I got the worst 2500k of all time!

[Rvenger]

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?

Did you try reseating that heatsink? Something doesn't sound right.

Stress it at stock to see if it errors. I am wondering if this is memory/mobo related.

 

[repoman0]

Did you try reseating that heatsink? Something doesn't sound right.

Stress it at stock to see if it errors. I am wondering if this is memory/mobo related.

Actually these temperatures are pretty much in line with others who are using around this voltage - saw one guy with the same cooler doing 1.38V and getting 75-80C core temps and 65C CPU temp. Only difference is his was at a higher clock speed, but that's a rather small heat increase compared to voltage.

I will try reseating anyway though.

It didn't error at stock, I checked it when I got it using the stock cooler. The memory was very stable in my Phenom rig so I'd be surprised if that was it. And the motherboard seems pretty much fine...

Tempted to walk over to Cambridge and get a $180 Microcenter 2500k and see what it will do - man I hate this overclocking game, 4.2GHz is incredibly fast yet I want more Reminds me of when I used to buy Opteron 165s seemingly by the tray from Monarch Computer to try and get a 3.2GHz chip. The difference was back then I could sell them on ebay for more than I paid :whiste:

 

[Rvenger]

If you sell that old 2500k for $185 on ebay, I bet someone will buy it in a heart beat. If it works good at stock speed then its a good CPU.

 

[BTRY B 529th FA BN]

Have you tried a BIOS update? or it could just be a bad board. My Gigabyte UD3R was way better than my Asus Maximus ROG board

 

[jimpatrick]

maybe w/ bios settings?before my 2600k needs 1.35vcore for 4.5 then i enabled pll,disabled spread spectrum then llc 1 after that it just needs 1.3,havent tried lower yet.

 

[Achilles97]

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.

This ^

I had bad temps with my 2500k at first, with a Mugen3 heatsink, until I realized that it was not flat. See my post with pics here:
http://www.overclock.net/t/1154002/scythe-mugen-3-base-is-concave#post_15486422

 

[Zap]

Reminds me of when I used to buy Opteron 165s seemingly by the tray from Monarch Computer to try and get a 3.2GHz chip.

Heh, I did that in the year before they went out of business (and floated money using customer's credit cards ). I purchased a bunch of Opteron 144 OEM chips from them. One of them turned out to be a terrible overclocker (around 2.2GHz barely) while most of the others were the "normal" 2.7GHz and one was near 3GHz. When I mentioned it here, everyone gave me crap about it and blamed my other parts or just outright called me a liar or noob. Thing was that I used same exact setup, and could swap out that "bum" chip for another one and have totally different results.

 

[Rvenger]

Heh, I did that in the year before they went out of business (and floated money using customer's credit cards ). I purchased a bunch of Opteron 144 OEM chips from them. One of them turned out to be a terrible overclocker (around 2.2GHz barely) while most of the others were the "normal" 2.7GHz and one was near 3GHz. When I mentioned it here, everyone gave me crap about it and blamed my other parts or just outright called me a liar or noob. Thing was that I used same exact setup, and could swap out that "bum" chip for another one and have totally different results.

Yep I had an Opteron 146 and 144. Actually I think I sold them both on here as well. The Opty 144 ran 2.7ghz and the 146 was able to hit 2.8ghz. They were really fun chips.

 

[grkM3]

this is with water cooling and 4 fans on a 2600k.Im also running windows 7 64bit with sp1 installed and avx enabled and using a version of intel burn test that supports avx.

Id like to see some of these colder cpus running this version of intel burn test and a temp program with the offsets set to 0,you will be suprised how hot your cpu will get when you run full avx

Here is the same exact computer running a version of intel burn test that dosnt support avx,keep in mind that this was done right after the first test ran for 13-15 min and was on a cold system with the water cooled and then re ran right after with it being heat soaked from the test before.

 

[Rvenger]

this is with water cooling and 4 fans on a 2600k.Im also running windows 7 64bit with sp1 installed and avx enabled and using a version of intel burn test that supports avx.

Id like to see some of these colder cpus running this version of intel burn test and a temp program with the offsets set to 0,you will be suprised how hot your cpu will get when you run full avx

Here is the same exact computer running a version of intel burn test that dosnt support avx,keep in mind that this was done right after the first test ran for 13-15 min and was on a cold system with the water cooled and then re ran right after with it being heat soaked from the test before.

Considering thats a 2600k with HT, I would give that a 10c advantage. My old 2600k had the same exact temps @ 4.3ghz w/1.35v. OILFIELDTRASH might be able to confirm similar temps as well.

grkm3- whats up with your activation?:hmm: :awe:

 

[grkM3]

Considering thats a 2600k with HT, I would give that a 10c advantage. My old 2600k had the same exact temps @ 4.3ghz w/1.35v. OILFIELDTRASH might be able to confirm similar temps as well.

grkm3- whats up with your activation?:hmm: :awe:

activation got wacked when I shut off hyper threading in the bios and windows dected new hardware and you know the rest.

I have 4 paid keys for windows 7 but dont feel like calling this in as it works like normal and im always changing my settings

 

[Rvenger]

activation got wacked when I shut off hyper threading in the bios and windows dected new hardware and you know the rest.

I have 4 paid keys for windows 7 but dont feel like calling this in as it works like normal and im always changing my settings

I hear ya, I have 2 paid keys and I am always tweaking my rig so I don't even activate anymore because I wipe it out before the 120 days anyways.

 

[repoman0]

So I did end up going to Microcenter to grab another one - $180 was too good of a price and I figured I could easily recoup losses selling it if needed.

This one is a half hour into LinX+AVX at 4.2GHz with BIOS set at 1.24V+LLC level 5, which results in 1.27V at full load as read by Gigabyte's software - supposedly very accurate. My max CPU temp as read by the software is 60C, and max core temps as read by HWMonitor are 69C for the hottest and 63C for the coolest. Definitely better and my ambients are pretty high right now - closed my window so that my apartments absurdly hot heat would really test the chip

So I guess this one will probably run 4.4-4.5GHz at the 1.37V that the other one needs for 4.2. Not a huge difference and still not an amazing chip but I bought it for the fun of testing it out!

 

[nyker96]

So I did end up going to Microcenter to grab another one - $180 was too good of a price and I figured I could easily recoup losses selling it if needed.

This one is a half hour into LinX+AVX at 4.2GHz with BIOS set at 1.24V+LLC level 5, which results in 1.27V at full load as read by Gigabyte's software - supposedly very accurate. My max CPU temp as read by the software is 60C, and max core temps as read by HWMonitor are 69C for the hottest and 63C for the coolest. Definitely better and my ambients are pretty high right now - closed my window so that my apartments absurdly hot heat would really test the chip

So I guess this one will probably run 4.4-4.5GHz at the 1.37V that the other one needs for 4.2. Not a huge difference and still not an amazing chip but I bought it for the fun of testing it out!

this is the reason I don't like to buy off ebay, most the chips are poor overclockers that's why they got sold off. I know some people goes out and buys 3 onsale then test them all, sell off the poor clockers on ebay.

 

[grkM3]

Repo you are not mounting the coolers right.even 69 is way to high fir 1.27 volts.that chip i sold u did 4.7-8 ghz at under 1.4 volts on my mive board.i had a gigabyte ud7 (380 dollar)board and it wouldnt clock over 4.4 and i returned that board fir an asus.

You dont have a good contact on that cooler bro.

 

[repoman0]

Repo you are not mounting the coolers right.even 69 is way to high fir 1.27 volts.that chip i sold u did 4.7-8 ghz at under 1.4 volts on my mive board.i had a gigabyte ud7 (380 dollar)board and it wouldnt clock over 4.4 and i returned that board fir an asus.

You dont have a good contact on that cooler bro.

Did you use Linx with AVX to test? I'm doing Prime95 now (linx passed for 100 minutes) and it's running at 51C max - 4.2GHz with 1.28V. Those AVX instructions are killer. That is the heatspreader temp as read by the gigabyte utility, core temps are about 10C higher - I've read that I should be using the heatspreader temp. I am very sure that the heatsink is mounted right, I've done it literally a hundred times in the past. I think it's just that people aren't used to seeing the AVX temps that linpack can put out, and also I've been posting the core temps which I guess don't really matter.

I think you might be right that the motherboard is a bit limiting, I have a GA-Z68XP-UD3. You were having the same results pretty much on the UD7 it seems - maybe these boards aren't great overclockers?

 

[grkM3]

Ill never buy a gigabyte board again.nothing clocks like my p67 mive and this board was also north of $360 i can bench this 2600k at 5.2-5.3ghz with 4 sticks at 2133 at tight 7-10-7 timings and the ud7 could not even boit into windows at 4.4 ghz it was max stable right around where you are now

 

[grkM3]

If you want a cheap great clocking board grab a p67 pro from asus.that little sucker clocked just like its older brother mive board for half the price.

 

[repoman0]

Might grab this one - http://www.microcenter.com/single_product_results.phtml?product_id=0371775

Should be $70 assuming I can get them to apply that $50 discount even though I already bought the CPU. I definitely want Z68 for the fast video encoding, and it looks like that board has everything I want. Goddamn this is addicting

 

[grkM3]

Might grab this one - http://www.microcenter.com/single_product_results.phtml?product_id=0371775

Should be $70 assuming I can get them to apply that $50 discount even though I already bought the CPU. I definitely want Z68 for the fast video encoding, and it looks like that board has everything I want. Goddamn this is addicting

you got me benching mine again also lol

try and see if you can return the gigabyte/I called newegg and they took mine back because I bought another board in the same price range even tho it said exchange only in the sale.

how high can you clock the new chip with the board? is it going past 4.4 at say 1.35 volts? and just remember that every single sandy will clock really well when its brand new and in a few week once it breaks in it will need a little more volts.

like if its stable at 1.27 now you will need about 1.315-1.320 in a few months if you keep benching it.

My 2600k did 5ghx at 1.44 and now it needs 1.5-1.520 to be at least 5 passes of full avx stable

 

[repoman0]

Yeah they will take it back for a 15% restock fee. Not bad since I would end up saving money with that Microcenter board anyway.

Neither chip will do 4.4GHz at 1.35V - actually they both failed at 1.37V and I didn't want to try it higher (just because I won't run it that high 24/7 due to power usage and efficiency).

Now I feel like an idiot for blaming the CPU when it might be the board!

All this for probably a 5% difference when all is said and done!

 

[grkM3]

I was at 4.2-4.3 ghz wall also with my ud7.before you give up on that board make sure its on the latest bios and dont install any of the software for tweaking the board in windows.just do your overclock in the bios and reboot.

You are safe up to 1.395 volts or so just dont stress test it a lot.your cpu in normal usage will never see that much load and will never get close to tye temps intel burn test gets.

Im also kinda local to you as im in haverhill and could run that 2500k on my mive board to show you how high it can clock.heck it was doing 4.6ghz for months without a single crash before i upgraded to a 2600k

 

[Rvenger]

repo, what board do you have?

Don't go with the low end Asus. If you want a cheap but good overclocking board get a Biostar TZ68a+, TZ68k+ or Asrock Extreme 3 Gen3.


http://forums.anandtech.com/showthread.php?t=2174468&highlight=

 

[repoman0]

repo, what board do you have?

Don't go with the low end Asus. If you want a cheap but good overclocking board get a Biostar TZ68a+, TZ68k+ or Asrock Extreme 3 Gen3.


http://forums.anandtech.com/showthread.php?t=2174468&highlight=

I ended up with the 70 dollar Asus - I checked out all of those and the only one that microcenter carried (tz68a+) was really lacking in the rear I/o department. Only 4 USB ports, I need more than that. Anyway windows booted right up after the swap and installed all drivers, currently testing the new chip at 4.2ghz with 1.24v full load according to hwmonitor and it seems stable. Llc is off so maybe it will help lower it. So 4.5-4.7 should be doable safely now not bad for half the price of my other board!

 

[grkM3]

I ended up with the 70 dollar Asus - I checked out all of those and the only one that microcenter carried (tz68a+) was really lacking in the rear I/o department. Only 4 USB ports, I need more than that. Anyway windows booted right up after the swap and installed all drivers, currently testing the new chip at 4.2ghz with 1.24v full load according to hwmonitor and it seems stable. Llc is off so maybe it will help lower it. So 4.5-4.7 should be doable safely now not bad for half the price of my other board!

isnt 1.24 volts even less than what the cpu will call for even at stock setting once turbo is on?

Try my old chip and run it at 1.370 and boot it at a 46x multy.It should be 1000% stable and dont worry about avx stress testing as you will never load the chip like that in real life and it should still pass at least 5 passes set to high stress.

and just another note to make on running long hours of intel burn test.You are straining the board just as much and if not even more with the cheap VR they put on the lower end boards.If you are stable for 10 min but then it locks up its the board and be carefull as there have been some boards buring up running intel burn test for too long.

just for fun feel the MB sinks and VRs when you run burn test to see how hot they get.