Silicon Lottery Percentages? What % is a good overclock?

[WelshDdraig]

Hey everyone, as the title suggests, Silicon Lottery Percentages.

 

I know what the silicon lottery is and how it works, but one thing I'm unsure of is the percentages - as in, how bloody lucky you are if your overclock is x% over the default clocks.

So I was wondering if anyone here could shed some information on this.

What percentage boundaries would be considered good, better, best, or Holy S**t?

 

Specifically as to why I want to know. 

I've had my i5 for around 23 months now and am looking to upgrade. Since it can overclock, I've had it overclocked since I've had it.

Up until now, its been at around 4.5GHz.

Only recently, I've decided to push it to what I find is its max on AIR. 

 

My CPU: Intel Core i5 4690K

Cooling solution: Phanteks PH-TC14-PE (Push/Pull)

Stock Speeds - 3500mhz (3.5GHz)

 

My overclock - 4898mhz (give it a little rounding, 4.9GHz)

Equating to a 40% overclock (I think) - rock solid stable (Well, I've experienced no crashes and I've been at this speed for about a week now), at 1.31v (I know CPUz is saying 0.952v, but I've probably left SpeedStep on )

 

Is this a good overclock for the voltage and for what the chip is?

 

Please let me know

Thanks.

 

 

[aisle9]

I'd say you're a winner. Now let's see it survive eight hours of RealBench and OCCT

[iBabySlapper]

 

5 minutes ago, Alexp10v2 said:

Hey everyone, as the title suggests, Silicon Lottery Percentages.

 

I know what the silicon lottery is and how it works, but one thing I'm unsure of is the percentages - as in, how bloody lucky you are if your overclock is x% over the default clocks.

So I was wondering if anyone here could shed some information on this.

What percentage boundaries would be considered good, better, best, or Holy S**t?

 

Specifically as to why I want to know. 

I've had my i5 for around 23 months now and am looking to upgrade. Since it can overclock, I've had it overclocked since I've had it.

Up until now, its been at around 4.5GHz.

Only recently, I've decided to push it to what I find is its max on AIR. 

 

My CPU: Intel Core i5 4690K

Cooling solution: Phanteks PH-TC14-PE (Push/Pull)

Stock Speeds - 3500mhz (3.5GHz)

 

My overclock - 4898mhz (give it a little rounding, 4.9GHz)

Equating to a 40% overclock (I think) - rock solid stable (Well, I've experienced no crashes and I've been at this speed for about a week now), at 1.31v (I know CPUz is saying 0.952v, but I've probably left SpeedStep on )

 

Is this a good overclock for the voltage and for what the chip is?

 

Please let me know

Thanks.

 

 

I'd say that's a really good OC'ing 4690k. Like, you sure that's stable?

 

[Enderman]

That's a really good OC.

How good your OC is also depends on the CPU. 4.9GHz on skylake is much more impressive than 4.9GHz on haswell.

[alihazzi7]

Cinebench r15 it and post the pictures

[WelshDdraig]

@alihazzi7 Just did the R15 run. I did get a 612cb, but my RAM OC made the system crash, put it back to 1866MHz, and did the run again, and got a 551cb.

Spoiler

 

[done12many2]

31 minutes ago, Alexp10v2 said:

@alihazzi7 Just did the R15 run. I did get a 612cb, but my RAM OC made the system crash, put it back to 1866MHz, and did the run again, and got a 551cb.

 

Wow.  That's what my laptop's 2760qm with 1600 MHz DDR3 scores.  552 to be exact.  The CPU starts the bench at 3.5 GHz but clock speed quickly drops down to roughly 3.2 GHz.  Hard to believe that hyperthreading would make up for that much difference in clock speed.  I need to install something like throttlestop and give it another go.

 

As far as whether or not you have a good chip, you'd need to show that it's stable before anyone could really tell you or not.  Cinebench isn't exactly a standard for stability.  It's just a benchmark.  Run OCCT or RealBench for a few hours and let us know how your OC handles it.

 

I for one could care less about clock speed.  The only thing that matter to me is how much the CPU produces.  If a 5 GHz chip encodes a video slower than a 4 GHz chip, I'll take the 4 GHz chip that does more work.

 

By the way, here are all the LTT Cinebench R15 submissions for the 4690k.  Your score isn't really inline with your clock speed.  I think your chip is throttling and clocking down during the benchmark, which would be a sign that you're pushing it too much.  I base this on the fact that the lowest score of 564 on this chart was done at 3.5 GHz which happens to be stock.

 

349	runit3	Intel Core I5 4690K	4.985ghz	765
354	TheGamingBarrel	Intel Core I5 4690K	4.869ghz	755
370	ProKoN	Intel Core I5 4690K	4.804ghz	736
379	harrynowl	Intel Core I5 4690K	4.8ghz	730
388	HazMatt	Intel Core I5 4690K	4.8ghz	726
394	TheGingerKid	Intel Core I5 4690K	4.713ghz	721
410	iyzyz	Intel Core I5 4690K	4.8ghz	707
414	Priller	Intel Core I5 4690K	4.6ghz	703
419	LSky	Intel Core I5 4690K	4.6ghz	701
422	TheElt	Intel Core I5 4690K	4.6ghz	698
423	legacy99	Intel Core I5 4690K	4.6ghz	697
426	Ragin Asian	Intel Core I5 4690K	4.6ghz	692
432	TehMasterSword	Intel Core I5 4690K	4.6ghz	687
434	App4that	Intel Core I5 4690K	4.6ghz	684
436	OCing Noob	Intel Core I5 4690K	4.509ghz	682
440	Micky Love	Intel Core I5 4690K	4.5ghz	681
453	Geralt	Intel Core I5 4690K	4.5ghz	668
454	SteelMK	Intel Core I5 4690K	4.5ghz	668
455	FilipSebik	Intel Core I5 4690K	4.5ghz	666
456	FirstimeFailboy	Intel Core I5 4690K	4.5ghz	664
461	SparkyRain	Intel Core I5 4690K	4.6ghz	660
468	Wyes	Intel Core I5 4690K	4.5ghz	645
469	Daring	Intel Core I5 4690K	4.5ghz	641
472	Nougatcookie	Intel Core I5 4690K	4.5ghz	638
476	Nalm	Intel Core I5 4690K	4.307ghz	635
482	Growlith1223	Intel Core I5 4690K	4.6ghz	631
533	lol1290	Intel Core I5 4690K	3.5ghz	564
417	harrynowl	Intel Core I5 4690K	4.7ghz	701

[WelshDdraig]

36 minutes ago, done12many2 said:

 

Spoiler

 

Wow.  That's what my laptop's 2760qm with 1600 MHz DDR3 scores.  552 to be exact.  The CPU starts the bench at 3.5 GHz but clock speed quickly drops down to roughly 3.2 GHz.  Hard to believe that hyperthreading would make up for that much difference in clock speed.  I need to install something like throttlestop and give it another go.

 

As far as whether or not you have a good chip, you'd need to show that it's stable before anyone could really tell you or not.  Cinebench isn't exactly a standard for stability.  It's just a benchmark.  Run OCCT or RealBench for a few hours and let us know how your OC handles it.

 

I for one could care less about clock speed.  The only thing that matter to me is how much the CPU produces.  If a 5 GHz chip encodes a video slower than a 4 GHz chip, I'll take the 4 GHz chip that does more work.

 

By the way, here are all the LTT Cinebench R15 submissions for the 4690k.  Your score isn't really inline with your clock speed.  I think your chip is throttling and clocking down during the benchmark, which would be a sign that you're pushing it too much.  I base this on the fact that the lowest score of 564 on this chart was done at 3.5 GHz which happens to be stock.

 

349	runit3	Intel Core I5 4690K	4.985ghz	765
354	TheGamingBarrel	Intel Core I5 4690K	4.869ghz	755
370	ProKoN	Intel Core I5 4690K	4.804ghz	736
379	harrynowl	Intel Core I5 4690K	4.8ghz	730
388	HazMatt	Intel Core I5 4690K	4.8ghz	726
394	TheGingerKid	Intel Core I5 4690K	4.713ghz	721
410	iyzyz	Intel Core I5 4690K	4.8ghz	707
414	Priller	Intel Core I5 4690K	4.6ghz	703
419	LSky	Intel Core I5 4690K	4.6ghz	701
422	TheElt	Intel Core I5 4690K	4.6ghz	698
423	legacy99	Intel Core I5 4690K	4.6ghz	697
426	Ragin Asian	Intel Core I5 4690K	4.6ghz	692
432	TehMasterSword	Intel Core I5 4690K	4.6ghz	687
434	App4that	Intel Core I5 4690K	4.6ghz	684
436	OCing Noob	Intel Core I5 4690K	4.509ghz	682
440	Micky Love	Intel Core I5 4690K	4.5ghz	681
453	Geralt	Intel Core I5 4690K	4.5ghz	668
454	SteelMK	Intel Core I5 4690K	4.5ghz	668
455	FilipSebik	Intel Core I5 4690K	4.5ghz	666
456	FirstimeFailboy	Intel Core I5 4690K	4.5ghz	664
461	SparkyRain	Intel Core I5 4690K	4.6ghz	660
468	Wyes	Intel Core I5 4690K	4.5ghz	645
469	Daring	Intel Core I5 4690K	4.5ghz	641
472	Nougatcookie	Intel Core I5 4690K	4.5ghz	638
476	Nalm	Intel Core I5 4690K	4.307ghz	635
482	Growlith1223	Intel Core I5 4690K	4.6ghz	631
533	lol1290	Intel Core I5 4690K	3.5ghz	564
417	harrynowl	Intel Core I5 4690K	4.7ghz	701

 

 

Thanks for the info. I'm just finishing a run of OCCT (Only 1 hour as I turn my PC off while I sleep).
With Cinebench, it has never recorded my overclock for some reason. it just always states 3.5GHz. But where you mentioned I maybe pushing it and its downclocking, I didn't know it would do that. i thought it would be fine as long as the voltages and the temps were fine - I'll look into that. I might bump the speeds down a little to see what happens, but its stable (For the moment.)

[done12many2]

1 minute ago, Alexp10v2 said:

37 minutes ago, done12many2 said:

Thanks for the info. I'm just finishing a run of OCCT (Only 1 hour as I turn my PC off while I sleep).

With Cinebench, it has never recorded my overclock for some reason. it just always states 3.5GHz. But where you mentioned I maybe pushing it and its downclocking, I didn't know it would do that. i thought it would be fine as long as the voltages and the temps were fine - I'll look into that. I might bump the speeds down a little to see what happens, but its stable (For the moment.)

 

 

You're welcome.  I like to overclock, but I need to see measurable improvement when i do so otherwise it's a wash.  That's why I benchmark so much.  I'm always tinkering with something and I need to know what that resulted in.  

[WelshDdraig]

8 minutes ago, done12many2 said:

 

You're welcome.  I like to overclock, but I need to see measurable improvement when i do so otherwise it's a wash.  That's why I benchmark so much.  I'm always tinkering with something and I need to know what that resulted in.  

I know its only 1 hour. I'll do a longer test while in lessons tomorrow, but it passed the 1 hour test with no errors.

[WelshDdraig]

@alihazzi7 Just re-ran my Cinebench R15. While my system was stable at 4.9GHz, the result was piss poor, so I reduced it to 4.8GHz, and got a score of 708cb. is that any better @done12many2

 

Spoiler

 

[done12many2]

2 minutes ago, Alexp10v2 said:

@alihazzi7 Just re-ran my Cinebench R15. While my system was stable at 4.9GHz, the result was piss poor, so I reduced it to 4.8GHz, and got a score of 708cb. is that any better @done12many2

 

 

  Hide contents

 

 

Let's see.  You reduced your overclock by 2% and gained a 22% increase in Cinebench R15.  I'd say that's great. 

 

Work on CPU cache and memory and watch it go up a bit more. 

 

Nice work! 

[Juggern0ttt]

4.7ghz on a i7 6700k. From what I've read I'm fairly lucky. 

[done12many2]

13 minutes ago, Juggern0ttt said:

4.7ghz on a i7 6700k. From what I've read I'm fairly lucky. 

 

 

Let this button be a part of your life.

 

 

[smokefest]

2 hours ago, Alexp10v2 said:

@alihazzi7 Just re-ran my Cinebench R15. While my system was stable at 4.9GHz, the result was piss poor, so I reduced it to 4.8GHz, and got a score of 708cb. is that any better @done12many2

 

  Reveal hidden contents

 

I highly doubt this OC is stable. But if, its awesome OP AS hell OC...

I had a 4690k with 1.31 vcore and was at 4.6 ghz

 

I suggest you run Realbench for 8 hours stress test, if it pass the test, then ur good, i doubt it will tho...

 

or

 

Launch bf1 and play for 3 hours , if no BSOD, ur also good, lol but again, i doubt you wil lnot BSOD for that OC, its quite too high for low voltage

[WelshDdraig]

38 minutes ago, smokefest said:

I highly doubt this OC is stable. But if, its awesome OP AS hell OC...

I had a 4690k with 1.31 vcore and was at 4.6 ghz

 

I suggest you run Realbench for 8 hours stress test, if it pass the test, then ur good, i doubt it will tho...

 

or

 

Launch bf1 and play for 3 hours , if no BSOD, ur also good, lol but again, i doubt you wil lnot BSOD for that OC, its quite too high for low voltage

I actually just lowered the voltage again, from 1.32 to 1.30

 

i'll run another stress test with OOCT as for some reason i can't get realbench to work

 

[smokefest]

Just now, Alexp10v2 said:

I actually just lowered the voltage again, from 1.32 to 1.30

 

i'll run another stress test with OOCT as for some reason i can't get realbench to work

 

Must stress test with more than 1 program, and real cpu game or application usage, not only synthetic

I cud get my 4690k stable in cinebench, and aida64,

But then I was gettin BSOD in bf1 after 1-2 hour of gaming. OC wasn't stable finally.. lolll

Good luck have fun

[done12many2]

1 minute ago, smokefest said:

Must stress test with more than 1 program, and real cpu game or application usage, not only synthetic

I cud get my 4690k stable in cinebench, and aida64,

But then I was gettin BSOD in bf1 after 1-2 hour of gaming. OC wasn't stable finally.. lolll

Good luck have fun

 

If it passes OCCT (CPU:OCCT tab with large data set) for even an hour, I'll be beyond impressed.

 

I agree with the multiple test approach.  You run into too many different CPU load conditions to say that an overclock is stable based off of any one test.  You may pass one, yet crash another.  It's about finding the correct voltage to pass everything.  When I stress test, I try to get it to crash.  Doing so exposes a weakness, which I'd rather know about when testing as opposed to being surprised in the middle of some important work.

[Juggern0ttt]

On 11/30/2016 at 9:05 AM, done12many2 said:

 

Let this button be a part of your life.

 

 

4.7ghz on a i7 6700k. From what I've read I'm fairly lucky. 

[done12many2]

32 minutes ago, Juggern0ttt said:

4.7ghz on a i7 6700k. From what I've read I'm fairly lucky. 

 

 

By posting that again, it's clear that you didn't understand my hint.  If you hit the "PrtScn" button located in the top right corner of your keyboard, you can copy/capture your screen and paste in into an editing application (Paint as an example).  The screen capture will be of much higher quality than your cell phone picture of a screen.

 

I've attached an example.