Closed Solved

OC i5-3750k Q's

Hi all,

Well, my i5-2500k decided to fail on me after a year and a half of use. I'm not having a lot of luck with parts lately (my Asus P8P67 Pro MoBo failed in Dec). I got the i5-3570k as a replacement as I need my comp running and couldn't wait for Intel's response for RMA. The i5-3570k was on sale and ended up being more cheaper than the i5-2500k, hence my decision to get it.

I know with SB to keep vcore around 1.2v and temps below 65-70'c. I am assuming this is the same for IB?

I also notice that this chip I have will not clock down to 1600 MHz when idle. I already reset CMOS and left everything stock, but it will still be stuck at turbo speeds. Any idea of what might be causing this as I didn't have this issue with the i5-2500k.

TIA.
44 answers Last reply Best Answer
More about 3750k
  1. the energy saving mode is a settings in the bios.

    what cooler are u using. also theres a overclocking section in toms could help me beyond my limit oc knowledge
  2. Sorry, thought I was in the OC forum when posting this.

    Using a Scythe Mugen-2 Rev.B cooler.

    Mods: please move to appropriate forum. Thanks.
  3. are u overclocking?
  4. Not yet.

    I plan to OC this to around 4.2GHz and hopefully still have the vcore well below 1.2v. I know IB runs hotter than SB and don't want to push it too hard. I'm just trying to figure out if IB's happy medium is to keep its vcore around 1.2v and temps below 65-70'c (same as SB) as a baseline.

    I figured out why my IB chip wasn't down clocking. Forgot to set the min processor state to 5% under performance power management.
  5. neograndizer said:
    Not yet.

    I plan to OC this to around 4.2GHz and hopefully still have the vcore well below 1.2v. I know IB runs hotter than SB and don't want to push it too hard. I'm just trying to figure out if IB's happy medium is to keep its vcore around 1.2v and temps below 65-70'c (same as SB) as a baseline.

    I figured out why my IB chip wasn't down clocking. Forgot to set the min processor state to 5% under performance power management.


    ivy Bridge is very different form Sandy in more than just heat. For one, Ivy Bridge does not like voltage as much.

    http://www.thinkcomputers.org/intel-ivy-bridge-overclocking-guide/

    There is a guide. You should be able to get 4.2-4.4GHz on stock 1.2v.

    As for temps, during stress-testing its normal for a 3570K to hit 92C, keep it under 100C in Intel Burn test and you are safe. Absolute max before dmg is done to the chip is 103C, but at that temp the chip will shut itself down automatically anyway.
  6. you killed your sandy?! that doesnt happen often

    how do you think it happened? 1.2 volts is very low for a sandy btw. ive been running mine at 1.41-1.44 for 2 years now without a hiccup
  7. Yeah, i5-2500k memory controller failed. Thought it was RAM or the SSD, but turned out to be the CPU. Got a lot of read/write access denied errors as well as memory faults.

    Just testing my OC on the i5-3750k with vcore set at offset -0.005v @ 42x 100 BCLK. P95 Blend at 1 hour and temps for 1 core maxed at 74'C. Vcore peaked at 1.128v. No errors yet, so it looks good. Will be using IBT soon afterwards.
  8. neograndizer said:
    Yeah, i5-2500k memory controller failed. Thought it was RAM or the SSD, but turned out to be the CPU. Got a lot of read/write access denied errors as well as memory faults.

    Just testing my OC on the i5-3750k with vcore set at offset -0.005v @ 42x 100 BCLK. P95 Blend at 1 hour and temps for 1 core maxed at 74'C. Vcore peaked at 1.128v. No errors yet, so it looks good. Will be using IBT soon afterwards.


    Seems a little hot, what cooler are you using.
  9. 1.65v or 1.5v ram?
  10. honestly i think temps are fine.. because Prime95 is the only time you will see temps that high. the average game like BF3 isn't gonna fully max his CPU like Prime does
  11. I am using a Scythe Mugen-2 Rev.B. My SB chip at 43x was going for 65'C max. IB at 42x is going for 74'C max. I guess they weren't kidding around when they say these chips run hotter. I'll probably tune it down to 40x to keep temps low.

    The RAM is rated at 1.6v, but I undervolted to 1.55v and tested for greater than 24 hours Memtest86+ with no errors on my previous OC with SB. Timings 7-8-7-24 2T @ 1600MHz.

    Edit: 2x4 GB RAM sticks
  12. if you havent checked it out, you should read some ivybridge delidding threads. theyre pretty awesome if you are into that sort of thing.

    even if you never work up the courage to do it to your own (its risky) it is still neat to know what sort of performance you could have.
  13. ud better better performance from a dh-14 from noctua ivy does heat more than sandy brig.
  14. Best answer
    neograndizer said:
    I am using a Scythe Mugen-2 Rev.B. My SB chip at 43x was going for 65'C max. IB at 42x is going for 74'C max. I guess they weren't kidding around when they say these chips run hotter. I'll probably tune it down to 40x to keep temps low.

    The RAM is rated at 1.6v, but I undervolted to 1.55v and tested for greater than 24 hours Memtest86+ with no errors on my previous OC with SB. Timings 7-8-7-24 2T @ 1600MHz.

    Edit: 2x4 GB RAM sticks

    Ivy is hotter because the IHS(cover directly above the CPU die that transfers heat) is not soldered on like Sandy, it used a type of thermal paste to conduct the heat. Intel did this to conserve costs.

    You may want to get your RAM to 1.5v as that is safest on Ivy, as they are quite sensitive to higher voltage RAM. 1.5v is suggested for Ivy and in many cases the chips lose lifespan when used with higher voltage RAM.
  15. yes intel specifications are 1600mhz and 1.5volts max for ram.é
  16. Novuake said:
    Ivy is hotter because the IHS(cover directly above the CPU die that transfers heat) is not soldered on like Sandy, it used a type of thermal paste to conduct the heat. Intel did this to conserve costs.

    You may want to get your RAM to 1.5v as that is safest on Ivy, as they are quite sensitive to higher voltage RAM. 1.5v is suggested for Ivy and in many cases the chips lose lifespan when used with higher voltage RAM.

    That was revised. they are soldered now like sandy
  17. ^this

    ram volted higher than 1.5v is the number one killer of sandys and ivys.

    considering you already killed a sandy memory controller, you might want to think about using different ram.

    edit: new ivys are soldered like sandys? so, are you saying a new ivy runs cooler than an older one?
  18. Rockdpm said:
    That was revised. they are soldered now like sandy


    SINCE WHEN??? Source please!
  19. It was when they did the revisions... also i get my 3770K no later than tuesday so i could check and see but pretty sure they switched out from thermal compound to soldered. and even if they still use compound... its not as hot as the originals. the originals were dangerous
  20. source!
  21. Rockdpm said:
    It was when they did the revisions... also i get my 3770K no later than tuesday so i could check and see but pretty sure they switched out from thermal compound to soldered. and even if they still use compound... its not as hot as the originals. the originals were dangerous


    Not good enough, source???

    You can't just say stuff like that! :pt1cable:
  22. what source. theres only one source. and its a cosmic gamma ray shooting waterfall. of course
  23. Do i look like I am someone who works for Intel? Ask the friggin Intel rep for toms he will tell you if you wanna know so bad
  24. no u look like a super spacecat man
  25. oh thats nice.. wouldn't care if you never asked me again.. just kinda don't care
  26. i fink yer as high as the cat in your avatar
  27. no i was refurring to the cat in ur avatar lol.

    why so serious?
  28. Ok thats cool.. you guys talking like you can't understand english really makes you look smart.. good day
  29. we arent the ones making unsubstantiated, and frankly false, claims. understand that.
  30. lol, we spaek guud english, totally bro.
  31. Rockdpm said:
    Ok thats cool.. you guys talking like you can't understand english really makes you look smart.. good day


    Nothing to do with your english. We go by one rule. Screenshot or it NEVER happened. This applies to news and conjecture as well. WE WANT PROOF!
  32. rofl, are english is quite good, even novuake english is better than yours and hes from africa.

    without proof its just a rumour or a word of mouth, show us the pudding!
  33. ^this

    ram volted higher than 1.5v is the number one killer of sandys and ivys.

    Thanks. I'll probably bring the timings to CAS 8 with 1.5v and see how stable it'll be. Didn't think the DIMM voltage would have killed SB that quickly though.

    Seeing how hot 42x was, I'll be bringing IB down to 40x and retest it after playing with the RAM timings and voltage.
  34. well intels specs are 1600mhz and 1.5volts id stay within that confine.
  35. Out of curiosity, how reliable is Memtest86+? I've currently set RAM to 8-8-8-24 2T @ 1.475v. I finished 1 pass with no errors. I'm going to see if I can get it to 1.45v at current timings or tweak some more. Might tighten timings back to 7-8-7-24 2T and undervolt to less than 1.5v if I'm lucky.
  36. memtest is pretty accurate.
  37. Thanks, ran 10 passes using 8-8-8-24 2T @ 1.45v. No errors, so I think I'll leave my RAM as it is. If I encounter any errors while running OS, I'll just bump voltage by a small bit and retest.

    Now going to bring my OC down to maybe 40/41x and see how vcore and temps go. If P95 temps are at 65'C or below after 20-30 mins, I'll be quite happy with the results and stress test for further stability.
  38. sounds not bad :)
  39. Yeah, IB is definitely a hot chip. My Scythe cooler was able to manage a modest clock (43x) with low temps on SB (max 65'C P95 24 hour run and 67'C under IBT maximum stress testing with 10 passes). But it can't seem to handle IB well. At 40x running P95 blend for 11 hours, temps peaked at 70'C. Running IBT at maximum with 10 passes peaked temps at 73'C. Voltage also being significantly lower with the IB chip.

    Both stress tests showed no errors, so the OC is mostly stable. I think the temps for stress testing are good (not great) and vcore set at offset -0.01v.

    I'll leave my OC as it is. Now have to see if the offset will play nice when coming out of idling. Realistically speaking, I don't think I'll do any better than what I've currently set (unless I plan on increasing the offset - further).

    Hopefully, I won't screw up this chip like I did my i5-2500k. LOL. Live and learn...
  40. Novuake said:
    Ivy is hotter because the IHS(cover directly above the CPU die that transfers heat) is not soldered on like Sandy, it used a type of thermal paste to conduct the heat. Intel did this to conserve costs.


    Regarding this, I think the heat is being retained by the smaller transistors and not being able to efficiently dissipate/transfer the heat to the heat sink. Probably due to it's smaller die/surface area. I don't think soldering would have helped in this situation either.
  41. Best answer selected by neograndizer.
  42. delidding an ivy drops the temp by 10 degrees or more.

    check out some ivybridge delidding threads, theyre neat
  43. neon neophyte said:
    delidding an ivy drops the temp by 10 degrees or more.

    check out some ivybridge delidding threads, theyre neat


    Yeah, will be looking at that soon enough now that I've finished testing the system. Definitely sounds like an interesting topic. :)
Ask a new question

Read More

CPUs Intel i5