Who's getting/got a 8086k?

[Christophe Corazza]

1 minute ago, DarkSmith2 said:

and he also forgot potentially a new RAM kit that clocks higher  

 

That is easily solved by going for RGB RAM 

[porina]

27 minutes ago, Christophe Corazza said:

Overclocking CPUs seems to be a very expensive hobby 

I mean... look at @porina... simply to overclock only 1 CPU 

Overclocking in itself isn't necessarily expensive. I could stop where I am now, and 6 cores at 5.2 isn't doing badly!

 

Competitive overclocking is another matter...

25 minutes ago, DarkSmith2 said:

and he also forgot potentially a new RAM kit that clocks higher  

I didn't forget. The system currently has 2x8GB TridentZ 3000C14. Now, that's not a fast kit, but it is B-die. I have previously overclocked it to 3866C17 so far, but I'd like to try lowering the latency some more. I think this is a weakness preventing me from reaching the upper levels in competitions.

20 minutes ago, Christophe Corazza said:

That is easily solved by going for RGB RAM 

My 7800X system is my RGB unicorn vomit theme system...  

[Christophe Corazza]

2 minutes ago, porina said:

Overclocking in itself isn't necessarily expensive. I could stop where I am now, and 6 cores at 5.2 isn't doing badly!

 

Competitive overclocking is another matter...

 

Fron your previous posts, I do sense some competitiveness when you're pushing the system to the max 

 

4 minutes ago, porina said:

My 7800X system is my RGB unicorn vomit theme system...  

 

Then why aren't you getting at least a 30 % performance increase? 

[DarkSmith2]

9 minutes ago, porina said:

I didn't forget. The system currently has 2x8GB TridentZ 3000C14. Now, that's not a fast kit, but it is B-die. I have previously overclocked it to 3866C17 so far, but I'd like to try lowering the latency some more. I think this is a weakness preventing me from reaching the upper levels in competitions.

well im struggling to get my kit past 3866CL17, well i can do 4000CL19 with 1.45v but thats it, anything better would take more voltage than i would like to use. Despite being samsung B-Die higher kits are just binned (well atleast those really high kits). Much less lottery involved. But you cant be sure if its not your IMC or Board that is limiting. Or maybe im just doing something wrong with RAM OC in general, but im getting really not the results im looking for. With 4133CL17 i would be happy for 24/7 use.

[porina]

50 minutes ago, Christophe Corazza said:

Then why aren't you getting at least a 30 % performance increase? 

I haven't calibrated the RGB ram to full rainbow yet.

47 minutes ago, DarkSmith2 said:

well im struggling to get my kit past 3866CL17

I'm happy enough to reach 3866 but if I can chip away at the timings that should help a lot. I've done secondary and tertiary tuning but they're not 100% stable, so can't be used for 24/7 running, but ok for lesser benchmarks. Note had any luck with lowering primary, and on the mobo I have it is a major pain to recover from unbootable state. The clr-cmos jumper is hard to reach, and the self recovery isn't great.

[DarkSmith2]

8 minutes ago, porina said:

I haven't calibrated the RGB ram to full rainbow yet.

I'm happy enough to reach 3866 but if I can chip away at the timings that should help a lot. I've done secondary and tertiary tuning but they're not 100% stable, so can't be used for 24/7 running, but ok for lesser benchmarks. Note had any luck with lowering primary, and on the mobo I have it is a major pain to recover from unbootable state. The clr-cmos jumper is hard to reach, and the self recovery isn't great.

before bios update i had to uninstall my GPU to reach the mobos Battery and remove it because clrcmos and safeboot didnt worked properly maybe 1 out of 3 times it worked. I know the pain  

[Christophe Corazza]

20 minutes ago, porina said:

I haven't calibrated the RGB ram to full rainbow yet.

 

That explains everything :-)

[porina]

Silicon lottery have their binned 8086k for sale now: https://siliconlottery.com/collections/coffeelake

As of 6/17/18, the top 14% of tested 8086Ks were able to hit 5.3GHz or greater.

As of 6/17/18, the top 60% of tested 8086Ks were able to hit 5.2GHz or greater.

As of 6/17/18, the top 92% of tested 8086Ks were able to hit 5.1GHz or greater.

As of 6/17/18, 100% of tested 8086Ks were able to hit 5.0GHz or greater.

 

For comparison to 8700k, following tweet summaries some recent stats.

 

Compared to the most recent 8700k numbers, it does seem the 8086k gets you as good a chance of 5.2 as the 8700k does at 5.1, likewise 5.1 to 5.0 respectively. So, the price difference buys you perhaps 100 MHz of OC headroom on average.

[DarkSmith2]

siliconlottery.com statistics are out. they look like this:

 

if you compare the percentages, we can guess that its a 1 tier shift. Note that the sample size of the 8700k is much bigger.

[DarkSmith2]

god damnit you had been faster

[porina]

33 minutes ago, DarkSmith2 said:

god damnit you had been faster

You did what I was too lazy to do and table it for easier comparison  I've given up on getting mine past 5.2 short of going sub-ambient, but that'll be a battle for another day.

[DarkSmith2]

2 minutes ago, porina said:

You did what I was too lazy to do and table it for easier comparison  I've given up on getting mine past 5.2 short of going sub-ambient, but that'll be a battle for another day.

well that, and mysteriously my internet connection sucks today. Well 5.2 is very nice either way you can still improve your score via RAM/Cache alot.

[Christophe Corazza]

45 minutes ago, porina said:

Silicon lottery have their binned 8086k for sale now: https://siliconlottery.com/collections/coffeelake

As of 6/17/18, the top 14% of tested 8086Ks were able to hit 5.3GHz or greater.

As of 6/17/18, the top 60% of tested 8086Ks were able to hit 5.2GHz or greater.

As of 6/17/18, the top 92% of tested 8086Ks were able to hit 5.1GHz or greater.

As of 6/17/18, 100% of tested 8086Ks were able to hit 5.0GHz or greater.

 

For comparison to 8700k, following tweet summaries some recent stats.

 

Compared to the most recent 8700k numbers, it does seem the 8086k gets you as good a chance of 5.2 as the 8700k does at 5.1, likewise 5.1 to 5.0 respectively. So, the price difference buys you perhaps 100 MHz of OC headroom on average.

 

Looking at those numbers... I would say that you still have reached a nice OC using air cooling and an Asrock Z370 Pro4.

However, siliconlottery.com uses liquid cooling. Perhaps a rematch with liquid cooling? 

[porina]

48 minutes ago, Christophe Corazza said:

Looking at those numbers... I would say that you still have reached a nice OC using air cooling and an Asrock Z370 Pro4.

However, siliconlottery.com uses liquid cooling. Perhaps a rematch with liquid cooling? 

I do have high end air, which is comparable to lower-mid range water anyway. I don't have spare wc kit and I don't want to get another AIO... I do have a broken AIO I might try to revive some day. It was already non-functional due to long term liquid loss. In looking to refurbish it, and change the tubing, I broke one of the plastic barbs on the block trying to get the hose off so not sure it'll be a simple fix.

[DarkSmith2]

20 minutes ago, porina said:

I do have high end air, which is comparable to lower-mid range water anyway. I don't have spare wc kit and I don't want to get another AIO... I do have a broken AIO I might try to revive some day. It was already non-functional due to long term liquid loss. In looking to refurbish it, and change the tubing, I broke one of the plastic barbs on the block trying to get the hose off so not sure it'll be a simple fix.

Yea AiO's suck, thats why im aiming to get a full custom loop.

[Christophe Corazza]

24 minutes ago, porina said:

I do have high end air, which is comparable to lower-mid range water anyway. I don't have spare wc kit and I don't want to get another AIO... I do have a broken AIO I might try to revive some day. It was already non-functional due to long term liquid loss. In looking to refurbish it, and change the tubing, I broke one of the plastic barbs on the block trying to get the hose off so not sure it'll be a simple fix.

 

Althugh I truly love the look of a full custom water loop, I think I'm going to stick woth my AIO.

It is too much of a hassle when building the loop and especially during maintanance of your PC.

 

8 minutes ago, DarkSmith2 said:

Yea AiO's suck, thats why im aiming to get a full custom loop.

 

What CPU cooler do you have right now? An air cooler I presume?

[DarkSmith2]

17 minutes ago, Christophe Corazza said:

 

Althugh I truly love the look of a full custom water loop, I think I'm going to stick woth my AIO.

It is too much of a hassle when building the loop and especially during maintanance of your PC.

 

 

What CPU cooler do you have right now? An air cooler I presume?

No, i just bought an AiO (alphacool Eisbaer LT360) when i upgraded because i didnt want to have the mechanical burden of my aircooler (1.2KG) on my new expansive Motherboard, i used it for 2weeks waiting for the AiO though. 

 

Temp difference was 6°C. Now i turned the AiO cooler/pump orientation to "goofy" So that its hoses are coming out on the upper part of it just to reduce noise/be sure the pump gets feeded with coolant, weakend the temps by 3°C (because of the fin orientation on the copper plating). So all in all i have an advantage of maybe 3-5°C left with my 360mm radiator AiO with 24/7 settings.

 

When it comes to higher vcore and max overclock the initial difference was a bit higher, max. 8°C if irc. But its also ~900-1000RPM with a max of 1500RPM on my 3x 120mm noctua black.swap chromax fans vs 2x the industrial 140mm variant with 900-1000RPM and a max of 2000rpm and a much higher static pressure. To be fair neither of those fans were used at max RPM. 

[O9B0666]

this is sort of off topic, but sort of not idk...

I can run cinebench at 5.2 multi and single with my 8700k, now i see some of these 8086k's with scores around 1750, which my 5.2's are just breaking 1700...what else goes into the scoring system for this? As technically they are the same chip except a 300mhz factory difference, but oc'd at the same speeds there's a decent difference in scoring.

[DarkSmith2]

22 minutes ago, O9B0666 said:

this is sort of off topic, but sort of not idk...

I can run cinebench at 5.2 multi and single with my 8700k, now i see some of these 8086k's with scores around 1750, which my 5.2's are just breaking 1700...what else goes into the scoring system for this? As technically they are the same chip except a 300mhz factory difference, but oc'd at the same speeds there's a decent difference in scoring.

Its pretty simple, cinebench is purely measuring the time your PC needs to render the image from start to finish. Afaik its not even real time, its windows time, you can actually even cheat a little by running the benchmark multiple times or using different windows versions / using the PC offline.

 

The reason for having only small gains with higher frequencies are diminishing returns, bandwidth bottlenecking and so on.

 

 

highest ive seen was 1690 score at only 5GHz without changing anything RAM related. Just by rerunning.

[O9B0666]

13 minutes ago, DarkSmith2 said:

Its pretty simple, cinebench is purely measuring the time your PC needs to render the image from start to finish. Afaik its not even real time, its windows time, you can actually even cheat a little by running the benchmark multiple times.

 

The reason for having only a small gains with higher frequencies are diminishing returns, bandwidth bottlenecking and so on.

 

 

highest ive been was 1690 score at only 5GHz without changing anything RAM related.

my ram has no XMP, i cheaped out for cheaper stuff otherwise no complaints about it atm. So my suspicion was right that other parts do factor into this then. If i knew how to oc ram id give it a try. Huh at 5.0ghz your score is higher then mine, mine is 1664.

[DarkSmith2]

3 minutes ago, O9B0666 said:

my ram has no XMP, i cheaped out for cheaper stuff otherwise no complaints about it atm. So my suspicion was right that other parts do factor into this then. If i knew how to oc ram id give it a try. Huh at 5.0ghz your score is higher then mine, mine is 1664.

well my RAM is running @ddr4-3600CL15 (basically XMP profile) But ill do an experiment. BRB!

[DarkSmith2]

15 minutes ago, DarkSmith2 said:

well my RAM is running @ddr4-3600CL15 (basically XMP profile) But ill do an experiment. BRB!

@O9B0666 so i just fiddled a bit with RAM timings and its pretty close to 1750 @5.2 GHz already, just ran the test twice.

[O9B0666]

hmm now i have to try these runs again myself. I bet it might be because i have 2666mhz patriot signature ram too

[DarkSmith2]

9 minutes ago, O9B0666 said:

hmm now i have to try these runs again myself. I bet it might be because i have 2666mhz patriot signature ram too

Can you link me your exact RAM kit? Or do you know if its Samsung B-Die? (probably not)

[O9B0666]

2 minutes ago, DarkSmith2 said:

Can you link me your exact kit? Or do you know if its Samsung B-Die? (probably not)

https://www.amazon.com/Patriot-Signature-2666MHz-PC4-21300-Channel/dp/B07BRWRFC8/ref=sr_1_3?tag=sciestud09-20&s=electronics&ie=UTF8&qid=1527988432&sr=1-3&keywords=16gb+ddr4+2666