|
fargom posted:I want to give a shoutout to my champion I5-750 2.6 ghz ancient CPU that has been holding 3.8gz since early 2010. Upgraded to a 6600k and some 3200 DDR4 today, hoping to win the lottery again. I'm air cooling and i got 4.6 stable at 75c and my only exhaust fan is broken but I'm using the Phanteks PH-TC14PE which is a beast. e: I forgot to mention this is in prime95. Otherwise, it doesn't even hit 60c. Regrettable fucked around with this message at 05:43 on Sep 18, 2016 |
#
?
Sep 18, 2016 04:41
|
|
|
|
| # ? Apr 26, 2024 15:04 |
|
|
4.2 is what i got for a stable OC on my 6600K with hyper 212, and as far as I know thats a pretty under average result? either way its thermals are fine and don't seem like the limiting factor, running prime95 its only peaking at 63 but keeps under 60 on avg
|
|
#
?
Sep 18, 2016 05:21
|
|
|
I had my 6600k at 4.7 with voltage 1.330 on air at 80c under stress testing load, 64c on regular load like games and stuff, with a hyper 212 Evo. I switched to an AIO liquid cooler recently and temps are lower but haven't attempted anything higher yet. I think you'll be just fine with air cooling using recommended voltages.
|
|
#
?
Sep 18, 2016 05:23
|
|
|
I think it's safe to expect the average 6600K to do 4.4 to 4.5 GHz on budget air cooling. Mine hits 4.4 at 1.3V (~1.26V under load due to no LLC support on my motherboard) and thermals are no problem using an Arctic Freezer i32, which performs similarly to a 212 EVO.
|
|
#
?
Sep 18, 2016 08:16
|
|
|
Hooray for weird instability bullshit! I've been sitting on what seems to be a nice stable OC at 4.6GHz with a vcore of 1.3V based on general use and an extended stress test with AIDA64. On the advice of someone in the PC building megathread, I double-checked my stability with 10 runs of ITB on max RAM. It failed on the third run, so I figured I'd bump up my vcore to 1.35. This seemed fine until I started playing Assassin's Creed 3 and watched the game become instantly CPU-limited. tl;dr gently caress you Intel Burn Test, I do what I want.
|
|
#
?
Sep 18, 2016 08:30
|
|
|
LogicalFallacy posted:Hooray for weird instability bullshit! I've been sitting on what seems to be a nice stable OC at 4.6GHz with a vcore of 1.3V based on general use and an extended stress test with AIDA64. On the advice of someone in the PC building megathread, I double-checked my stability with 10 runs of ITB on max RAM. It failed on the third run, so I figured I'd bump up my vcore to 1.35. This seemed fine until I started playing Assassin's Creed 3 and watched the game become instantly CPU-limited. Intel burn test won't get you a completely stable overclock either, obviously. You want to use it for a short initial stability test but then use prime95 for at least a few hours and after you pass that you'll still have to up your vcore by 10-25mv to ensure stability in everything.
|
|
#
?
Sep 18, 2016 22:58
|
|
|
LogicalFallacy posted:tl;dr gently caress you Intel Burn Test, I do what I want.
|
|
#
?
Sep 18, 2016 23:41
|
|
|
e: nevermind
|
|
#
?
Sep 19, 2016 00:13
|
|
|
Ninkobei posted:GPU OC question - The Sapphire Trixx 6.0 software doesn't let me use voltage control on my Sapphire Fury. MSI Afterburner works just fine though - is there a trick to Trixx to unlock that setting? Is yours the OC that has a switch on the card to toggle between the two firmwares, one of which has a higher power limit? Edit: I've been using 5.2.1 w/o any problems
|
|
#
?
Sep 19, 2016 00:57
|
|
|
I LIKE TO SMOKE WEE posted:Is yours the OC that has a switch on the card to toggle between the two firmwares, one of which has a higher power limit? Yep! Using the blue light special bios. I guess the new 6.0+ are targeted at the RX folks and their fancy LEDs
|
|
#
?
Sep 19, 2016 12:44
|
|
|
AEMINAL posted:I've got my 6600K at a 1.35 vcore and 4.4 GHZ OC, how much can i push the vcore if i want to try going for 4.5 or 4.6 GHZ? Honestly, I'd leave it there. Skylake hasn't been out long enough to set definitive limits, but general consensus is never to go above 1.4V for long-term use. Personally, I tried to stay significantly below it with my OC. 1.35 is probably a good place to stop unless you can get to 4.5 GHz with minimal extra voltage.
|
|
#
?
Sep 20, 2016 16:01
|
|
|
What is the real-world performance difference between 4.4 GHz and 4.3? I could not get my i5-6600k to be stable at 4.4 unless I put the voltage to 1.4v, and then the temperatures were spiking to 95-99 under load. I'm using a Cryorig m9i in a Thermaltake V1, btw. Now I'm running at 4.3 at 1.2v and it's also much cooler (75-80 under load)
|
|
#
?
Sep 20, 2016 16:20
|
|
|
CharlieFoxtrot posted:What is the real-world performance difference between 4.4 GHz and 4.3? I could not get my i5-6600k to be stable at 4.4 unless I put the voltage to 1.4v, and then the temperatures were spiking to 95-99 under load. I'm using a Cryorig m9i in a Thermaltake V1, btw. Not much, find the sweet spot for your CPU and don't cry over the last 100MHz.
|
|
#
?
Sep 20, 2016 16:24
|
|
|
CharlieFoxtrot posted:What is the real-world performance difference between 4.4 GHz and 4.3? I could not get my i5-6600k to be stable at 4.4 unless I put the voltage to 1.4v, and then the temperatures were spiking to 95-99 under load. I'm using a Cryorig m9i in a Thermaltake V1, btw. 2.33%
|
|
#
?
Sep 20, 2016 16:26
|
|
|
CharlieFoxtrot posted:What is the real-world performance difference between 4.4 GHz and 4.3? I could not get my i5-6600k to be stable at 4.4 unless I put the voltage to 1.4v, and then the temperatures were spiking to 95-99 under load. I'm using a Cryorig m9i in a Thermaltake V1, btw. Almost nothing, don't sweat it, the longevity of your processor is much more important than that tiny bit of extra performance.
|
|
#
?
Sep 20, 2016 16:28
|
|
|
CharlieFoxtrot posted:What is the real-world performance difference between 4.4 GHz and 4.3? I could not get my i5-6600k to be stable at 4.4 unless I put the voltage to 1.4v, and then the temperatures were spiking to 95-99 under load. I'm using a Cryorig m9i in a Thermaltake V1, btw. CPUs often have different sweet spots. I guess you've found your limit, so I'd definitely stick with 4.3.
|
|
#
?
Sep 20, 2016 16:38
|
|
|
I only got my 5820K to 4.13 GHz at 1.3V Vcore with everything else at defaults/auto. Any point to messing around with uncore or memory timings or anything like that to get it a little farther? Is it safe/productive to push up to say 1.35V on Haswell-E? My idle power consumption went way up too. Is there any way to make it so that it'll only step up the voltage when it advances the multiplier? Ideally it would be nice to have a separate Vcore for each multiplier setting...
|
|
#
?
Sep 20, 2016 22:29
|
|
|
Paul MaudDib posted:I only got my 5820K to 4.13 GHz at 1.3V Vcore with everything else at defaults/auto. Any point to messing around with uncore or memory timings or anything like that to get it a little farther? Is it safe/productive to push up to say 1.35V on Haswell-E? My motherboard had an options to set a base vcore and then add a positive offset. So base vcore is 1.3 + 0.035. It still drops the voltage below 1 when its not loaded.
|
|
#
?
Sep 20, 2016 22:45
|
|
|
B-Mac posted:My motherboard had an options to set a base vcore and then add a positive offset. So base vcore is 1.3 + 0.035. It still drops the voltage below 1 when its not loaded. There's a setting called Load-Line Calibration that when enabled should give you a constant voltage. there's controversy I guess, but basically as long as you don't pump too many volts through your CPU it will be fine.
|
|
#
?
Sep 20, 2016 23:02
|
|
|
Ninkobei posted:There's a setting called Load-Line Calibration that when enabled should give you a constant voltage. there's controversy I guess, but basically as long as you don't pump too many volts through your CPU it will be fine. Yeah, it's to help stabilize it without having to raise vcore higher to compensate for spikes and you should use it in conjunction with an offset. Paul MaudDib posted:I only got my 5820K to 4.13 GHz at 1.3V Vcore with everything else at defaults/auto. Any point to messing around with uncore or memory timings or anything like that to get it a little farther? Is it safe/productive to push up to say 1.35V on Haswell-E? I personally wouldn't mess with memory timings unless you're trying to go beyond the xmp profile, but you'll probably want to set the VCCSA to make sure you're getting the proper timings and you'll have to increase the VCCIO along with it. Regrettable fucked around with this message at 04:10 on Sep 21, 2016 |
|
#
?
Sep 21, 2016 04:04
|
|
|
So I just bought new DDR3 2133 RAM. I can't seem to get it to run at 2133 though, 1600 is as fast as it would get and still boot. Is there anything special I need to do in BIOS besides enabling XMP? I tried that, and setting the timing manually and neither seemed to work. My MOBO is this and 2133 is supposedly the max speed it can handle: http://www.newegg.com/Product/Product.aspx?Item=N82E16813128487 It's still a big upgrade to go from 8 GB of 2x2 and 1x4 GB of different brands (whoops) to 2x8 GB of the same stuff, but I'd like to get what it's actually rated for from it if possible. Edit: found out I may have put the ram in the wrong slots (put them right next to each other, manual says something about spacing them different, some dual channel jargon) so I'll give that a shot and play with the settings again. Suburban Dad fucked around with this message at 14:54 on Sep 23, 2016 |
|
#
?
Sep 23, 2016 13:37
|
|
|
Larrymer posted:So I just bought new DDR3 2133 RAM. I can't seem to get it to run at 2133 though, 1600 is as fast as it would get and still boot. Is there anything special I need to do in BIOS besides enabling XMP? I tried that, and setting the timing manually and neither seemed to work. My MOBO is this and 2133 is supposedly the max speed it can handle: Are you sure your cpu can handle 2133? Also, the online manuals I checked say they go next to each other like a lot of Gigabyte motherboards, but one says they should go in 2 and 4 and another says 1 and 3. e: 2 and 4 - https://www.manualslib.com/manual/685508/Gigabyte-Ga-P67a-D3-B3.html?page=16 1 and 3 - https://www.manualowl.com/m/Gigabyte/GA-P67A-D3-B3/Manual/207424 (page 16) 
Regrettable fucked around with this message at 05:40 on Sep 24, 2016 |
|
#
?
Sep 24, 2016 05:11
|
|
|
Regrettable posted:Are you sure your cpu can handle 2133? Also, the online manuals I checked say they go next to each other like a lot of Gigabyte motherboards, but one says they should go in 2 and 4 and another says 1 and 3. Thanks for the reply. I initially had them in 2 and 4, but moved them to 1 and 2. Neither way seems to work, I'll try 2 and 4 next. I have an i5 2500k @ 4.4 which lots of people seem to run RAM at this speed, even though the specs on intel's site don't list it. It doesn't list 1600 for that matter, either. http://ark.intel.com/products/52210/Intel-Core-i5-2500K-Processor-6M-Cache-up-to-3_70-GHz Here's a couple BIOS screens. 3 screens, bios shows up on my left hence the weird angle.  Timings set to what the RAM is supposed to be. Default on left, manual inputs on right. I didn't touch the "Channel A Advanced Timing Control" values.  XMP enabled and other settings. Won't boot above 1600 still, tried 18XX and 2133 (only other options above 1600). The channel and rank settings didn't seem to have an effect either way on getting it to boot so I left them on.  Some voltages that I don't think have anything to do with this, I didn't see a spot where I could input the voltage but the mobo detected the sticks at 1.6 (which it is). 
|
|
#
?
Sep 29, 2016 17:44
|
|
|
Larrymer posted:Some voltages that I don't think have anything to do with this, I didn't see a spot where I could input the voltage but the mobo detected the sticks at 1.6 (which it is). Last image, try manually setting the DRAM Voltage to 1.6 and see if that helps. (I hate auto voltage)
|
|
#
?
Sep 29, 2016 18:39
|
|
|
Ninkobei posted:Last image, try manually setting the DRAM Voltage to 1.6 and see if that helps. (I hate auto voltage) That didn't work, neither did switching the slots up for the RAM. I registered for the gskill forum and hopefully they can help. I guess it can be a bitch on older motherboards and might take some tweaking of the advanced timings. I don't think I'll buy a gigabyte board again though, it has few and cryptic options for overclocking the cpu and RAM both and not much documented from what I can tell. Granted it's 5 years and a few generations old but still.
|
|
#
?
Sep 29, 2016 19:20
|
|
|
Have you tried to update the BIOS for the motherboard? You can find the latest one here: http://www.gigabyte.com/products/product-page.aspx?pid=3800#bios Sometimes later BIOS revisions improve RAM compatibility, although it's still a crapshoot. ColTim fucked around with this message at 20:18 on Sep 29, 2016 |
|
#
?
Sep 29, 2016 20:16
|
|
|
Yep, running F7. :/ Edit: FINALLY, got it to work. I had to manually input the tRFC value that I grabbed from HWinfo64 and that seemed to do it. I also put the voltage to 1.6 which didn't do it on it's own, but maybe it was a combo of the two. I don't want to try and tweak the timings or CAS or any of that poo poo, I can finally just let it be. Suburban Dad fucked around with this message at 02:20 on Sep 30, 2016 |
|
#
?
Sep 29, 2016 20:19
|
|
|
So apparently my 4670K really needed a re-paste badly, I decided that since I was cleaning every other thing in my place that I'd clean out the PC too. Originally had the CPU clocked to 4.2Ghz, I had it stable at 4.3 for a short time but weirdly saw temperatures jumping into the low 90's when running IBT even though I have a 240mm dual fan CLC. I had already de-lidded my CPU about a year after getting it and re-pasted at that time, but didn't think much and figured if 4.2 was my max, that was alright with me. After cleaning up the CPU and block and reapplying paste on the die and IHS, my temps are now barely hitting mid 70s under load in IBT in the exact same stress tests. Also was able to bump up to 4.4Ghz stable @ 1.32v, not a huge increase but nice to get a ~30% boost overall with better temps and stability. Maybe I'll see if I can push a little further into the 4.5-4.6 range... 
|
|
#
?
Oct 2, 2016 06:15
|
|
|
I noticed with standard paste that would happen with my delid. After I moved to liquid ultra metal on the die it lasted forever.
|
|
#
?
Oct 2, 2016 15:33
|
|
|
i was convinced to relook at my OC and try and get better results even at 1.375 volts i cant get my 6600k stable at 4.4, the second core just fails instantly on prime95, but the other three seem to chug along fine looking at thermals they seem fine, theres a very rare spike into low 70s on one core, but it keeps 50s-60s is it safe to push voltage higher? is there something I could be missing? or did i really just get the bad silicon lottery draw
|
|
#
?
Oct 23, 2016 00:16
|
|
|
RME posted:i was convinced to relook at my OC and try and get better results What motherboard do you have? There are other settings that might be the culprit here. 4.4 is really quite low for skylake. However, 1.375v is no longer "too high" either, so there is that. But for some reference my 6600k gets 4.6 ghz dead stable at 1.325 vcore and I'm certain that's a little too much voltage. That also runs at 90+*C with Prime95 because its just using a tiny AR06 cooler (for a mitx RVZ02) First thing I'd do is set LLC (load line calibration) to High or Max or Extreme and see if anything changes If you're temperatures are that low I'd just keep cranking up the voltage. Also I'd make sure your vcore settings are actually sticking, those temps + max OC are a little suspect and perhaps your mobo is not applying settings (or attempting to autovolt). Someone else will have to chime in on voltage monitoring software because it seems to be all over the place with Skylake but I believe CoreTemp showed me the most accurate numbers
|
|
#
?
Oct 23, 2016 02:43
|
|
|
Asrock z170 pro4S, so def more on the budget side of things and maybe thats the culprit in theory i could keep exactly core 2 at a lower clock because it does allow setting the multiplier for each individual core but i dont know how advisable that even is I havent touched any LLC settings, but im kind of unclear how they work and ive been using HWMonitor to check on temps and volts
|
|
#
?
Oct 23, 2016 03:36
|
|
|
RME posted:Asrock z170 pro4S, so def more on the budget side of things and maybe thats the culprit When a cpu is put under load, the voltage drops slightly, actually by design (vdroop) to make sure when the cpu comes off load, the voltage doesn't spike too high. However, modern boards can often switch fast enough, so sometimes vdroop simply causes instability if your overclock is right on the edge of being stable. Load line calibration is a setting designed to prevent vdroop under load. It's often better for the system to bump that setting a bit instead of generally raising voltage, but as with all things, your mileage may vary.
|
|
#
?
Oct 23, 2016 11:40
|
|
|
RME posted:i was convinced to relook at my OC and try and get better results Probably just bad luck at silicon lottery. I had same thing happen with my i5-6600k, needed 1.35V for 4.2ghz stable. I exchanged it for i7-6700k that does 4.5Ghz at 1.28V
|
|
#
?
Oct 23, 2016 23:05
|
|
|
I'm getting an i5 4690k to replace my g3258 and I had a question about overclocking + RAM When it comes to RAM, can the i5 even use bigger speeds, or does it come down to what the motherboard supports? And does it matter what speed I pick? Like would DDR3 2400 be that much better than 1600? I have old 1067 (I think, its about that speed anyway) ram and was planning on upgrading anyway. EDIT: I'm asking because the difference is $5 between 1600 and 2400 RAM. Lord Windy fucked around with this message at 11:21 on Oct 24, 2016 |
|
#
?
Oct 24, 2016 11:05
|
|
|
Lord Windy posted:I'm getting an i5 4690k to replace my g3258 and I had a question about overclocking + RAM Short answer: Almost unnoticeable boost, even more diminishing returns after 2133.
|
|
#
?
Oct 24, 2016 11:29
|
|
|
Do I have to touch anything to make 2400 work? The board supports it and for $5 I may as well just get the faster one.
|
|
#
?
Oct 24, 2016 11:29
|
|
|
If your board has it , enable XMP in the BIOS's RAM settings.
|
|
#
?
Oct 24, 2016 11:38
|
|
|
Sidesaddle Cavalry posted:If your board has it , enable XMP in the BIOS's RAM settings. Awesome, thanks. That is easy to do and my board supports it. (Z97 Anniversary Edition)
|
|
#
?
Oct 24, 2016 11:41
|
|
|
|
| # ? Apr 26, 2024 15:04 |
|
|
If your motherboard isn't stupid, that is. That's all you should have to do but if you scroll up you can see my issues with RAM and having to manually adjust some settings to get it to work.
|
|
#
?
Oct 24, 2016 12:19
|
|