Who's getting/got a 8086k?

[porina]

 

Just arrived. My sample serial number ends in 666. Tonight I will bench in hell!

[Kalm_Traveler]

45 minutes ago, porina said:

 

Just arrived. My sample serial number ends in 666. Tonight I will bench in hell!

Very nice! I ordered from Newegg within 2 minutes of the listing going live but they only shipped yesterday, and it won't arrive until Thursday, 6 days later... 

 

Really wish I would have ordered from Amazon and got it 1 day later. Oh well, still waiting on the radiator to show up. 

 

Be sure to let us know your results. I have a Rockitcool IHS and Conductonaut waiting for mine... 

[Princess Luna]

On 6/10/2018 at 4:39 AM, porina said:

They stated they're only making 50000 of them, and 8086 of them are to be used for the giveaway.

Just a small correction, they are re-labeling 50.000 i7 8700K's as i7 8086K

[porina]

1 minute ago, Kalm_Traveler1 said:

Very nice! I ordered from Newegg within 2 minutes of the listing going live but they only shipped yesterday, and it won't arrive until Thursday, 6 days later... 

I could have got one as early as Saturday if I fancied a 3 hour round trip. I didn't want it that badly...

1 minute ago, Kalm_Traveler1 said:

Be sure to let us know your results. I have a Rockitcool IHS and Conductonaut waiting for mine... 

I will probably be limited in time tonight, my first priority is multi-core scaling in Prime95 as I suspect the ram interface will bottleneck badly, but I never proved it on a 6 core ring bus CPU. It already does that on quad cores... once I'm happy it is working, then I'll delid+LM later in the week and go for max clocks.

[Christophe Corazza]

57 minutes ago, porina said:

Just arrived. My sample serial number ends in 666. Tonight I will bench in hell!

 

Nice one 

[porina]

I've ran out of time for one evening. Only have time for some quick impressions.

 

CPU is with stock TIM, not delid (yet), cooled by Noctua D15 so about as good as it gets on air. At +100mV offset I can play in the low 5.x GHz region and a quick try with more voltage decreased stability, suggesting thermal problems. Stock temps running P95 small FFT were mid 70's. 

 

Pure clock, I can get a hwbot submission in at 5.4, but 5.5 is instant crash. Don't ask about stability at 5.4, it only needed to survive long enough to save the result.

 

Cinebench R15, I can do repeat runs at 5.2 if I leave a good amount of time between runs, presumably as stuff cools. I wasn't leaving monitoring going since I wanted max bench performance. Running too quickly after another run often resulted in BSOD. After a few tries and tweaks, I couldn't improve on my score of 1720.

 

While I was testing at 5.1 I did have voltage and temp monitoring on, and that was into 1.4x (the mobo seems to have some adaptive voltage with clock I'll have to fix later) and temps during runs well into the 80's.

 

Using Prime95 I did some testing into compute workload scaling. Most here wont care, but comparing 8086k to 7800X, it confirmed my suspicion that the 8086k is crippled with only a dual channel ram configuration. If you keep within the L3 cache, the ring cache is more effective than the mesh cache of 7800X even when that was overclocked 50%. For large sizes exceeding on CPU cache, the quad channel on 7800X runs away with it.

 

Tomorrow, delid, liquid metal, and more fun...

[Christophe Corazza]

6 minutes ago, porina said:

CPU is with stock TIM, not delid (yet), cooled by Noctua D15 so about as good as it gets on air.

 

Do you consider more exotic cooling to puch the CPU even further?

[porina]

18 minutes ago, Christophe Corazza said:

Do you consider more exotic cooling to puch the CPU even further?

I think it'll be fun to see how far I get on air. There's only 3 Cinebench R15 submissions for this CPU on hwbot for now, but I managed to take the lead against the other two on water. Basically the D15 shouldn't be underestimated and it is comparable to 240mm AIOs if not more, even if it wont keep up with a full custom loop. I do have a 280mm AIO that needs a refill but I'm not convinced that'll buy me more than air would.

[Christophe Corazza]

22 minutes ago, porina said:

I think it'll be fun to see how far I get on air. There's only 3 Cinebench R15 submissions for this CPU on hwbot for now, but I managed to take the lead against the other two on water. Basically the D15 shouldn't be underestimated and it is comparable to 240mm AIOs if not more, even if it wont keep up with a full custom loop. I do have a 280mm AIO that needs a refill but I'm not convinced that'll buy me more than air would.

 

I'm eager to see how far you'll be able to get it tomorrow 

[AntiTrust]

9 hours ago, porina said:

Tomorrow, delid, liquid metal, and more fun...

Please post results if you have time, id be really interested to see how it went. 

[NMS]

Honestly, I'd love to get it just for its value, whatever it may be. But at that price, I'm gonna pass.

[porina]

1 minute ago, NMS said:

Honestly, I'd love to get it just for its value, whatever it may be. But at that price, I'm gonna pass.

Apart from the name and box, you'll probably get near enough the same out of a 8700k.

 

12 minutes ago, AntiTrust said:

Please post results if you have time, id be really interested to see how it went. 

I'll try to summarise delid results and bench scores here from time to time, but for a chance of seeing them sooner I'm mackerel on hwbot. http://hwbot.org/user/mackerel/

 

Since it is early days for the CPU, it is interesting seeing how others do on hwbot. I think I've decided, I want to see if I can be top on air. No chance of keeping up with LN2 users, but I'll give any non sub ambient a run for their money.

[NMS]

6 minutes ago, porina said:

Apart from the name and box, you'll probably get near enough the same out of a 8700k.

I don't plan on putting it in the system anyway:) But you are correct, the barely higher clock speed just wouldn't cut it for me. And honestly, for such an anniversary Intel could've released something much better.

[O9B0666]

2 hours ago, porina said:

Apart from the name and box, you'll probably get near enough the same out of a 8700k.

 

 

im stable at 1.30v w 5ghz...anything more like 5.1 im in the 1.4 range, its interesting seeing your results. Clearly this is a higher binner 8700k.

[O9B0666]

i would like to see delid results, air vs air i dont think a 8700k can go as far as a 8086, your numbers are nice with air, im using aio and in the 5.1 area im finally breaking 70c, with a delid that thing will be a monster.

[DarkSmith2]

From what ive heard and seen until now, the 8086k isnt any better than a good 8700k, but its really binned so easier to get a really good one, especially now after Intel most likely binned them out of 8700k production, i wouldnt buy a 8700k anymore unless you can make sure to get an older revision. Also have seen first 4,8GHz max. potatoes on the internet. So unfortunately there are still huge outliers with the 8086k.

 

Im still happy with my 8700k as far as core clock goes, can do 5.2 safe, maybe 5.3 on water? But im curious that i might have a bad IMC, tried RAM OC, and cant get past DDR4-3866, might be the board, might be my RAM-Kit or my CPUs IMC, dont know how to find out. I can do 3600 CL14 thats for sure.. im lacking a bit of RAM OC knowledge to know if that means anything or not.

[porina]

1 hour ago, O9B0666 said:

im stable at 1.30v w 5ghz...anything more like 5.1 im in the 1.4 range, its interesting seeing your results. Clearly this is a higher binner 8700k.

One thing I don't like about the mobo I'm using is it likes to do its own thing on voltage... although I have been lazy and left it on auto, I didn't expect it to still adjust when I change multipliers in software as opposed to bios.

5 minutes ago, DarkSmith2 said:

Im still happy with my 8700k as far as core clock goes, can do 5.2 safe, maybe 5.3 on water? But im curious that i might have a bad IMC, tried RAM OC, and cant get past DDR4-3866, might be the board, might be my RAM-Kit or my CPUs IMC, dont know how to find out. I can do 3600 CL14 thats for sure.. im lacking a bit of RAM OC knowledge to know if that means anything or not.

I find higher speeds to be hit and miss, and you're doing better than I have previously. With a 3000C14 B-die kit I can do 3600C16 or 3866C17 but not managed to lower latencies on either.

[DarkSmith2]

4 minutes ago, porina said:

I find higher speeds to be hit and miss, and you're doing better than I have previously. With a 3000C14 B-die kit I can do 3600C16 or 3866C17 but not managed to lower latencies on either.

yea well i bought a 3600 cl15 kit. Maybe its slightly binned, but its just not satisfying to do RAM OC, when i bought it i thought i gonna do 4133 CL17 easy, seems i was a little optimistic, only changing small things when OC'ing just isnt fun. I also cant exclude that im to stupid for RAM OC.. so well yeah, more fiddling is required i think. Maybe should try something like 4000 with cl19 or so to see if that boots...

[porina]

3 minutes ago, DarkSmith2 said:

yea well i bought a 3600 cl15 kit. Maybe its slightly binned, but its just not satisfying to do RAM OC, when i bought it i thought i gonna do 4133 CL17 easy, seems i was a little optimistic, only changing small things when OC'ing just isnt fun. I also cant exclude that im to stupid for RAM OC.. so well yeah, more fiddling is required i think. Maybe should try something like 4000 with cl19 or so to see if that boots...

I couldn't find any stability at 4000 by relaxing primary timings and generous volts. Maybe something could be done with secondary/tertiary timings but don't have long enough to fiddle with those too!

[DarkSmith2]

48 minutes ago, porina said:

I couldn't find any stability at 4000 by relaxing primary timings and generous volts. Maybe something could be done with secondary/tertiary timings but don't have long enough to fiddle with those too!

thats why, its the easiest to just copy paste timings from Raja's Apex profile. 

 

And i came this far yesterday on my own (just changing timings and CR):

[O9B0666]

5 hours ago, porina said:

One thing I don't like about the mobo I'm using is it likes to do its own thing on voltage..

lmao do you have the same one i have? mine was doing that too, still does haha 

[arnavvr]

I was considering getting one until I realized that Z390 and Ice Lake is coming out soon. My 7700K @5.1 is fast enough anyways, and I'd rather have double the core count. 

[porina]

15 minutes ago, O9B0666 said:

lmao do you have the same one i have? mine was doing that too, still does haha 

Mine is an Asrock, yours is Gigabyte according to sig? I do expect varying voltages with auto in bios, but not in XTU for example but it seems to be happening. I don't see that on my Asus boards.

[porina]

I did the delid and LM application. Before and after, Prime95 12x64k FFT core temps went from 76-79 to around mid 60's. So... around 10C+ drop there.

 

5.2 seems bench stable at +150mV offset in XTU. I didn't get a crash during the bench session under that condition. I still need to work out what actual voltage the mobo is supplying and fix it, as voltage monitoring software seems erratic. I see reported peaks of 1.48v which is on the high side, but under load it settles to around 1.36v. I've done a small selection of benches on hwbot, but my XTU score sucks compared to others (~2100 at 5.2) so I haven't uploaded it yet. Can't figure out why...

 

Can I push clock further?... not really. 5.3 was not stable enough to complete Cinebench R15 even with +250mV offset.

[Christophe Corazza]

4 minutes ago, porina said:

Can I push clock further?... not really. 5.3 was not stable enough to complete Cinebench R15 even with +250mV offset.

 

So that means it's over and out in terms of increasing the clock speed?