|
United Kingdom20285 Posts
Heya, i can edit some more stuff into OP later~
I wanted to make a thread to discuss haswell overclocking, performance, temperatures, etc. In particular, the weird really high temperatures with AVX instructions enabled. I spent some time putting some numbers together:
ALWAYS THE SAME:
Gigabyte z87x-ud3h
4.5ghz core and uncore
2000mhz 10-10-10-28 RAM
1.2v uncore
1.22vcore, +0.15 analog io, +0.15 digital io, +0.1 system agent, default llc < volts need some work
~20c ambients open case because exhaust fan broken, xigmatek utgard, 120mm front intake
170mm top exhaust, silver arrow SE two ty-147 fans @1200rpm
6 minutes IBT temp, 4gb RAM:
HT on, avx on - 132 gflops - Max temps 92, 93, 91, 85*****
HT on, avx off - 67.2 gflops - Max temps 73, 74, 72, 66
HT off, avx on - 129 gflops - Max temps 83, 84, 82, 79
HT off, avx off - 67.6 gflops - Max temps 67, 67, 67, 61
x264 Benchmark HD v5.0.1
HT on, avx on - Pass 1; 88.12fps - Pass 2; 19.65fps - Max temps 71, 71, 70, 66
HT on, avx off - Pass 1; 88.08fps - Pass 2; 19.36fps - Max temps 70, 71, 70, 64
HT off, avx on - Pass 1; 70.98fps - Pass 2; 17.30fps - Max temps 63, 63, 62, 58
HT off, avx off - Pass 1; 67.39fps - Pass 2; 17.29fps - Max temps 63, 63, 62, 60 ???*
Starcraft 2: HOTS, maxed physics and effects, reflections on. Med shaders etc. 5 min replay segment, x4 speed
HT on, avx on - 2013-06-13 13:32:37 - SC2
Frames: 4022 - Time: 53290ms - Avg: 75.474 - Min: 35 - Max: 114
HT off, avx off - 2013-06-13 13:58:14 - SC2
Frames: 4032 - Time: 53368ms - Avg: 75.551 - Min: 35 - Max: 113
Note; CPU temps ridiculously low when playing sc2. Three cores dipping below 40c - one around 50-53c with ht AND avx.
No gain from HT and lack of load on other 3 cores = cool as shit fully loaded core
*Ambients varying within reason; no temperature control
*****STOP 0x0101 on third run, possible thermal failure?
Discussion here; It seems pretty ridiculous to me, and shocking that most people are tuning their overclocks around stress tests such as IBT and Prime with AVX enabled. That has been cause for much misinformation and confusion about Haswell's overclocking performance; It actually seems pretty damn good. A bit hotter than Ivy perhaps, but people are definitely doing more on lower vcores. I'll be tuning OC around the most stressful loads i expect to be doing; Not around IBT with AVX, which somehow pushes my OC, which sits in the high 60's with CPU at 100% load from x264 encoding, into the 90's and to the point of crashing system with silver arrow on 1.22v.
The sc2 benchmark i used was this replay: http://drop.sc/342789
TGeuSohel vs SeleCT: Medium shaders, maxed physics and effects, reflections on, SeleCT PoV (after selecting an orbital command) played 15 minute to 20 minute at x4 speed (after playing through a few times to remove stutters) and ingame performance is kinda silly.
|
Pretty good work! I'll update my post with some results (although I don't have all night to test like Cyro  ) but my results seem pretty similar. I'm testing to get 4.4Ghz stable, I'll post my benchmarks there and see if I can go for 4.5Ghz sometime soon! Thanks for your work Cyro o:
Also for anybody building a new PC, we can leave detailed 'guides' on how we tuned ours I guess. Not so much guides as saying what settings we changed, since they're different for each motherboard. I have an asrock, and I think cyro has a gigabyte. Just need an Asus
|
United Kingdom20285 Posts
I'm just getting started on tuning. My uncore was set completely blindly, i didn't experiment with digital/analog io or system agent properly yet, and because of some weird behaivour i have hopes that i can get better clocks on lower vcores with them. My vcore is really excessive right now, i added a ton to make sure i wouldn't fail during benching. I've passed IBT max ram with a lot lower. RAM also; i'l try to push to 2400 cas12, since jinglehell's old blog showed scaling with frequency but not timings.
FPS Shown in Min/Max/Average
2.8Ghz CPU Clock
1066 CL9 FPS: 28/133/61.6--95%
1066 CL6 FPS: 28/134/62.6--96.6%
1333 CL9 FPS: 30/139/64.8--100%
1333 CL6 FPS: 30/139/65.2--100.6%
1600 CL9 FPS: 31/142/67.2--103.7%
1600 CL6 FPS: 31/142/67.3--103.8%
|
Bunch of questions time.
Played with Base Clocks yet? If so any difference at all?
45 x 100?
36 x 125?
27 x 166?
Any testing uncore?
Are you delidded?
Hm can't think of more questions at this time.
|
United Kingdom20285 Posts
Bunch of questions time.
Played with Base Clocks yet? If so any difference at all?
45 x 100?
36 x 125?
27 x 166?
I tried 27x167 quickly, but it didn't seem to give better results initially (maybe voltages have to be tuned differently) and there's some issues with changing bclk straps.. RAM multipliers are weird and i couldn't set it how i wanted IIRC, and you have to cut power and mess around in bios for it to accept the bclk strap setting and boot properly, like it would stick on 167mhz even if you set 100x45 in bios and fail to boot over and over again until i just cut power for a few minutes, and i heard that's a common problem
Any testing uncore?
It didn't seem to make much difference in a few cases, i'l try it with sc2.
Are you delidded?
No, maybe in the future, i bought CLU and i was going to at least do some testing, get down overclock first, but i'm having some doubts now - who wants to hit a $430 CPU with a hammer? I could replace it, but then i'd have to face the silicon lottery again (reshuffling the deck when you had a decent hand) and i wouldnt be able to get a graphics card. I'm pretty damn sure that i wouldn't mess it up, but there's always a chance for something silly to go wrong, and that would really suck.
(:
|
I've tried like 40x 112.5 (something like that, it was 4.5Ghz) and had a bunch of problems, so I just went back to normal. I've heard some people having better success at 38x1.25, but I'm going to start with normal stuff and then go back to BCLK changing whenever I'm done finding it with just multipliers.
It seems like the old stress tests are performing terribly for haswell. It wouldbe interesting if they behaved more like x264. Then again, I'm nit an expert so maybe that's expected.
|
United Kingdom20285 Posts
If anybody has a sandy or ivy bridge CPU (oc'd) and a bit of time to spare, it would be VERY interesting to see temperatures with and without avx enabled for IBT. It's pretty easy to do
And Alryk, that's a terrible bclk to use ;p
Your straps are 100mhz, 125, 166.6. You should stick VERY close to those. I wouldn't expect 112.5 bclk to do anything at all
|
Could you grab a recent build of x264 and redo the benches with encoding some test file (@crf whatever)? Just make sure that you're not getting I/O bottlenecked.
Not that there are huge gains from AVX2 (so far), considering where most of the code time actually is and all of the supporting operations (see links here)
http://forum.doom9.org/showthread.php?t=168037
So that means that likewise, temps shouldn't be all that much higher when using it, I think.
|
On June 13 2013 22:57 Cyro wrote:No, maybe in the future, i bought CLU and i was going to at least do some testing, get down overclock first, but i'm having some doubts now - who wants to hit a $430 CPU with a hammer? I could replace it, but then i'd have to face the silicon lottery again (reshuffling the deck when you had a decent hand) and i wouldnt be able to get a graphics card. I'm pretty damn sure that i wouldn't mess it up, but there's always a chance for something silly to go wrong, and that would really suck. (:
You can use the LiquidUltra you bought instead of your current thermal paste. It should do a few degrees C of improvement. You'll get some stains reportedly, but it's supposedly fixed through metal polish and half an hour of rubbing. You'll also only have to do that polishing if you care about a stain, or perhaps when selling the CPU.
Don't use the sponge you get in the CLU package for anything! It's a metal sponge and will scratch stuff easily. It will destroy the printing on the CPU and the mirror finish of the heat-sink's base. I think it's only in the package because it's what's best to clean old LiquidPro.
|
United Kingdom20285 Posts
I bought it to use under the IHS, not above, i have mx-4 and didn't really want to deal it above IHS unless gains were big. I probably wouldn't do it unless maybe if i delidded, for extra temp drop, and because CPU is "messed up" anyway
Thanks for the info
|
I also was pretty interested in what happens on Ivy Bridge with the temperatures and AVX and IBT, so I checked what happens with my i5.
I'm sitting at an i5-3570k at 4.6 GHz, offset VCore set to result in about 1.255 V, LLC "High" which is a tiny drop, C1E and EIST both on and Windows power profile is "balanced", RAM at 2000 MHz.
The numbers I see for temperature with AVX on or off are not very different compared to what you've shown for HyperThreading off.
3570k @ 4.6 GHz:
AVX on: 125 GFlops -- max temp 75, 84, 82, 80 C
AVX off: 67 GFlops -- max temp 67, 73, 71, 71 C
Here's what you showed in the OP:
On June 13 2013 22:16 Cyro wrote:
6 minutes IBT temp, 4gb RAM:
HT off, avx on - 129 gflops - Max temps 83, 84, 82, 79
HT off, avx off - 67.6 gflops - Max temps 67, 67, 67, 61
I turned AVX off and on with "bcdedit /set xsavedisable 1" and "bcdedit /deletevalue xsavedisable" (and rebooting).
EDIT: screenshots:
AVX on: http://i.imgur.com/kqyZnIi.png
AVX off: http://i.imgur.com/o2M1KFz.png
|
On June 13 2013 23:08 Cyro wrote:
If anybody has a sandy or ivy bridge CPU (oc'd) and a bit of time to spare, it would be VERY interesting to see temperatures with and without avx enabled for IBT. It's pretty easy to do
And Alryk, that's a terrible bclk to use ;p
Your straps are 100mhz, 125, 166.6. You should stick VERY close to those. I wouldn't expect 112.5 bclk to do anything at all
Blame the OCN post xD but that explains it. At least I don't pretend to be an expert 
|
United Kingdom20285 Posts
On June 14 2013 00:42 Ropid wrote:I also was pretty interested in what happens on Ivy Bridge with the temperatures and AVX and IBT, so I checked what happens with my i5. I'm sitting at an i5-3570k at 4.6 GHz, offset VCore set to result in about 1.255 V, LLC "High" which is a tiny drop, C1E and EIST both on and Windows power profile is "balanced", RAM at 2000 MHz. The numbers I see for temperature with AVX on or off are not very different compared to what you've shown for HyperThreading off. 3570k @ 4.6 GHz: AVX on: 125 GFlops -- max temp 75, 84, 82, 80 C AVX off: 67 GFlops -- max temp 67, 73, 71, 71 C Here's what you showed in the OP: Show nested quote +On June 13 2013 22:16 Cyro wrote:
6 minutes IBT temp, 4gb RAM:
HT off, avx on - 129 gflops - Max temps 83, 84, 82, 79
HT off, avx off - 67.6 gflops - Max temps 67, 67, 67, 61
I turned AVX off and on with "bcdedit /set xsavedisable 1" and "bcdedit /deletevalue xsavedisable" (and rebooting). EDIT: screenshots: AVX on: http://i.imgur.com/kqyZnIi.pngAVX off: http://i.imgur.com/o2M1KFz.png
Thanks a ton. Wanted to make sure i wasn't losing my mind on avx temperature gap but it seems like there's some funny buisness going on with Haswell*
*Edit: Maybe this is wrong to say? 17c instead of 11c gap - but the complete lack of being able to reproduce this gap (which is 20c with ht on) makes it a bad idea to tune around it i think
|
On June 13 2013 22:43 Cyro wrote:
2.8Ghz CPU Clock
1333 CL9 FPS: 30/139/64.8--100%
1333 CL6 FPS: 30/139/65.2--100.6%
1600 CL9 FPS: 31/142/67.2--103.7%
1600 CL6 FPS: 31/142/67.3--103.8%
The difference between 1600 at CL9 and CL6 really is small. I have no idea: how often does the CPU query RAM during SC 2? The “total latency” of 1600 at CL9 is higher by a factor of 1.5, when compared to 1600 at CL6. Shouldn’t this add up more clearly?
|
United Kingdom20285 Posts
I dunno, im a RAM newbie (:
I remembered that blog from years ago though
|
Those latencies are just to kick off the data transfer, I think. The transfer itself might be what's taking most time overall.
|
Will you make a list of FPS numbers at different memory settings, Cyro? That would be neat. Haswell might be very different compared to the CPU used in that old blog.
|
On June 14 2013 01:35 Ropid wrote:
Those latencies are just to kick off the data transfer, I think. The transfer itself might be what's taking most time overall.
Sounds plausible.
|
On June 14 2013 01:21 blueslobster wrote:Show nested quote +On June 13 2013 22:43 Cyro wrote:
2.8Ghz CPU Clock
1333 CL9 FPS: 30/139/64.8--100%
1333 CL6 FPS: 30/139/65.2--100.6%
1600 CL9 FPS: 31/142/67.2--103.7%
1600 CL6 FPS: 31/142/67.3--103.8% The difference between 1600 at CL9 and CL6 really is small. I have no idea: how often does the CPU query RAM during SC 2? The “total latency” of 1600 at CL9 is higher by a factor of 1.5, when compared to 1600 at CL6. Shouldn’t this add up more clearly?
It depends on the task. In gaming, you usually don't need a short latency, as the bottlenecks are nearly always on the GPU and CPU. What may limit is data loading, which is usually kept to 4K+ chunks, usually higher when talking about texture loading. Keep in mind that approximately 98% of the RAM read commands given to the CPU end up reaching the cache, usually L1 and L2, and that the memory controller aggressively pulls bigger pages from the RAM into the cache to avoid this anyways, the latency is less critical for anything that doesn't require random access to big data sets (> 8MB). Good game engine design makes sure that at level loading, the most heavily used data will be cached on the CPU anyways, as this is predictable information. Beyond that there are some very clever hardware tricks in modern CPUs to semi-cache information (TLB, predictive memory controller, etc), which can help keep the latency down to nano-seconds even with CL12.
There is the "memory wall", a phenomena caused by a slower rate of increase in memory speeds vs. CPU speeds over approximately three decades, so modern CPU design does a lot to work around that, and specifically, latency for most common application usage.
tl;dr: for gaming, memory latency doesn't really matter.
|
On June 14 2013 00:42 Ropid wrote:I also was pretty interested in what happens on Ivy Bridge with the temperatures and AVX and IBT, so I checked what happens with my i5. I'm sitting at an i5-3570k at 4.6 GHz, offset VCore set to result in about 1.255 V, LLC "High" which is a tiny drop, C1E and EIST both on and Windows power profile is "balanced", RAM at 2000 MHz. The numbers I see for temperature with AVX on or off are not very different compared to what you've shown for HyperThreading off. 3570k @ 4.6 GHz: AVX on: 125 GFlops -- max temp 75, 84, 82, 80 C AVX off: 67 GFlops -- max temp 67, 73, 71, 71 C Here's what you showed in the OP: Show nested quote +On June 13 2013 22:16 Cyro wrote:
6 minutes IBT temp, 4gb RAM:
HT off, avx on - 129 gflops - Max temps 83, 84, 82, 79
HT off, avx off - 67.6 gflops - Max temps 67, 67, 67, 61
I turned AVX off and on with "bcdedit /set xsavedisable 1" and "bcdedit /deletevalue xsavedisable" (and rebooting). EDIT: screenshots: AVX on: http://i.imgur.com/kqyZnIi.pngAVX off: http://i.imgur.com/o2M1KFz.png
I'm getting similar temps and I'm running 4.7ghz on a 3570k with the same voltage I figure until I get my phase change cooling its the best I'll do with a h60 in push/pull config T_T
|
United Kingdom20285 Posts
Well the h60 is SIGNIFICANTLY worse than the silver arrrow. Linus's google doc puts the silver arrow extreme 20c below the h80i, which is superior.. Even h80i push/pull with nf-f12's, 10c worse than silver arrow, but you probably didnt buy a budget offering and then strap $50 of fans to it so you'd have worse results. You can make large upgrades without spending an insane amount
|
United Kingdom20285 Posts
nsfw?+ Show Spoiler +
As a slight update to this i 0x0124 bluescreened while writing post, put system agent to +0.165v, 0x0124'd while on windows load screen, so put it to +0.18v and doing extended testing soon. Default values might be different for different boards, and i'm sure different CPU's will like different amounts of everything. VRIN, system agent, and i hear analog+digital IO can all have large effects on the vcore you need to be stable, or if you are stable at all at a certain overclock - I've failed 4.5 at 1.21vcore before but i'm seemingly comfortably below 1.16 now which is not a small gap. I had a LOT of trouble stabilizing 4.6 without really big vcore increases, and new knowledge+understanding will probably make it easier
The more i come to understand Haswell, the more it seems like first impressions of myself and pretty much everybody else were wrong.
|
I've been thinking and have you tried turning off speedstep (or whatever it's called these days, but the option to drop down to 1.6ghz and very low V on idle)? I'm thinking that because Haswell can switch vcore ~5x faster than Ivy that it's currently dropping vcore while stress testing in between requests(however fast they may be).
From IDF
![[image loading]](http://images.anandtech.com/reviews/tradeshows/IDF/2005/Fall/Day3/Keynote/coarse.jpg)
![[image loading]](http://images.anandtech.com/reviews/tradeshows/IDF/2005/Fall/Day3/Keynote/fine.jpg)
|
That's something different, I think. It's not about C1E, SpeedStep and other power saving.
When the CPU works on different instructions, it will use more or less parts of the core in processing those instructions. Resistance will change depending on how much of the core is used. If you keep the supplied power constant, this means the voltage drops. See here:
http://en.wikipedia.org/wiki/Ohm#Power_as_a_function_of_resistance
The outside of the CPU will notice the voltage changing, and will react to that by increasing or decreasing the supplied current. The voltage regulator modules (VRM) have to do the things shown in the slide even for completely normal operation. Voltage has to be kept inside the range where the transistors inside the CPU work correctly.
I think the slide is about comparing the integrated VRM of Haswell to old external VRM on the motherboard. It also might not be a comparison with actual motherboards for Ivy Bridge etc., but a comparison with Intel's specification for minimum requirements that might be used in the worst possible motherboards.
|
I think the above slides are relevant and about switching power states. The system agent / uncore / whatever logic for power needs to ask the VRMs (be they on chip or on the motherboard) for different voltages in real time as demands shift. However, if the core is busy, even if stalled, I think it should be at the higher power state and getting the higher voltage, I think, so actually this shouldn't be happening under heavy load.
I don't know much about semiconductor devices, but I'd suggest not even thinking about it in terms of resistance. The behavior of the load is very much not Ohmic (resistive), linear. When you're loading up more processing resources of the CPU, that's causing more transistors to have to switch (wasting more charge stored on gates) as logic gates switch states and more current to flow (more resistive losses in those flows) to make those changes.
The voltage regulators are buck converters, so they're switching, not linear either.
|
|
|
United Kingdom20285 Posts
Yea i disabled speedstep and c-states anyway. It's incredibly frustrating how you can stress test for an hour on 1.158v and then game/browse for hours completely fine, and then later in the day 1.19vcore is chain crashing and you can't reach 2 minutes uptime even without any load. I need to understand more
|
My Ivy Bridge seems much simpler. It's really just Vcore that determines if it's running stable or not. Even with that being the case, I still had those problems with stress tests seemingly running fine, but programs or Windows crashing later with the same Vcore. Voltage needed to be raised a lot to make the PC completely stable.
The amount of that increase in voltage was still always smaller than what a stress test at the next higher multiplier needed. This means if IBT ran for ten minutes at x47 multiplier, the settings used were definitely stable at x46. Perhaps try that if you are a little tired and feel somewhat desperate. Play with stress tests on the next higher multiplier, then use those settings for your current multiplier.
|
United Kingdom20285 Posts
Any assistance?
Perhaps try that if you are a little tired and feel somewhat desperate. Play with stress tests on the next higher multiplier, then use those settings for your current multiplier. I just slept a ton man. I have weeks, months to tune these overclocks, but i need to understand. I've had 45x@1.158 be fine for 3 hours (stress test lightly, then gaming) but i've also had 46x fail at 1.26. It would be ridiculous to overvolt vcore by more than 0.1v just for stability. Actually it seems like the CPU doesnt give a damn what my vcore is, it'll fail anyway, which i need to fix.
I mean it baffles me that i can run IBT for 40 minutes and browse/game for a couple hours again - and then reboot, and fail IBT in 30 seconds with identical settings, reboot again, crash opening firefox, reboot again, down in less than 20 seconds, like damn, what am i doing so seriously wrong?
|
In my notes, it's an 0.05 V increase to get things running perfect when starting with the voltage needed for a short test with IBT.
It was only short tests, I think, but I have this recorded:
IBT:
4.6ghz -> 1.215v
4.7ghz -> 1.265v
4.8ghz -> 1.320v
I'm actually running 4.6 GHz at 1.255 V now. At 1.250 V, WHEA warnings show up in the Windows event log.
Those tests I did, that's really not very stable. I didn't actually record the failures. I just recorded that it has a good chance to pass ten runs of IBT at 1024 MB memory use (or something like that).
|
United Kingdom20285 Posts
|
An update for me and what I'm stable at!
4.4Ghz
44 Multiplier, 100 BCLK @ 1.14V
4.0Ghz Uncore/Cache @ 1.12V
+.15V Digital, System, and Analog IO voltage
1.85VRIN
4.5 Ghz
45 Multiplier, 100 BCLK @ 1.18V
4.0 Ghz Uncore/Cache @ 1.12V
+.15V Digital, System, and Analog IO voltage
1.85 VRIN
Max temps at 4.4: 82, 79, 76, 72
Max temps at 4.5: 89, 87, 84, 80 Less toasty, but still toasty.
Failed
4.5Ghz all settings the same as above but at 1.16V. IBT failed on me, although I didn't BSOD or get any freeze ups.
4.5Ghz at 1.17V, all the same. Same as above - IBT failed me but didn't BSOD or lock up. Testing 1.175V
4.5Ghz at 1.175V failed. See above
I haven't tuned 4.5 yet, while I'm pretty sure 4.4 can't get any lower, I've tried a lot of combinations (mostly from Cyro's suggestions, I've ruled out other things as well. Lower VRIN works much better for me). On the other hand I'll see how low I can get 4.5, and I'll probably test different BCLK's today as well.
Edit 1: 1.18V passed compared to 1.2 original
Edit 2: Failed IBT at 1.16V
Edit 3: Failed IBT at 1.17V, no lockup or BSOD (on fourth test of Very high (4096MB) test)
Edit 4: 1.175V fails on the 5th Very High test. Seems like 1.18V is my stable point.
|
Any advice on how hotI can let my 4670k get before I should start to get worried? I dont want my case to keep sounding like a jet engine (which the damn H100i does at higher levels...)
|
90 celsius stress testing is probably fine. I hit 90 stress testing but SC2 doesn't go over 55 celsius. Stress testing puts an unnecessary amount of temperature when it's using AVX it seems. For example x264 and IBT will both put you at 100% load, but IBT will be 25 celsius hotter when you have AVX on.
Edit above: 1.18V passed at 4.5 Ghz. I'm really surprised as to how close it is to 4.4, I really couldn't seem to get 4.4 any lower.
|
On June 16 2013 03:00 Alryk wrote:
90 celsius stress testing is probably fine. I hit 90 stress testing but SC2 doesn't go over 55 celsius. Stress testing puts an unnecessary amount of temperature when it's using AVX it seems. For example x264 and IBT will both put you at 100% load, but IBT will be 25 celsius hotter when you have AVX on.
Edit above: 1.18V passed at 4.5 Ghz. I'm really surprised as to how close it is to 4.4, I really couldn't seem to get 4.4 any lower.
Ah, thanks, and for later idle and load just when playing games?
|
Anything under 75 load while playing games is fine, if a bit unnecessary. That's totally safe. Idle temps don't really matter though.
Edit 3 above: 1.17V also canceled IBT on me. No lockup. Testing 1.175V!
Edit 4: 1.175V fails on the 5th Very High test. Seems like 1.18V is my stable point.
|
United Kingdom20285 Posts
For an update, i made some wrong assumptions. IBT with avx off doesn't seem to do it's job very well at all, and i wasted a lot of time overanalyzing stuff etc. There was no way i could ever do anything at 1.16vcore - even though IBT probably passed several hours of tests.
I'l get this sc2 benchmark with uncore down at stock (gotta try it multiple times, cause sc2 likes to stutter which throws off average fps a bit and records minimum fps as 0 instead of your actual minimum) and then post where i am at
|
im currently using an antec 1200 case, h100i setup and the Asus ROG VI Hero motherboard.
Im currently running (what i pressume is stable even though prime95 forces it to high 80's low 90's and caused it to crash 19minutes which i dont actually think was because of the 100c ceiling) 4.4ghz with 1.216v (as far as i can tell is excessive and have not tried pushing it down yet). only messed with multiplier since im new to overclocking and manually setting voltage (the auto sets the voltage at 1.3v+ with 4.2ghz 0.o)
Also, the 1.3v+ ran for around 18minutes before i cancelled the prime95 due to it hitting 95c.
I could try with other similar settings, but i will probably just look at seeing either how far i can push the overclock on this voltage, or how far i can drop the voltage with this overclock. I also left the thermal paste that comes with the h100i on instead of using the artic silver 5 that i bought, but i will be removing it soon when i get some cleaning stuff to see if the artic silver is better.
|
United Kingdom20285 Posts
I think i'll write a small guide after a while for OC-ing Haswell
Do you have i5 or i7? Is this with AVX on or off?*
*You can check by running IBT, avx off should give you ~65gflops at 4.5ghz (or a bit more/less depending on ram etc or other things using cpu) and avx on will give like twice as high.
|
United Kingdom20285 Posts
Got bench @4.5ghz, 2000mhz 10-10-10-28 RAM, stock uncore - min FPS the same, average FPS a quarter of a percent lower (3.5ghz uncore vs 4.5), think it's safe to say that uncore probably doesn't really matter for sc2. I'l run RAM at 1066 and see what kind of performance gap there is. I'm kinda amazed at the reproducability of this benchmark (is that a word? :D) i can run it twice on the same settings and get the same result down to tiny tiny fractions.
|
United Kingdom20285 Posts
![[image loading]](http://i.imgur.com/t5zm2Km.png)
Taken from the benchmark i was doing before: 5 min replay segment at med shaders, maxed physics and effects reflections on at x4 speed, but i'm pretty sure the performance would carry over to regular gameplay.
I'l drop my core clock to 3.5ghz and run bench again (:
Frames, Min, Max, Avg
3048, 24, 89, 56.161
i couldn't set 3.5ghz because ud3h bios even though it says x35, 3.5ghz right there after you set it, somehow decides to boot into x25 2.5ghz (gg ud3h), so i benched with 3.6:
3.6 to 4.5ghz on Haswell
25% increase in frequency
37.5% increase in minimums
27% increase in averages
^Both numbers from 1600mhz 10-10-10-28 RAM and stock uncore.
Oh and also:
Running 4.5ghz core stock uncore 2000mhz 10-10-10-28 1.45v 1.05vring 1.86vrin +0.15 digital +0.15 analog +0.1 system agent 1.195 vcore
Through just a ton of painful trial and error it became obvious that IBT does not do its job well at all at least with avx off, and after passing probably what added up to hours of IBT under 1.16v (some 1.155) i had to raise vcore slowly up to 1.195 for 4.5ghz for stability to hold up in system, under high loads, x264 encoding etc, and to make whea erros go away. I have NEVER before seen a CPU pass 20 minutes of IBT max ram, and then reboot with +0.02vcore, and bluescreen multiple times without even reaching desktop - even if it's common to need a bit more voltage than to pass IBT, this was extremely confusing to me. Been rock solid at 1.195 (i initially set 1.2, and then dropped when i had no sign of issues whatsoever) and i passed 3 and a half hours IBT max ram at similar settings before (1.195vcore) i'm pretty sure i'l be fine at this voltage, or maybe a tiny tiny bit higher. I made some wrong assumptions about what worked for stability and what i "should" be able to do, which cost me a lot of time and caused a lot of confusion.
System agent voltage down to +0.1 now (i didn't try to lower it further; not really bothered about +0.1, can maybe experiment a bit)
When i set Digital i/o and Analog i/o to +0.5 instead of +0.15, i got a 0x0124 bluescreen before windows even loaded - might nudge them down a bit, but it's obvious they are relevant for stability. By far the most important things i'd say though would be Vcore and VRIN (IVR input voltage), people seem to be using vastly different numbers with different boards, my best is around 1.86 (at least for this vcore) but i've seen some people who know what they are doing with different motherboards set 1.9 - 2.22 VRIN, and i've been told up to ~2.35v should be fine on air (which means basically any cooling with your cpu temperature around normal +ambient cooling, or at least in normal air-water temp ranges) though i personally wouldn't push limits like that for long term when short term it already killed some (though apparently non retail) CPU's
VRIN and a bit of tweaking digital+analog i/o voltages can shave quite a bit off the vcore you need to be stable, or turn unbootable into impervious to all stress testing failures quite easily
|
Aren't slower memory kits usually tighter timings? Even the cheapest 1066 kits available today are CAS 7.
|
United Kingdom20285 Posts
Sc2 doesn't really care about timings AFAIK.
I remembered this:
FPS Shown in Min/Max/Average
2.8Ghz CPU Clock
1066 CL9 FPS: 28/133/61.6--95%
1066 CL6 FPS: 28/134/62.6--96.6%
1333 CL9 FPS: 30/139/64.8--100%
1333 CL6 FPS: 30/139/65.2--100.6%
1600 CL9 FPS: 31/142/67.2--103.7%
1600 CL6 FPS: 31/142/67.3--103.8%
I'l tighten the timings to cas8 or something at one point and see if the game cares, but i don't really think so. 8 vs 12 might be relevant
|
On June 17 2013 00:11 SoulWager wrote:
Aren't slower memory kits usually tighter timings? Even the cheapest 1066 kits available today are CAS 7.
For whatever reason, that doesn't seem to help very much. The memory speed is what determines bandwidth. The various caches of the CPU hide higher latency better than low memory speed.
There's also this:
7 / 1066 = 0.0066
10 / 2000 = 0.005
10 / 2000 < 7 / 1066
Those 10 clock cycles at 2000 MHz take a lot less real time than those 7 clock cycles at 1066 MHz.
Interestingly, the BIOS of my Gigabyte board increases latencies through that formula by itself if I try to overclock memory speed. In that 7 at 1066 MHz example, if I would set that memory to 2133 MHz and leave latencies on Auto, the board would change the 7 into 14. The latency in real time would stay the same, but the memory would still have double the bandwidth if it would actually work without errors at that higher speed.
|
United Kingdom20285 Posts
I like this scaling i'm gonna try and grab 2400mhz
I'm uncertain of IMC's capabilities (ivy bridge hit lots of trouble around 2200-2400 two-stick, or even ~2000-2200 four-stick often i hear) people report haswell at stock can take massive memory frequencies, well into the 3000's for two stick, but loses memory frequency headroom when you OC. It's unclear if core frequency costs this headroom, or uncore etc; I saw almost no performance gain in sc2 from increasing uncore clocks, so if i had to leave it at stock, that's not really a big deal if it doesn't play a major part in performance for anything else i want to do. 1600 cas10 to 2400cas12 could give pretty groundbreaking increases on minimum and average FPS, as much as 15-20% even perhaps. That's a LOT more than i would have previously expected from RAM
|
Cyro, your graphs are so pretty thanks for doing all the work! Have you done all of this on OCN yet?
I'm also stable at 4.5 Ghz. I just passed seven and a half hours of AIDA64, along with a long IBT at the beginning, so I think I'm good to go. My stats are:
Asrock Z87M Extreme4
44x multiplier @ 1.18 Vcore
40x cache/uncore @ 1.1V
1.85 VRIN
+.15 Digital, Analog IO and System Voltage offset
1.65V 1600mhz RAM
Max temps in IBT were about 87, 84, 79, 76 iirc, and AIDA was 84, 82, 80, and 74. I might try and up the voltage to 1.2 or 1.21 and see if I can hit 4.6 Ghz, but if I can't hit it at 1.21, I don't think I'll be able to reach it. I might be able to go a little bit lower with AVX off and testing using x264 though, I'll possibly try that.
|
United Kingdom20285 Posts
Cyro, your graphs are so pretty thanks for doing all the work! Have you done all of this on OCN yet?
Lol, random online chart tool, and np
Probably not going to post anywhere else, not much to gain. I dont see any point in getting into an online argument about how silly temperatures are with AVX when i can't provide absolute evidence of all of the whats, hows, why's etc. I'm sure beyond reasonable doubt that it's some weird artificial temperature gain and i think it's silly to overclock around it when it does not happen outside of specific stress tests under the condition of them heavily using avx instructions, but i can't exactly convince people of that without more data, evidence and understanding of everything (unless something were to happen for instance vcore readout point showing +0.12v with a multimeter only when running ibt/prime and only with avx enabled, at which point it's obvious what's going on) so i'm not going to try. I showed my temps in one thread, pretty much nobody even bothered to comment on it (~70c with x264 absolute max, ~74c ibt as expected absolute max in shorter time with avx off, but thermal failure with avx on) so i'l hold back unless i can say things for certain. And the other data, it's more relevant to sc2, very few people on OCN know or care about sc2 in particular (the ones that do are more likely to visit teamliquid than OCN)
1.21, I don't think I'll be able to reach it. I might be able to go a little bit lower with AVX off and testing using x264 though, I'll possibly try that.
With AVX off, your temps will drop like a rock in everything, to pretty much where they SHOULD be in the first place. Nothing you would actually run changes temp even a notable amount with avx on/off (from what i've seen so far) it's just some stress tests that go wild. You should be able to run quite a bit more than 1.2v on an i5 with a u12s. On ivy bridge you could probably do 1.35.
|
By the way, do you have a power meter? How much power is consumed during IBT with AVX enabled and then with it disabled? How much under say maxed out x264 encoding? Why not just set a power or current cap such that you throttle slightly down only if under overclocked synthetic AVX2 load? (but high enough such that other real-world loads are unaffected) Then you can leave AVX on and maybe even with lower voltages / higher clocks.
Load-line calibration on Haswell is for IVR output and not IVR input, right? So it's something clearly defined by Intel and implemented by the IVR?
|
I'm successfully (so far) testing 4.6Ghz on 1.23Volts with AIDA... I'm 50 minutes in and temps are all 88celsius. It seems core temp has had a problem reporting the temps on my CPU, Aida seems much more accurate. All settings the same as my previous posts other than the vcore which is up .05V.
Haswell seems like it's fantastic for overclocking if you know what to do. With AVX on I'm pretty much at the limit of what I think I can test on air with AIDA and the like, although I could always use x264 to test. (Without delid of course). It seems like I should be able to get 4.6 stable here, which I think would be a pretty good result. Cyro is testing like 4.9Ghz right now, but I'll leave it to him to give specifics, other than him being crazy ^^
|
United Kingdom20285 Posts
my chip is too voltage hungry
MOAR POWERRRRRR
By the way, do you have a power meter? How much power is consumed during IBT with AVX enabled and then with it disabled? How much under say maxed out x264 encoding? Why not just set a power or current cap such that you throttle slightly down only if under overclocked synthetic AVX2 load? (but high enough such that other real-world loads are unaffected) Then you can leave AVX on and maybe even with lower voltages / higher clocks.
Very interesting and i'l look into it
|
|
|
United Kingdom20285 Posts
![[image loading]](http://i.imgur.com/CSgBnOz.png)
Gotta stabilize higher frequencies.. Passing lots of stuff but hitting 0x0101 bluescreens seemingly regardless of vcore and not sure what to adjust
|
I'm curently at 4.3 right now, 1.15v vcore, everything else is auto. No problems running 24/7 but keeps blue screening after 2-3 hours on prime95. I usually let prime95 run and then go do something else so not sure if it is a thermal or voltage problem yet, the CPU takes a while to reach over 80C.
|
Try making some of the changes we have... They make a huge difference. What motherboard do you have?
|
I disagree with few points made in this thread.
1. Stress testing, as name suggests, is supposed to push the limits of the platform. By disabling AVX/HT you just make it lighter on the stress. What's the point?
2. There is a reason that HT/AVX is implemented in this chip. And that is to improve performance in specific cases. Assumption that:
"disabling specific parts of the chip makes it OC better"
is being equal to:
"the chip is OCing better"
is simply not viable, isn't it?
3. AVX raising temps is not strange at all. It has been very distinctive since the first benchmarks of Sandy Bridge, where it was implemented by Intel for the first time.
|
United Kingdom20285 Posts
^That's all kinda one point in one.
I never planned to leave AVX disabled, my temperatures under full CPU load with AVX enabled are lower than IBT without AVX. They are barely affected at all under such real world load - while enabling it for IBT makes temps soar 20-25c higher which is just unrealistic - There was no other CPU that i have seen that would have such a massive temperature gap between the most intense real world loads you can come up with, and synthetic loads. Having spent days each overclocking three architectures (and learning temperature and overclocking characteristics in depth) i think it's really silly to base a Haswell overclock around synthetic-load-with-avx temperatures. It behaves differently to other architectures, the gap is much much wider.
Oh and HT, i wouldn't run it disabled with any actual overclock. Kinda silly to buy a 4770k and do that, no?
Is there a good argument for setting an overclock designed for stability, gaming and x264 load, based on the temperatures of synthetic-with-avx loads, which are >20c hotter than synthetic-without-avx loads, that are already hotter than full CPU load of x264-with-avx?
^Bit tricky to use terms, but that's what i am trying to say.
I believe in being stable against any and all loads etc; it just seems absurd in the very specific case of haswell with synthetic-avx to hold yourself back so much for something that, specifically, will never happen.
If there's a good argument for not doing this, I'd like to hear it, but it seems to me that the intuition going against it is down to misunderstanding with how Haswell behaves, or sticking to the roots of conventional CPU overclocking even when it has disastrous results for this architecture, which is why i did what i did.
Also, i am sure a lot of my old posts look pretty bad/uninformed; I was learning on the fly, and it seems i have a complete new understanding of how things work every single day. Stuff will probably be a lot clearer a month from now, for everyone
|
Its been documented more than once by several people that its possible for has well to override your voltage settings and request additional voltage on the order of .1 or more with unrealistically high AVX loads - some in "adaptive" mode if your motherboard has it, others normally. The other CPU architectures certainly don't do this. In this case it doesn't make sense to test with AVX in cases where it limits your stress testing by temperatures. While the extra .1v isn't bad by itself, it is when it arbitrarily limits your overclock by artificial temperature increases.
I've gotten 4.6ghz stable with AVX hitting about 89c on IBT at 1.23ghz, passing an hour and a half of AIDA. on the other hand, x264 crashed on me until 1.245V, but even then it was only hitting 75c or so. So no, we aren't necessarily limiting the stress put on the platform by disabling avx or using x264. And in this case, conventional wisdom seems to fail. No other CPU architecture afaik overrides your fixed voltage and takes you from 1.2 to 1.35V on abnormally high AVX loads.
|
On June 17 2013 20:42 Alryk wrote:
Try making some of the changes we have... They make a huge difference. What motherboard do you have?
I'm running an MSI Z87 MPower Max. It has an ez mode overclock where you push a button and it does 4.2Ghz automatically, setting the vcore at 1.2v. I used it for a few days then try manual overclock. I wouldn't want to have to use 1.2v+ until maybe 4.4-4.5. I'll look into other settings as well.
|
On June 17 2013 14:09 JDI1 wrote:
I'm curently at 4.3 right now, 1.15v vcore, everything else is auto. No problems running 24/7 but keeps blue screening after 2-3 hours on prime95. I usually let prime95 run and then go do something else so not sure if it is a thermal or voltage problem yet, the CPU takes a while to reach over 80C.
Does it necessarily matter that you BSoD after such a long prime95 session?
I've been using 4.4ghz @1.152V for quite a few days now. It's passed pretty hard IBT but only 30 minutes of prime95. I don't know if there's something you normally do that'd actually be equivalent to 30 minutes of torture test, let along 2-3 hours. What is the normal bench mark for stability testing? I thought IBT is most likely enough.
|
On June 18 2013 12:27 Spec wrote:
Does it necessarily matter that you BSoD after such a long prime95 session?
I've been using 4.4ghz @1.152V for quite a few days now. It's passed pretty hard IBT but only 30 minutes of prime95. I don't know if there's something you normally do that'd actually be equivalent to 30 minutes of torture test, let along 2-3 hours. What is the normal bench mark for stability testing? I thought IBT is most likely enough.
That's what I think too, even though I thought IBT is supposed to be more strenuous than prime95. I learned how to overclock when IBT still wasn't around yet and prime95's blend test was the stress test of choice. Not sure how the new AIDA 64 and others stack up though.
|
On June 14 2013 16:06 Cyro wrote:nsfw? + Show Spoiler +![[image loading]](http://i.imgur.com/DPKbeLN.png) As a slight update to this i 0x0124 bluescreened while writing post, put system agent to +0.165v, 0x0124'd while on windows load screen, so put it to +0.18v and doing extended testing soon. Default values might be different for different boards, and i'm sure different CPU's will like different amounts of everything. VRIN, system agent, and i hear analog+digital IO can all have large effects on the vcore you need to be stable, or if you are stable at all at a certain overclock - I've failed 4.5 at 1.21vcore before but i'm seemingly comfortably below 1.16 now which is not a small gap. I had a LOT of trouble stabilizing 4.6 without really big vcore increases, and new knowledge+understanding will probably make it easier The more i come to understand Haswell, the more it seems like first impressions of myself and pretty much everybody else were wrong.
where did you get that background? pls, let me know. lol.
|
United Kingdom20285 Posts
Secret 
|
On June 18 2013 12:42 JDI1 wrote:Show nested quote +On June 18 2013 12:27 Spec wrote:
Does it necessarily matter that you BSoD after such a long prime95 session?
I've been using 4.4ghz @1.152V for quite a few days now. It's passed pretty hard IBT but only 30 minutes of prime95. I don't know if there's something you normally do that'd actually be equivalent to 30 minutes of torture test, let along 2-3 hours. What is the normal bench mark for stability testing? I thought IBT is most likely enough. That's what I think too, even though I thought IBT is supposed to be more strenuous than prime95. I learned how to overclock when IBT still wasn't around yet and prime95's blend test was the stress test of choice. Not sure how the new AIDA 64 and others stack up though.
There are some people that like AIDA64 the most.
I think there's various client programs for Folding@Home, for graphics cards and for the CPU. That could be neat and make the testing feel somewhat useful.
This is about Ivy Bridge:
After prime95 ran for a day, the PC seemed to run perfect, no programs suspiciously crashing anymore, etc. There were still those "WHEA" warnings in the Windows event viewer. Those disappeared after adding about 0.02 V.
It looked like this:
at 1.215 V, IBT ran fine
at 1.235 V, prime95 ran for a day
at 1.255 V, WHEA warnings are gone
Despite that increase in temperature with IBT, prime95 found errors while IBT ran fine. The PC also was definitely not stable in practice. It sometimes crashed with a BSOD or programs stopped.
I have a hunch it's because I did run prime95 for hours and hours, while IBT for a much shorter time. IBT could perhaps be just as good as prime95, but it just never ran for something like 12 hours. I kind of don't want to use it for that long. You can still use the PC to do light stuff like web browsing while prime95 is running.
In hindsight, I should have done a lot less testing. The perfect lowest possible voltage to run prime95 still had those WHEA warnings. There are four 0.005 V steps in between that test and the PC running perfect. I should have done testing in much larger steps, perhaps 0.02 V.
|
On June 18 2013 22:00 Ropid wrote:Show nested quote +On June 18 2013 12:42 JDI1 wrote:On June 18 2013 12:27 Spec wrote:
Does it necessarily matter that you BSoD after such a long prime95 session?
I've been using 4.4ghz @1.152V for quite a few days now. It's passed pretty hard IBT but only 30 minutes of prime95. I don't know if there's something you normally do that'd actually be equivalent to 30 minutes of torture test, let along 2-3 hours. What is the normal bench mark for stability testing? I thought IBT is most likely enough. That's what I think too, even though I thought IBT is supposed to be more strenuous than prime95. I learned how to overclock when IBT still wasn't around yet and prime95's blend test was the stress test of choice. Not sure how the new AIDA 64 and others stack up though. There are some people that like AIDA64 the most. I think there's various client programs for Folding@Home, for graphics cards and for the CPU. That could be neat and make the testing feel somewhat useful. This is about Ivy Bridge: After prime95 ran for a day, the PC seemed to run perfect, no programs suspiciously crashing anymore, etc. There were still those "WHEA" warnings in the Windows event viewer. Those disappeared after adding about 0.02 V. It looked like this: at 1.215 V, IBT ran fine at 1.235 V, prime95 ran for a day at 1.255 V, WHEA warnings are gone Despite that increase in temperature with IBT, prime95 found errors while IBT ran fine. The PC also was definitely not stable in practice. It sometimes crashed with a BSOD or programs stopped. I have a hunch it's because I did run prime95 for hours and hours, while IBT for a much shorter time. IBT could perhaps be just as good as prime95, but it just never ran for something like 12 hours. I kind of don't want to use it for that long. You can still use the PC to do light stuff like web browsing while prime95 is running. In hindsight, I should have done a lot less testing. The perfect lowest possible voltage to run prime95 still had those WHEA warnings. There are four 0.005 V steps in between that test and the PC running perfect. I should have done testing in much larger steps, perhaps 0.02 V.
I have a similar example... 4.6Ghz at 1.23V passed the Very high test x10 on IBT no problems, at 88c. But it failed x264 around 20% of the second pass even though it was only around 75c. I needed 1.245V to finish getting 4.6 completely stable as far as I could tell. I have since dropped down to 4.5 @ 1.18v, but IBT seems to only get you on the edge of stability, but both Cyro and I seem to need a couple more voltage UPS before getting fully stable. When I had 4.4 on like 1.1 (pass IBT but would have failed others) I crashed in SC2. So if you can't pass thirty minutes of Prime, id consider upping your vcore a bit.
@JD: the CPU will reach its "terminal temperature" so to speak after maybe half an hour of prime or any stress test imo. If you haven't crashed by then its probably a voltage issue. Have you only touched vcore?
|
On June 18 2013 20:17 Cyro wrote:Secret
pl0x halp. I want to has it pleeeze ^_^
|
So I am running the i5 at 44x 1.152v with Asus Z87-A mobo.
Temp at idle is low 40s and max load around mid 70s.
The thing is that I saw someone with 44x with 1.14v. So I tried to go back in to drop the voltage to see if it'd be stable, but even though my bio is set at 1.14v, cpu-z still shows 1.152v.
Is anyone familiar with Asus Z87-A? Am I overlooking something?
|
The voltage display jumps in pretty large steps. It's also only a guess. You have no choice but to trust there's a change with a different setting in the BIOS.
|
United Kingdom20285 Posts
Trust what you set in the bios, not what CPU-Z says. If you can hold 44x at 1.14v (pass 10x run ibt max ram, other stress tests etc) then that's pretty nice
|
|
|
United Kingdom20285 Posts
Yea man i jumped on it straight away
I guess i didn't update this thread very much recently, i didn't want to flood etc, i should throw that in OP, and make a big update to it at some point. I've felt busy, ever since i got new build stuff (:
talked to sin a little nearer the end of the thread too
|
Cyro hurry up and do updates and stuff for all of us lazy kids
|
United States7481 Posts
cross-posting this here in case anyone has some ideas.
CPU VRIN Override LLC: Extreme
CPU VRIN Override Voltage: 1.9
VCore: 1.238 (I typed in 1.24 and 1.241 but both times it went to 1.238)
Ring Voltage: 1.15
CPU Multiplier: 44x
c1e, c3, c6/c7, and eist: on
My GPU is at stock.
I tried some brief OCing a couple weeks ago and validated it with a little while of AIDA64 before booting up some games and noticing lower FPS, so I reverted to stock and forgot about it for a few weeks until I was less busy.
I'm back today and testing with 3dmark Firestrike (not extreme)
When I had my CPU at stock, I had Graphics Score 8162 and Physics score 10547. Pretty much what I expected looking elsewhere.
When I Overclocked, my physics score went up to 12489, but my graphics score plummeted to 4676.
I logged CPU temps, and they went about 5c higher under load. I think that's reasonable. I also logged everything GPU-Z logs and nothing looks out of order there either. I can share these logs on request.
Anyone have any ideas about this?
4770k, gtx 770 2gb windforce
|
Is your GPU still hitting the right clocks and temps?
|
United States7481 Posts
|
United Kingdom20285 Posts
I got no idea sorry (or i would have posted in i think the ocn thread where i saw it before)
Im still hitting wall getting past ~4.5 with ht on, or ~4.7 with ht off
Prime 24 hours passed at 1.33vcore with avx disabled (i never saw anything but 0x0101 bluescreens past ~1.32vcore), i got an x264 fail at 1.38 with low cpu load, temps didnt hit 75. AVX enabled 4.7 ht off passes x264 below 1.3v, avx enabled or disabled toggling ht on fails it at 1.38vcore?? http://i.imgur.com/oS1e0ii.pngThere you see a pass, and my temps with x264 (current) and max temps (from ibt 10x standard pass) with avx off. Toggle avx on, no change to x264 temps or stability. Works fine. Toggle ht on, add 0.05vcore and drop 100mhz? fails within a minute, 0x0101 and locks up before or during the crash dump
|
I've been trying to overclock my 4770k on a asus z87m-pluss but I'm having issues.
I started with something simple: x42 with 1.25 volts. After 1 hour of OCCT I had no issues, but when I load of SC2 and plays a game for 12 minutes the PC just reboots without any BSOD. I've always experienced the same with just having my computer almost idle (watching streams) and the computer just reboots without any errors. It seems like standard clocks is the only thing that works for me, what can be the issue of this?
Thanks.
|
United Kingdom20285 Posts
What cooling?
And i'd lower your vcore. Try 1.15 first if you can, manually set the cache/uncore ratio to 34. If you can as well, set the CPU voltage regulator input voltage or VRIN to 1.7v (NOT THE VCORE, IT IS A SEPERATE SETTING, by taking any tips you agree that i am not liable for any damages in miscommunication because 1.7vcore would break stuff if you set that..)
And also, you can apply llc to cpu input voltage (vrin), i'm not sure what that one is called on Asus. The same voltage that you set to 1.7. Another thing you can do is increase digital io voltage by +0.05v offset.
I'd guess what's happening is just that you didn't control the cache/uncore manually and it's doing stuff that you don't want it to do
|
You're probably correct, going to try that.
I have an h80i (bad choice, lol)
|
So CPU core voltage is vcore and CPU cache voltage override is vrin?
I also have one that is CPU input voltage (VCCIN) I think that's more likely to be the vrin
You wanted me core ration to 42 but the cache ration to 34? I have one min cache ration and one max cache ratio
|
United Kingdom20285 Posts
Core voltage is vcore. Cache/uncore/ring voltage is uncore.
VCCIN is IVR input voltage (VRIN) i think, but please confirm before you set anything to 1.7 volts. It should be 1.8 by default - if you see that value, just bump it down to 1.7 (:
You wanted me core ration to 42 but the cache ration to 34? I have one min cache ration and one max cache ratio
Yes, min/max it to 34 and give cache/uncore/ring 1.1 volts
I've never seen asus z87 before and i'd rather not break anything so make sure you got the right ones (:
|
My mobo is very secretive. Only shows auto for everything x.x
EDIT: Says that it is 1.776, I guess thats kinda 1.8
|
@Evire:
There's an Asus guide for overclocking Haswell, I think. You might want to search for that. Perhaps look in their forums or those "asus rog (republic of gamers)" forums.
|
On August 12 2013 20:37 Ropid wrote:
@Evire:
There's an Asus guide for overclocking Haswell, I think. You might want to search for that. Perhaps look in their forums or those "asus rog (republic of gamers)" forums.
I did and I watched the youtube video. He never chanced VCCIN in the video or cache voltages.
I booted into windows, running tests.
|
|
|
United Kingdom20285 Posts
It's important to change vrin especially at high or low overclocks, it can make you require significantly less vcore for stability, the default of 1.8 + vdroop doesn't work perfectly for everything
I still don't know asus settings, bios and names very well at all though. I know Gigabyte like the back of my hand, easily spent a hundred hours or two talking about them and messing around in bios/os in the last few months
|
You should probably update the first post with everything you've learned so far :p
|
United Kingdom20285 Posts
Probably.. It's a wreck T-T
|
I'm running 4.3ghz with 1.15 vcore, really awesome. Do I just try setting the multiplier higher until it fails, and then add more vcore? If I increase vcore with 0.05, then how much would you increase uncore with? And would you increase the uncore ration proportionally with the vcore ratio?
|
United Kingdom20285 Posts
I'm running 4.3ghz with 1.15 vcore, really awesome. Do I just try setting the multiplier higher until it fails, and then add more vcore? If I increase vcore with 0.05, then how much would you increase uncore with? And would you increase the uncore ration proportionally with the vcore ratio?
I'm running 4.3ghz with 1.15 vcore, really awesome. Do I just try setting the multiplier higher until it fails, and then add more vcore?
^Yea. Keep VRIN 0.5 above Vcore if in doubt too, with llc on it.
If I increase vcore with 0.05, then how much would you increase uncore with?
^Don't increase vcore that much at once, use smallest amounts you can. Leave uncore at 34x for now and ring/cache volts at 1.1. You gotta stress test properly
And would you increase the uncore ration proportionally with the vcore ratio?
^You don't have to clock it as high as the CPU, and it's more difficult to clock to a certain point (like i'm running 4.7 core but getting uncore past 4.4 would be a lot of trouble)
Focus on core stability first, see what you can do, what temperatures you have, make sure performance is correct with cinebench 11.5 (and set it above normal or high/realtime priority) and then raise uncore later
|
I had to pump the vcore really hard to get a stable 4.5GHz, right now I have 1.2 volts because it crashed with 1.195. Maybe I should change some of the other settings? I notice in your first post that you changed a lot of other different voltages, can I do the same to achieve stability at a lower voltage? My CPU is close to 85celcius
|
United Kingdom20285 Posts
+0.05 for 200mhz is not pumping vcore hard, it's actually really good, and 4.5 stable on 1.2vcore is great assuming it's actually stable. What are you testing with, and where are you getting the 85c from?
Not much else to do aside from adding +0.05 on digital io volts and 1.7 vrin
|
I'm running CCT (? program you recommended) and getting 82C in HWmonitor. My CPU fan is really weird because one core is above 80, second one is below, and third and forth are closer to 70C
|
United Kingdom20285 Posts
Aa, did i say that to you? Sorry lots of people and confusing
You running linpack with avx unchecked in OCCT?
|
On August 13 2013 07:07 Cyro wrote:
Aa, did i say that to you? Sorry lots of people and confusing
You running linpack with avx unchecked in OCCT?
Yes. Computer actually crashed after 2 hours testing
EDIT: Looks stalbe 4.5GHz with 1.215 vcore
EDIT2: Crashed around 2 hours, lol. Changing to 1.218
EDIT3: The 4.5GHz at 1.25 seemed stable after 2 hours of stress testing, but the moment i load up OBS and try to stream som PayDay 2 it crashed after 5 minutes. I thought stress tests would be way more demanding that streaming.
|
Sorry for making two posts in a row but
Seems like my OC is stable with stress test but streaming games seem to crash it. Is this a problem with OBS, or am I looking at other parts of my computer failing? I have not OC'd anything else, stock on GPU.
|
Yeah, you're wasting your time with stress tests, frankly. You should look at x264 video encoding or other real programs that you can use easily for a test. I've seen people mention the Folding@Home CPU client. Someone uses a chess engine. Also, when things run seemingly stable, take a look at the Windows event viewer and see if there's "WHEA warnings" showing up. Those don't crash the PC but they can be made to disappear by tweaking the overclock settings.
Old experiences from Ivy Bridge and Sandy Bridge are probably a bit worthless, but even with that, stress tests were not that good when only running short tests. For me with Ivy Bridge, there were errors that only showed up after 12 hours of prime95 and things like that. Only after I got prime95 to run indefinitely were games running perfectly stable. In hindsight, it was probably a waste of time with the hours of stress tests.
|
On August 14 2013 00:43 Ropid wrote:
I've seen people mention the Folding@Home CPU client.
Tried it and it ran for 30 minutes without any problems or errors that I could see. Still crashed when I ran OBS together with SC2. Its not OBS nor SC2's fail cause it crashes in other games too, I even tried 4.2GHz without any luck
|
On August 13 2013 16:37 Evire wrote:Show nested quote +On August 13 2013 07:07 Cyro wrote:
Aa, did i say that to you? Sorry lots of people and confusing
You running linpack with avx unchecked in OCCT? Yes. Computer actually crashed after 2 hours testing EDIT: Looks stalbe 4.5GHz with 1.215 vcore EDIT2: Crashed around 2 hours, lol. Changing to 1.218 EDIT3: The 4.5GHz at 1.25 seemed stable after 2 hours of stress testing, but the moment i load up OBS and try to stream som PayDay 2 it crashed after 5 minutes. I thought stress tests would be way more demanding that streaming.
Lol you should ask Cyro about getting x264 stable.
If you are going to be streaming, your stress-testing will need to include encoders. It's going to require a bit more conservative clock then say a game-stable OC :p
|
United Kingdom20285 Posts
EDIT3: The 4.5GHz at 1.25 seemed stable after 2 hours of stress testing, but the moment i load up OBS and try to stream som PayDay 2 it crashed after 5 minutes. I thought stress tests would be way more demanding that streaming.
You're confused because linpack/prime don't actually work too well for determining stability, i had to go from ~1.285 to ~1.315 after my ibt(linpack) pic and when i was unstable before, i could pass 48h prime. People are reporting everywhere being able to pass stress tests, but failing in games, and also x264
If you are going to be streaming, your stress-testing will need to include encoders. It's going to require a bit more conservative clock then say a game-stable OC :p
Some games aparantly a lot harder on the CPU than linpack. As a basis for stress testing now i think i throw avx off and use linpack for cpu (to see if everything works, i didn't find anything good for determining a final vcore), prime custom blend with 7000MB RAM (lower fft lengh if i want to short test, like 1 min if i'm gonna leave it an hour or 10 min for overnight) for RAM stability (didn't try memtest yet, but bad ram settings seem to mess up prime really fast, wasn't sure what to test with)
And then, avx toggled back on for real world usage, x264, gaming, benchmarks, etc.
For you i'd say either just add a bit of vcore or drop a core multiplier. It's confusing though - if you can pass 2 hours @1.21vcore but still fail at 1.24? I didn't see that happen personally. Did you increase VRIN a little bit again (like 1.7 to 1.75)
Only after I got prime95 to run indefinitely were games running perfectly stable.
The difference was with ivy, if you could pass all fft lenghs, you were generally good for everything. With haswell, you pass all fft lenghs in a 24 hour test, and then open x264 and fail within 30 seconds
|
On August 13 2013 16:37 Evire wrote:Show nested quote +On August 13 2013 07:07 Cyro wrote:
Aa, did i say that to you? Sorry lots of people and confusing
You running linpack with avx unchecked in OCCT? Yes. Computer actually crashed after 2 hours testing EDIT: Looks stalbe 4.5GHz with 1.215 vcore EDIT2: Crashed around 2 hours, lol. Changing to 1.218 EDIT3: The 4.5GHz at 1.25 seemed stable after 2 hours of stress testing, but the moment i load up OBS and try to stream som PayDay 2 it crashed after 5 minutes. I thought stress tests would be way more demanding that streaming.
stress tests test heat tolerance and certain instruction sets that really push CPU load and the stress tests are more demanding than streaming PayDay 2 using OBS, but the difference is that streaming PayDay 2 using OBS probably makes use of more different instruction sets than the stress test
I also think using stress tests with AVX is overrated because it does not represent real world at all and is just useful for really pushing core temperatures which helps reveal temperature related overclock issues, but doesn't help reveal issues for real world applications
|
United Kingdom20285 Posts
Any suggestions for what i should edit OP to?
|
Not really sure what you can add to the OP other than some links as to some guides on how to overclock a Haswell or cooling/temps.
For the stability testing people are looking at, a 10x pass on high or very high on IBT should get you past the 1st gate. A 12 hour run on p95 (blend or 8Ks, blend if you have overclocked RAM, 8Ks if you know that RAM/GPU is stable and CPU is the only unknown factor) will get you past the 2nd gate. Play a game for a few hours and check it out (make sure it is on the highest settings and a fairly new game, not Doom 3 quality) will get you past the 3rd gate. Then you can try folding, which should check it out whether or not the overclock is completely stable.
Of course, a run of p95 + some gaming should do the trick for normal usage so you don't have to go through all the hassle of taking that long.
Here are some links to contribute:
This link gives detailed information on the architecture and the adventures of a Haswell overclock:
http://www.overclock.net/t/1401976/the-gigabyte-z87-haswell-overclocking-oc-guide
This link talks about the benefits of Haswell and the amazing performance that you can get from your $50 ram kit:
http://www.overclockers.com/3step-guide-to-overclock-intel-haswell
|
United Kingdom20285 Posts
I found p95 to be quite bad for CPU stability, in that prime stable for me doesn't mean game stable, or x264 stable. I passed prime >12 hours multiple times and failed within 30 seconds in x264, cause you can have OC now that'll pass all prime fft lenghs, but be game unstable, and also, i need like 0.03v more for gaming than i do for linpack stable
Thanks for the post
With a very good air cooler or all-in-one water cooler, you’re looking at a heat limited voltage cap of about 1.25 V. At that voltage with air or AIO cooling, you’ll be seeing temperatures in the upper 80′s to lower 90′s (°C) range under normal full processor load.
I don't like that second link a lot, it misses out a lot of specifics and that temperature range seems like they are testing for avx1 (but not avx2 or normal loads)
|
I mostly looked at the RAM performance. Nothing really about the rest, the 1st link gets really into detail with the chips which is what gives the info needed to look at the 2nd link.
|
United Kingdom20285 Posts
There's a lot of misinformation there about not being able to run fast RAM with OC though. AFAIK that's not true at all you just gotta bump the right voltages. I mean Belial set 4.8ghz/3000cas13 no problem, people run fast ram at more than stock speeds all the time
|
United States7481 Posts
On August 15 2013 20:44 Military wrote:Not really sure what you can add to the OP other than some links as to some guides on how to overclock a Haswell or cooling/temps. For the stability testing people are looking at, a 10x pass on high or very high on IBT should get you past the 1st gate. A 12 hour run on p95 (blend or 8Ks, blend if you have overclocked RAM, 8Ks if you know that RAM/GPU is stable and CPU is the only unknown factor) will get you past the 2nd gate. Play a game for a few hours and check it out (make sure it is on the highest settings and a fairly new game, not Doom 3 quality) will get you past the 3rd gate. Then you can try folding, which should check it out whether or not the overclock is completely stable. Of course, a run of p95 + some gaming should do the trick for normal usage so you don't have to go through all the hassle of taking that long. Here are some links to contribute: This link gives detailed information on the architecture and the adventures of a Haswell overclock: http://www.overclock.net/t/1401976/the-gigabyte-z87-haswell-overclocking-oc-guideThis link talks about the benefits of Haswell and the amazing performance that you can get from your $50 ram kit: http://www.overclockers.com/3step-guide-to-overclock-intel-haswell
Yeah man, like Cyro said I see almost no benefit in running hours of straight cpu stress tests with prime95 or anything similar. You can run a very quick one if you're just trying a new OC, but just playing a game or running a x264 will reveal issues almost right away.
|
I see it more of as a extreme measure people take. Some people really want to make sure that they are 99% stable (could never be 100% stable) but running like 1 ~ 2 hours of P95, game, some x264 and some IBT (mainly for temperatures) is enough for some.
I think with the RAM, Haswell's performance relationship with RAM is similar to APUs from AMD.
|
United Kingdom20285 Posts
The thing is, with ivy bridge, if you could prime through all of the fft lenghs for a decent period of time, it was very difficult to make the CPU have problems in other areas, while Haswell is routinely passing prime/linpack with a chunk less voltage than required for stability in x264 and games
|
So I'm finally in the process of overclocking my i7-4770K.
+ Show Spoiler [Hardware] +- Intel Core i7-4770K Processor
- Gigabyte GA-Z87X-UD3H, Socket-1150
- Noctua NH-D14 Cooler
- EVGA GeForce GTX 780
- Corsair Dominator DDR3 1600MHz 8GB
- Corsair AX 860i, 860W PSU
- CM Storm Trooper Gaming Big Tower
I've thoroughly read Overclock.net's Haswell Overclocking Thread and The Gigabyte Z87 Haswell OC Guide, meticulously following all advice, and I'm starting off with a basic 42x template, very conservative. I've also watched and followed this and this for a basic overview of the BIOS and some example settings.
Right now I'm running Prime95 in Large FFT mode. It's been going for about 40 minutes so far, and this (screenshot taken like half an hour ago because I'm busy with other stuff as well) is the current output. It all looks fine to me, but if someone can take a look at my monitors to make sure there's nothing amiss I would appreciate it.
Also, what other benchmarking software do you guys recommend?
Edit: Following this guide I was able to determine that my CPU is quite handily in in the top 50th percentile group. It booted in 4.6GHz, 1.20V, and it is currently Prime95ing away using Large FFT @ 4.6GHz (4.2 Uncore) 1.25V, reaching 75 degree temperatures. If I were to use IntelBurnTest the temperatures would increase pretty dramatically, though.
Edit 2: AIDA64 is weird. The interface in AIDA claims the temperatures are a stable ~60 with peaks (spikes) of 76. HWMonitor flails rather wildly from value to value, from 63 to 72 to 59 to 78.
Edit 3: After an hour+ of prime95 and then another hour with AIDA64, my CPU seems stable as hell at 4.6GHz. I will mess with it a bit more tomorrow, but the temperatures are approaching 85° even on my Noctua NH-D14 (with the system pushed to 1.260V because of AIDA64), so I'll probably just be looking for ways to reduce the temps more than I am trying to squeeze some more MHz out of the system.
Any information you guys can give me would be appreciated.
|
On August 18 2013 16:40 Aylear wrote:So I'm finally in the process of overclocking my i7-4770K. + Show Spoiler [Hardware] +- Intel Core i7-4770K Processor
- Gigabyte GA-Z87X-UD3H, Socket-1150
- Noctua NH-D14 Cooler
- EVGA GeForce GTX 780
- Corsair Dominator DDR3 1600MHz 8GB
- Corsair AX 860i, 860W PSU
- CM Storm Trooper Gaming Big Tower
I've thoroughly read Overclock.net's Haswell Overclocking Thread and The Gigabyte Z87 Haswell OC Guide, meticulously following all advice, and I'm starting off with a basic 42x template, very conservative. I've also watched and followed this and this for a basic overview of the BIOS and some example settings. Right now I'm running Prime95 in Large FFT mode. It's been going for about 40 minutes so far, and this (screenshot taken like half an hour ago because I'm busy with other stuff as well) is the current output. It all looks fine to me, but if someone can take a look at my monitors to make sure there's nothing amiss I would appreciate it. Also, what other benchmarking software do you guys recommend? Edit: Following this guide I was able to determine that my CPU is quite handily in in the top 50th percentile group. It booted in 4.6GHz, 1.20V, and it is currently Prime95ing away using Large FFT @ 4.6GHz (4.2 Uncore) 1.25V, reaching 75 degree temperatures. If I were to use IntelBurnTest the temperatures would increase pretty dramatically, though. Edit 2: AIDA64 is weird. The interface in AIDA claims the temperatures are a stable ~60 with peaks (spikes) of 76. HWMonitor flails rather wildly from value to value, from 63 to 72 to 59 to 78. Edit 3: After an hour+ of prime95 and then another hour with AIDA64, my CPU seems stable as hell at 4.6GHz. I will mess with it a bit more tomorrow, but the temperatures are approaching 85° even on my Noctua NH-D14 (with the system pushed to 1.260V because of AIDA64), so I'll probably just be looking for ways to reduce the temps more than I am trying to squeeze some more MHz out of the system. Any information you guys can give me would be appreciated.
Priming on Large FFTs isn't a good way to use P95. Prime on 8Ks or Blend for a few hours.
Then you need to play games, different types for an hour or two per game and see if the CPU fails.
I don't really think your CPU is stable with such light stress testing.
|
run P95 blend. I use ~6 hours as my stress time length.
typical timeframes for when I've seen BSODs using P95 blend are <1 hour, 3-4 hours, 5-6 hours
the stress tests are more for trying to find your max overclock speed ASAP for whatever max voltages you are willing to feed into the CPU imo
or you can just just overclock to what you think is "stable" and use your computer for real world activities. if you don't get any instability for 1 week, then increase speed by 100 MHz. if you experience instability, then increase voltages or decrease OC speed.
you can also run a mix of stuff to see if that will produce any kind of instability like
Cinebench
Uniengine
3dMark
http://www.primatelabs.com/geekbench/download/windows/
any any other type of benchmark stuff out there
|
Small FFT is alright, but as it doens't stress RAM/IMC, it's fairly academic in nature. Read - not really useful for 24/7 OCing. It's for this reason that I still value Large FFT more than small, but only to augment custom blend. I usually start with Large FFT followed by overnight Blend. However - I would still follow-up both of this with x264 work at the very end; as x264 stresses the CPU differently than P95.
I used to bother with Unigine in the past, but considering I've had x264-unstable OCs easily pass on these, I would not waste my time with this in the end. If you're x264-stable, you're probably more stable than most overclocks on the Net. People who think Linpack/IBT represent stable overclocks are being silly. These are fast ways to check, but will consistently fail in P95 and eventually in real-world use.
|
On August 19 2013 03:32 mav451 wrote:
Small FFT is alright, but as it doens't stress RAM/IMC, it's fairly academic in nature. Read - not really useful for 24/7 OCing. It's for this reason that I still value Large FFT more than small, but only to augment custom blend. I usually start with Large FFT followed by overnight Blend. However - I would still follow-up both of this with x264 work at the very end; as x264 stresses the CPU differently than P95.
I used to bother with Unigine in the past, but considering I've had x264-unstable OCs easily pass on these, I would not waste my time with this in the end. If you're x264-stable, you're probably more stable than most overclocks on the Net. People who think Linpack/IBT represent stable overclocks are being silly. These are fast ways to check, but will consistently fail in P95 and eventually in real-world use.
If he isn't overclocking his RAM, or he knows that his RAM is stable... few hours of small FFTs + few hours of blend would be the better combo than large FFTs. That's if he is not OC'ing his RAM or he knows it is stable. If he is also overclocking his RAM or does not know if it is stable or not, then he can do blend for a longer time.
Priming + x264 + play games + folding will pretty much tap into almost every little corner of different types of calculations that the CPU will ever need to do and decide if it is 99.9% stable or not.
|
I did say 'augment' in my post. Since the advent of the IMCs, I've felt that small FFTs have not had much if any relevance for stress-testing. Real world loads (gaming or x264) will be moving and utilizing lots of RAM. Small FFTs will basically only test your CPU/cache - what good is that for real-world relevance? :p
|
On August 19 2013 03:47 mav451 wrote:
I did say 'augment' in my post. Since the advent of the IMCs, I've felt that small FFTs have not had much if any relevance for stress-testing. Real world loads (gaming or x264) will be moving and utilizing lots of RAM. Small FFTs will basically only test your CPU/cache - what good is that for real-world relevance? :p
Honestly, it does nothing when it comes to real-world usage. I mean, that's pretty much every synthetic stress test algorithm. None of us will be crunching the same algorithm endlessly (at least most of us won't be).
|
United Kingdom20285 Posts
If he isn't overclocking his RAM, or he knows that his RAM is stable
With haswell if you leave the secondary voltages as they are, you can become unstable because of a RAM setting just by changing your core multiplier, it's not as simple as that. Also, i think prime95 with avx is a really terrible stress test for Haswell
Aylear, your VIN (VRIN) is far too high, apply LLC and manual it to 1.75 if you don't know what to do with it and you'd be stable on less vcore, especially for the lower clocks
Your uncore is also at 39x, probably on auto ring voltage, which can present instability ESPECIALLY if you are pushing core clock up so i'd always suggest to just manual it to like 34x/30x (not 35x) and manually set 1.15 ring voltage or something like that for OC
Both of those are really important. You should really keep ram/uncore down and mess with core for a while - lots of testing, and bring stuff up slowly. I tried like you to throw up core/uncore/ram, passed a ton of stress tests and was laughing but actually not stable in the loads i actually wanted to use the system for (gaming and x264) until i spent weeks revising overclock
It'd be easier to help with OC if we could discuss and try settings etc, if you want to do that
|
On August 19 2013 04:00 Military wrote:Show nested quote +On August 19 2013 03:47 mav451 wrote:
I did say 'augment' in my post. Since the advent of the IMCs, I've felt that small FFTs have not had much if any relevance for stress-testing. Real world loads (gaming or x264) will be moving and utilizing lots of RAM. Small FFTs will basically only test your CPU/cache - what good is that for real-world relevance? :p
Honestly, it does nothing when it comes to real-world usage. I mean, that's pretty much every synthetic stress test algorithm. None of us will be crunching the same algorithm endlessly (at least most of us won't be).
I don't know why you're so stuck on using small-FFTs. My point is that anything not testing the IMC will have 0 real-world relevance. You are welcome to waste your time using small FFTs, but I imagine most users will want to be as efficient with stress-testing as possible.
*This is also EXACTLY why I don't recommend IBT or LinX (or its equivalents) as being useful stress testers. They don't touch the IMC at all, and guess what, once you start using your PC for actual applications (games or encoding), your IMC is going to get stressed hard. So unless you play games that can fit into JUST your CPU cache (e.g. SuperPI), then you are wasting your time.
|
Not true about IBT/LinX. It uses a lot of RAM and you can configure how much. It's just a bad stress test for another mysterious reason, seems to run fine with lower voltage than x264 encoding or just starting up a random game and playing that.
|
In my experience, large FFTs found RAM errors when I was consistently passing LinX. Just cuz you are using a large amount of RAM does not necessarily equate to high IMC loading.
I've never seen IBT/LinX as anything more than quick thermal tests. Thermal testing is important of course, but I wouldn't classify them as anything more useful than that.
|
On August 19 2013 04:00 Cyro wrote:
It'd be easier to help with OC if we could discuss and try settings etc, if you want to do that
I'm game. Do you want to do it through the forums, or Skype or whatever?
|
Yeah just talk to Cyro on Skype; PM him your handle so he knows who you are haha.
|
I've been talking to him all night haha, he gave me a ton of good settings advice and my CPU looks stable as hell at 4.6GHz at lower temps than I was originally getting. Thanks Cyro. ♥
|
Aylear,
Can you share some of this advice? Haswell is my first CPU and this is also my first custom built computer, and would like some basics on how to OC.
EDIT: Cyro, feel free to jump in and help a newb too, if you have extra time
|
United Kingdom20285 Posts
Oh god don't bump this it's ancient and terrible
What motherboard and cooling do you have?
|
Z87x-D3H, Stock cooling. Can get a fan.
|
United Kingdom20285 Posts
Yea you can't really do much on stock cooling. Some guys on the build resource thread could probably recommend a decent cooler in US, i hate the US CPU cooler market
|
Hahaha oh hi, it's this thread!
Yeah, the OP is pretty much wrong on most counts, partly because Haswell goes against some established overclocking conventions and we didn't know that before we started working on it.
Cyro's the man to talk to, though. I could talk about OCing a Haswell but I'd simply be parroting his advice (and with a weaker understanding of the underlying mechanics behind said advice). Good luck with your OC, hope you won the silicon lottery.
I can say one thing about it: The Hyper 212 Evo is a beast of a cooler, both in size, price, and efficiency. If your computer case can fit that monstrosity it's worth checking out.
|
United Kingdom20285 Posts
The Hyper 212 Evo is a beast of a cooler, both in size, price, and efficiency. If your computer case can fit that monstrosity it's worth checking out.
It's lower end, though.
In the UK, i5 costs ~£165, i7 costs ~£250.
212 costs about ~£20
better stuff such as true spirit 140 power or hr-02 macho come in around ~£33
you can get some of the best air coolers for ~£55
If you're spending £265-350 on CPU and mobo, then saving £13 or even £35 on cooling isn't a big deal, while some people are quite bothered about losing 200mhz. As for US pricing though.. they have less choice in the midrange (?) and i think it was worse compared to CPU/s + Mobo's. Maybe i'm just not used to dealing with higher dollar values.
My personal choice for cooling i5 cheaply and efficiently would probably be a true spirit 140 power with a ty-143 added on it - but that's both not available in the US, and quite noisy when all CPU cores are very highly loaded. It's like 10-15c better than 212 evo though. Very cost efficient way of approaching higher voltage overclocks.
Looking over temperature data, i think 212 is usable but also that some people might be unhappy with it:
Pre-devil's canyon users with undelidded CPU's - 4670k is usually ~6-8c hotter than 4690k
i7 users, especially 4770k - they run ~10c hotter than i5 with the extra threads loaded
people who want to max OC in general with good temperatures
If you wanna save money and you're running an i5 4690k.. it's quite hard to argue against it in US i guess - but that's a good case for it. I haven't been a big fan of it since we moved away from Nehalem and Sandy Bridge being easy to cheaply cool.
|
Both of you guys avoided my question D: But thank you for the advice on purchasing a fan! Hopefully I can find a good deal.
How do I go about overclocking my 4670K? If possible, both generic newbie advice (This will be my first one) and Haswell-specific advice would be appreciated.
:D
EDIT: In other words, a moderate overclock will suffice, I don't need to race the clock. Assume I know nothing about OC :D
|
United Kingdom20285 Posts
It's best to list stuff or maybe talk about it in an interactive way (though that's not needed for a basic OC if you can follow a little in-depth instructions) once you have cooling - i'l write a basic list of stuff when i've got 5 min spare though
ok..
Most important voltages for OC:
Input V (aka VRIN) - the voltage supplied to the CPU which is then turned into a huge list of other voltages by the integrated voltage regulator. Usually best around 0.6 higher than vcore, there are some rules. Good range is probably ~1.7 to 2.0 - some people use significantly more at times but there are questions of what this means for the chip, performance and also lifespan - and using lower values past a certain point might not be necessary nor helpful in any meaningful way. That is to say, there might not be any point going below ~1.7 for a very low voltage, or if you have a 1.25vcore, 1.85 VRIN overclock, there is little to no perceived or even proven benefit that i've seen from trying to stabilize say 1.25vcore, 1.75 VRIN. This is probably the most important voltage because it sources everything else, and a lot of novice overclockers and reviewers don't respect it properly and get a sub-par experience as a direct result of that.
Vcore - Core voltage, has to be above a certain point for X frequency to be stable. If there's no other problem causing stability other than too low vcore, it can be safely increased, usually to ~1.35 or so.
Uncore/Ring/Cache voltage, known as Ring on gigabyte boards. Used somewhat for general chip stability(?) but mainly increased for raising the uncore multiplier. Best to use ~0.1v below Vcore on a general OC, because it's not efficient in terms of performance or lifespan of the chip to push it higher. A ring voltage at 1.4 will be more dangerous than a vcore at 1.4.
---
For a basic, simple Haswell OC, you can get some software (i wrote more here, but eh, simple is better):
An x264 package, set up somewhat to increase CPU load compared to just randomly encoding something - https://mega.co.nz/#!3tAGnAqZ!QbCz2r1fG0WjM8DgGYeExngGypaHftAzPUgTSn2kAdk
important settings:
VRIN - 1.85
Vcore - 1.23
Ring - 1.15
Under advanced power settings(?), there is an LLC setting for VRIN. You can set it to "turbo" out of the options available.
^After that, set core multiplier to ~42. under advanced core settings(?), set uncore multiplier to 33.
I think that's everything to do before booting up - run that test, "x264 Stability Test (64bit - log)" is what i have to run, set it to use 8 threads, 3 loops. Check temperatures and clock speeds in hwinfo sensors tab - www.hwinfo.com - to make sure that they are correct (core temps under 80 or so, core clock speed at 4200mhz, vcore on the sensor near the bottom called "vcore" at ~1.25, etc) and if they're fine, wait for it to complete.
You can keep doing that on different settings. If 4200mhz core passes, reboot and try 4300mhz. If that passes, try 4400mhz etc.
One of them will crash with some bluescreen error message or just crash the encoder, at that point just go down 100mhz, set it to pass 10-20 loops and when you've done that, increase vcore and VRIN by 0.02 (to 1.25 vcore, 1.87 VRIN)
There's a lot more to do, but i think that's a pretty foolproof way to get a basic OC where it's very unlikely to have any problem aside from vcore/core multiplier being off.
After that's done, go to bios and manually enable c3, c6/c7 and EIST. Boot into windows and make sure the power plan is on balanced. Open Hwinfo, it should show core speeds, uncore speeds and further down, that "vcore" sensor - ignore the readings labeled VID. If that worked as expected, core speeds and Vcore should drop when your system is idle, but uncore will be locked at 3300mhz. Just leave it that way for a while and use your system for like a week to make sure stuff works, then it can be adjusted to work at a variable speed (such as 800mhz-4000mhz) - it's not important for performance, so don't worry about it
If you want a very very basic like how do CPU's and voltages and clock speeds work, then i'm not your person for that unfortunately :D
|
Cyro.
This is exactly what I needed. Thanks so much )
|
United Kingdom20285 Posts
np, if you have any questions/want tips etc then just ask, i can't guess what kind of issues or questions you might have
|
|
|
|
|
|
EPT![[image loading]](http://i.imgur.com/Ayt8Ukw.png)
