVX testing which supposedly stresses these chips more than anything. All my other voltages are on auto, but I experimented with them a little and they didn't really seem to change anything. Just needs a whole lot of Vcore compared to other chips I'm seeing.

Anyone else have one this bad? :awe:

Gives me an excuse to get Ivy when it comes out!

 

[Arkaign]

Wow that's terrible. Have you tried cleaning/reapplying thermal compound and reinstalling the HS?

Sometimes (rarely) a CPU or a HS will make it through to a customer that the surfaces are really uneven on. Every one comes with tiny tiny defects in that regard, but some are actually bad enough to cause temp problems like this.

If nothing helps, then damn I'd almost want to sell that SOB lol. That is fairly bad.

 

[richierich1212]

You just got a cpu that needs more TLC, lol.

 

[repoman0]

haha yeah.

I'm sure enough that I installed the heatsink fine to not bother doing it again - temps are pretty good for the high voltage I'm running (1.37 exactly). With prime95 it doesn't even hit 65C, it's just Linx+AVX that kills it.

I got the CPU used on the forums - I think the previous owner might have gone part way to killing it somehow. I guess I can't complain, still got +900MHz and it's way faster than my X4 955, and even with this terribly high vcore probably still draws less power. Still though... :twisted:

 

[Arkaign]

I moved from a 955BE to 2500k as well I've got some cheap Rosewill jumbo cooler w/120mm heatpipe tower and it runs at 4.5Ghz with peak temps around 60-62 (BF3 is the most demanding thing I do really).

 

[Charlie98]

Remember, the 212+ has a fairly uneven bearing surface and you might do well pulling it off and smoothing it out.

 

[Soulkeeper]

just be thankful it's not a llano !
that 1.35V would get you 3.2GHz instead of 4.2GHz

 

[Stuka87]

Boy, thats almost an AMD voltage. My 965BE is at 1.41v to hit 4GHz.

 

[khon]

I know the feeling, since I have a pretty bad 2500k as well.

It doesn't run hot (below 60C at full load), but it simply is not at all stable above 4.2GHz, no matter what I set the voltage to. Any higher than that and it crashes quickly when stress tested.

Still, I had an E6400 previously, so it was a huge upgrade at a very reasonable price.

 

[Puppies04]

Looks like it needs about 1.35V to get 4.2GHz, and hits temps of nearly 75C under a Hyper 212+ Granted that is with LinX AVX testing which supposedly stresses these chips more than anything. All my other voltages are on auto, but I experimented with them a little and they didn't really seem to change anything. Just needs a whole lot of Vcore compared to other chips I'm seeing. Anyone else have one this bad? Gives me an excuse to get Ivy when it comes out! repoman0 is offline Report Post

I know the feeling, since I have a pretty bad 2500k as well. It doesn't run hot (below 60C at full load), but it simply is not at all stable above 4.2GHz, no matter what I set the voltage to. Any higher than that and it crashes quickly when stress tested. Still, I had an E6400 previously, so it was a huge upgrade at a very reasonable price.

What speed RAM are you guys using, also how many DIMM slots have you filled? The point I am trying to make is have you made sure it isn't your RAM that is causing the stability problems before you blame the CPU.

Also are you sure your PSUs are up to the OC?

 

[Rvenger]

You might actually have a defective chip worthy of an RMA. Those are high temps on a Hyper 212+ at only 4.2ghz.

I would try to run the stock heatsink at stock speeds and stress it to see if it exceeds the Tcase temp of 72.5c.

 

[BallaTheFeared]

If it was the ram instability would be occurring at any cpu clock speed.

Just luck of the draw, just look at overclocking as a bonus. I'm probably going to side grade into Ivy, but since my i5 clocks really well I'm actually concerned I might back peddle in performance if the replacement doesn't clock nearly as well.

risk vs reward and all that

 

[Puppies04]

If it was the ram instability would be occurring at any cpu clock speed.

Just luck of the draw, just look at overclocking as a bonus. I'm probably going to side grade into Ivy, but since my i5 clocks really well I'm actually concerned I might back peddle in performance if the replacement doesn't clock nearly as well.

risk vs reward and all that

You sure about that? Is that why my rig won't go over 4.5ghz unless I drop my ram to 1600mhz down from its rated 1866mhz or start tweaking other voltages than vcore....

 

[Rvenger]

You sure about that? Is that why my rig won't go over 4.5ghz unless I drop my ram to 1600mhz down from its rated 1866mhz or start tweaking other voltages than vcore....

Less stress on the memory controller could give you better OC results.

 

[BallaTheFeared]

My OC voltage never changed going from 2 dimms to 4, are you raising VCCIO?

My experience has been that it doesn't matter, but tbh I've never tried my 2x4gb 1600 kit without my 2x2gb kit in. The 4gb is rated at 9-9-9 @ 2000MHz and does 2200MHz 8-10-9 and the other is rated at 1600 9-9-9 but with all four dimms together they do 1866 8-10-9.

 

[Zap]

You might actually have a defective chip worthy of an RMA.

No.

If the CPU does not work at advertised speeds (3.3GHz, "up to" 3.7GHz Turbo) then it is defective and worthy of RMA.

I think you guys are all spoiled. :twisted: Complaining about a 900MHz overclock. Really! This is why "overclocking = YMMV." Doesn't matter what "everyone says."

It could also be something in the platform. Only way to find out would be to use another 2500K and see what it gets. Same clocks? Then something in the platform (mobo/RAM usually). If not, then the CPU just isn't as good.

Understand that even if 4.5GHz is "the average" that means some will clock better, and some worse.

 

[repoman0]

I know the feeling, since I have a pretty bad 2500k as well.

It doesn't run hot (below 60C at full load), but it simply is not at all stable above 4.2GHz, no matter what I set the voltage to. Any higher than that and it crashes quickly when stress tested.

Still, I had an E6400 previously, so it was a huge upgrade at a very reasonable price.

What voltage are you using for your 4.2GHz? I was just tempted to try to lower mine until I remembered that it crashed pretty quickly with 0.05V less than what I'm running now (1.37)

No.

If the CPU does not work at advertised speeds (3.3GHz, "up to" 3.7GHz Turbo) then it is defective and worthy of RMA.

I think you guys are all spoiled. :twisted: Complaining about a 900MHz overclock. Really! This is why "overclocking = YMMV." Doesn't matter what "everyone says."

It could also be something in the platform. Only way to find out would be to use another 2500K and see what it gets. Same clocks? Then something in the platform (mobo/RAM usually). If not, then the CPU just isn't as good.

Understand that even if 4.5GHz is "the average" that means some will clock better, and some worse.

Relax dude, all in good fun! I'm not going to RMA it (never went above 73C on the stock cooler with stock voltages - apparently that's Intel's criteria?) and I know it's luck of the draw. And I'm not actually complaining because this was a pretty cheap upgrade from my 955 and is probably 30% faster including overclock

Wasn't planning on getting Ivy Bridge but I will definitely consider it now however!

Also for the guys who are asking, I'm running 2 4GB sticks at 1600MHz with 9-9-9 timings.

 

[Rvenger]

No.

If the CPU does not work at advertised speeds (3.3GHz, "up to" 3.7GHz Turbo) then it is defective and worthy of RMA.

I think you guys are all spoiled. :twisted: Complaining about a 900MHz overclock. Really! This is why "overclocking = YMMV." Doesn't matter what "everyone says."

It could also be something in the platform. Only way to find out would be to use another 2500K and see what it gets. Same clocks? Then something in the platform (mobo/RAM usually). If not, then the CPU just isn't as good.

Understand that even if 4.5GHz is "the average" that means some will clock better, and some worse.

Zap,

Please read my second sentence. I said if his chip is running too hot at stock speeds, it could have a poorly mounted heat spreader. Hyper 212+ and 2500k @ 4.2ghz should not hit 75c, even at his said voltage.

 

[Puppies04]

Zap,

Please read my second sentence. I said if his chip is running too hot at stock speeds, it could have a poorly mounted heat spreader. Hyper 212+ and 2500k @ 4.2ghz should not hit 75c, even at his said voltage.

Depends completly on ambients and case cooling but I agree it does sound a bit toasty for that setup

 

[repoman0]

Zap,

Please read my second sentence. I said if his chip is running too hot at stock speeds, it could have a poorly mounted heat spreader. Hyper 212+ and 2500k @ 4.2ghz should not hit 75c, even at his said voltage.

I lowered the chip to 4GHz and reduced the voltage by about 0.07V (any lower and it BSODs after a half hour to an hour of LinX) and now the two hottest cores reach 70C. Although my motherboard reads an "overall" CPU temperature of 60C. Still too hot? Which should I trust?

 

[repoman0]

Depends completly on ambients and case cooling but I agree it does sound a bit toasty for that setup

Case cooling is good as are ambients, my X4 955 ran overclocked (though slightly undervolted) at about 50C

 

[douglasb]

What speed RAM are you guys using, also how many DIMM slots have you filled? The point I am trying to make is have you made sure it isn't your RAM that is causing the stability problems before you blame the CPU.

The RAM would have absolutely no bearing on the high temperatures he is complaining about. It might contribute to his stability issues, but it wouldn't make his chip run so much hotter than it should be. I think he might just have a below-average overclocker. i5 quad at 4.2 is nothing to sneeze at, though, and should be fast for almost any application. OP should be happy with that.

 

[Maximilian]

Lousy temps are just a side effect of high volts no?

My 2500k gets those temps at 1.35v, it can hit 80*C on core 2.

OP play around with the different voltages, VTT is one that could help. Even if you dont get to 4.5 you are only 300mhz off at 4.2 so it dosent matter so much

 

[nyker96]

I'm surprised, from most report, the newer made 2500k seems to clock even better than when it first came out. But 4,2 isn't bad at all. It packs a hefty punch at 4,2ghz. Besides when you upgrade again couple of years down the line, the difference looking back between a 4,2ghz sandy and a 4,6-4,8ghz sandy isn't much.

 

[Subyman]

Damn, there are some high temps here. I was pushing 1.35V with a H60 and only hit 50C at load. From what I have seen, most poor OCs are due to bad motherboards or PSUs that aren't giving clean voltage or they are drooping. If you are running a lower end motherboard, then that may be your problem. I'd also reseat the heatsink making sure to use proper thermal compound and really clean the surface. Use 90% alcohol with Q-tips. You can check the flatness of the CPU and heatsink surface by finding a thick, flat piece of glass and laying them on top to see if they will rock back and forth.

You can check your voltage drop by turning on fixed multiplier in the bios, turning turbo off, and making sure all the power save options are off. Then go to windows, pull up hardware monitor and check your idle voltage. Load prime95 and fire up the workers. See if that voltage drops much. I've seen poor PSUs/Motherboards drops as much as .1-.2V, which will kill an OC.