Intel's 8th Generation Core Family - Coffee Lake (LGA 1151, 6C/12T)

Where do you expect Core i7-8700K's Turbo to land?

  • 3.8/3.9 GHz

    Votes: 0 0.0%

  • 4.0/4.1 GHz

    Votes: 3 23.1%

  • 4.2/4.3 GHz

    Votes: 6 46.2%

  • 4.4/4.5 GHz

    Votes: 3 23.1%

  • 4.6/4.7 GHz

    Votes: 1 7.7%
  • Total voters
    13
  • Poll closed .
Fry's has the 8100 and 8400 in stock as well but they are charging a 10 dollar premium. I imagine in a month stock will be better, certainly by the time low end boards come out all pricing will be coming down to earth where it should be
 
misterbobby said:
He either has SSAA on in Tomb Raider or his system is jacked up. Essentially he is running at over 4k if he has SSAA maxed at 1440p. Of course at those settings the cpu is not going to make any difference and was a very stupid test to show.
Click to expand...
Yes, that was with SSAA 4x on and Direct x 11 for Tomb Raider. Still I expected more improvement in the other benchmarks compared to my old computer. At least cpu intensive games such as x plane showed a huge improvement and future games that take advantage of all the cores will run better.
 
nanotube said:
Yes, that was with SSAA 4x on and Direct x 11 for Tomb Raider. Still I expected more improvement in the other benchmarks compared to my old computer. At least cpu intensive games such as x plane showed a huge improvement and future games that take advantage of all the cores will run better.
Click to expand...
Well those other benchmarks are gpu benchmarks made to stress the gpu so of course there is not too much difference. There are some games that a 920 cant even maintain 60 fps in and if you test in normal conditions without crippling levels of SSAA/MSAA then you will see massive gains in most games.
 
nanotube said:
I finally upgraded my computer from an i7-920 to an i7-8700k. I did a few graphic benchmarks to show the differences between the two processors. 3dmark showed barely an increase in frame rates except for the cpu test, Superpositon benchmark showed about a 7% increase, and Rise of the Tomb Raider had a little increase in the overall frame rates but the min rates were a lot better in the geothermal section. These test were all done with no overclocking of the cpu or the graphics card on both systems The graphics card is a Gigabyte GTX 1080 G1, the 920 had 12 GB of ram and the 8700k had 16 GB of ram.. I thought I would get a little better increase in performance in these benchmarks but it doesn't look like it. The biggest increase in performance happened with X-plane 11 using Flyinside and Oculus. I went from about 20 FPS to around 55 FPS.
Here are the results:

Tomb Raider benchmark- all maximum settings except ambient occlusion on hbao+. Resolution 1440p.

i7-920
mountain peak 47.42 FPS (min 20.48, max 79.36)
syria 32.51 FPS (min 16.22, max 48.71)
geothermal 35.66 FPS (min 9.96, max 49.89)
overall 38.88 FPS

i7-8700k
mountain peak 53.08 FPS (min 23.22, max 111.53)
syria 33.26 FPS (min 22.43, max 62.71)
geothermal 34.97 FPS (min 23.75, max 49.75)
Overall 40.85 FPS

View attachment 40539 View attachment 40540 View attachment 40541 View attachment 40542 View attachment 40543
Click to expand...


You must be running at 8k resolution or your memory is DDR4-100mhz or god knows what in the 8700k.
I can understand the i7 920 overclocked was a nice beast but I mean it's literally 7 or 8 years old, 4 core vs 6.
Something isn't right with that, it makes no sense
(I do see your low fps in Geothermal over doubled)
 
AbRASiON said:
You must be running at 8k resolution or your memory is DDR4-100mhz or god knows what in the 8700k.
I can understand the i7 920 overclocked was a nice beast but I mean it's literally 7 or 8 years old, 4 core vs 6.
Something isn't right with that, it makes no sense
(I do see your low fps in Geothermal over doubled)
Click to expand...

Something's weird about nanotube's settings, because he states 1440p resolution but the GTX 1080 framerates he's seeing are more like the 4K benchmarks of Rise I've seen than the 1440p ones. Everything maxed out may include SSAA(which is really just running at higher resolution and downsampling) or some other ridiculously expensive effect.

Regardless, it's not really surprising when you are running at framerates below 60fps in a single player game that even a huge CPU upgrade doesn't change average framerates that much. DX11(if that's what he's running) also sees no benefit from quad -> six core in single player scenarios. Only DX12 can make use of 6 cores for rendering. CPU matters for minimums, and for higher framerates like 100+. If you want to consistently maintain 144fps for your 144hz monitor in competitive multiplayer games, for example, CPU matters A LOT.
 
Sancus said:
Something's weird about nanotube's settings, because he states 1440p resolution but the GTX 1080 framerates he's seeing are more like the 4K benchmarks of Rise I've seen than the 1440p ones. Everything maxed out may include SSAA(which is really just running at higher resolution and downsampling) or some other ridiculously expensive effect.

Regardless, it's not really surprising when you are running at framerates below 60fps in a single player game that even a huge CPU upgrade doesn't change average framerates that much. DX11(if that's what he's running) also sees no benefit from quad -> six core in single player scenarios. Only DX12 can make use of 6 cores for rendering. CPU matters for minimums, and for higher framerates like 100+. If you want to consistently maintain 144fps for your 144hz monitor in competitive multiplayer games, for example, CPU matters A LOT.
Click to expand...
He already said he was using SSAA x4 and DX11.

And DX12 or DX11 makes no performance difference as for as cores used in the benchmark and overall it's actually a couple fps slower with HT on than it is with it off on my 4770k.
 
At this stage of the game you have two competitors with a product line up from enthusiast to entry level that have compelling options capable of suiting gamer needs. I said it before I will make a much bigger gain going from a 980ti to 1080ti than I will going from a 4790K to 8700K.

As for 4K gaming, very little pushes that well enough for high FPS, nothing out the box and the 1080ti is taxed at that resolution, 4K is a niche market for now so Graphics technology still needs to catch up to that resolution.
 
misterbobby said:
And DX12 or DX11 makes no performance difference as for as cores used in the benchmark.
Click to expand...

DX12 sees a higher number of draw calls going from 4->6 cores, or at least it should if implemented properly. Of course the maximum number of draw calls for your CPU is irrelevant if you're doing 1440p SSAA x 4 on a GTX 1080, because the GPU is bottlenecking so ridiculously hard that there aren't that many draw calls that need to be made.

But my main point was that, in general, if you are driving your GPU so hard that it can't even spit out >60fps your CPU has hardly any work to do, especially in a relatively simple single player game like Rise of the Tomb Raider that is probably not doing that much with the CPU other than rendering and your typical single threaded main game logic.
 
Well I'm cherry! haha And not switches.

My Microcenter overpriced gouged and scalped $500.00 8700K is doing 5.1ghz stable at 1.295 volts. I ran out of time to keep testing it but I am going to push volts lower tonight when I get home from classes.

Cache is at 4400mhz

I am going to push voltage as low as possible and then see where she floats at 5.1 lowest voltage.

As far as 5.2 ghz I will try that once I hammer 5.1 ghz down.

I may have won the Si Lottery and if this is the case It is worth the price premium to me.

Using a Strix ITX.
 
OrangeKhrush said:
At this stage of the game you have two competitors with a product line up from enthusiast to entry level that have compelling options capable of suiting gamer needs. I said it before I will make a much bigger gain going from a 980ti to 1080ti than I will going from a 4790K to 8700K.

As for 4K gaming, very little pushes that well enough for high FPS, nothing out the box and the 1080ti is taxed at that resolution, 4K is a niche market for now so Graphics technology still needs to catch up to that resolution.
Click to expand...

You should consider a 5775c. 8700k gaming performance without the need for a whole platform upgrade. Throw in a 1080ti and you'd be set for another couple of years.

Or just sell all your stuff (sell me that board....so hard to find one that isn't priced to the moon) and upgrade for not a whole lot out of pocket. :)
 
wickedld9 said:
You should consider a 5775c. 8700k gaming performance without the need for a whole platform upgrade. Throw in a 1080ti and you'd be set for another couple of years.

Or just sell all your stuff (sell me that board....so hard to find one that isn't priced to the moon) and upgrade for not a whole lot out of pocket. :)
Click to expand...

Lots of options though i want my airsoft rifle
 
tangoseal said:
I hear that ... there is a .300 black out SBR conversion I want to do my Colt 6920 but these damn computers have hognobbed all my money I was going to dedicated to that project.
Click to expand...

I want to buy a SVD and a official PSO-1 scope, currently using a G18C and MP5
 
wickedld9 said:
You should consider a 5775c. 8700k gaming performance without the need for a whole platform upgrade. Throw in a 1080ti and you'd be set for another couple of years.

Or just sell all your stuff (sell me that board....so hard to find one that isn't priced to the moon) and upgrade for not a whole lot out of pocket. :)
Click to expand...

I thought about doing that myself but... the few 5775cs available are sold at a high price and then there's the fact that it isn't the king for all games, but only for some of them (for example Crysis 3 prefers a 4790k). Plus it's very picky about motherboards and a poor overclocker in general so it's a pretty risky buy nowadays. And anyway, recent Intel CPUs paired with fast DDR4 (3.6ghz+) will beat it easily, in absolutely every situation (not just games).

It was a great buy for gamers back in the days to be sure (I wish I had known better). But all the recent reviews I have seen putting it in a good light were comparing it with Skylake or Kaby running with sluggish DDR4.
 
I appear to have somewhat lost the silicon lottery. O well.

4.8 all cores is stable at 1.25 volts (Asus LLC setting 6, load voltage around 1.232). To hit 4.9 I need around 1.28, and 5.0 is somewhere above 1.3 (I stopped trying as I hit a cooling wal there - I am not delidded).

Not too sad though. 4.8 is pretty good, and 6 / 12 is sweet.
 
Cali3350 said:
I appear to have somewhat lost the silicon lottery. O well.

4.8 all cores is stable at 1.25 volts (Asus LLC setting 6, load voltage around 1.232). To hit 4.9 I need around 1.28, and 5.0 is somewhere above 1.3 (I stopped trying as I hit a cooling wal there - I am not delidded).

Not too sad though. 4.8 is pretty good, and 6 / 12 is sweet.
Click to expand...

4.8 is plenty fast and unless your running a benchmark you would never be able to tell the difference from 4.8 to 5.0.
 
Guy seems unhappy with his CaseKing?

1.41V 5 GHz premium chip lol.

StRWuA9.jpg
 
Hagrid said:
Yeah, he could of used some of the money to learn patience..... :)

1.41 is too high isn't it for 24/7?
Click to expand...

Not if you live in a Igloo, but for everyone else yeah that thing with thermally throttle unless you have extreme cooling like chilled water. Also at that voltage the lifespan of that chip will be much shorter.
 
1.4v 24/7 on a delidded watercooled cpu is perfectly fine, on water with a delid i've gone as high as 1.544v before without any damage or degradation on kaby lake, that certainly isn't something you'd want to run 24/7 but its very doable i've seen pppl push as far as 1.6v on ambient on hwbot for bench runs. the thing is past 1.4 ish theres big diminishing returns on what you can get clock wise without going sub ambient on cooling. that guy should only be mad at himself those cpus are sold as overclocking cpus not daily use ones he was even dumb enough to go with the ultra edition which has a pure silver IHS for that super duper slight increase in thermal conductivity lol.
 
Gideon said:
4.8 is plenty fast and unless your running a benchmark you would never be able to tell the difference from 4.8 to 5.0.
Click to expand...
So I encountered a weird issue I have not while overclocking before.

When stress testing, My CPU starts out at 4.8GHZ as expected however ultimately seems to come down to around 4.2-4.3ghz. Ive been racking my brain trying to figure out why as it does not appear to be thermal (temperatures are 65-70C at worst).

I have read that you sometimes have to change the CPU Current setting in the BIOS. It can apparently sometimes TDP limit the CPU when running. I will have to look into that tonight.
 
Cali3350 said:
So I encountered a weird issue I have not while overclocking before.

When stress testing, My CPU starts out at 4.8GHZ as expected however ultimately seems to come down to around 4.2-4.3ghz. Ive been racking my brain trying to figure out why as it does not appear to be thermal (temperatures are 65-70C at worst).

I have read that you sometimes have to change the CPU Current setting in the BIOS. It can apparently sometimes TDP limit the CPU when running. I will have to look into that tonight.
Click to expand...
Yes bios setting for tdp limit
 
TaintedSquirrel said:
Guy seems unhappy with his CaseKing?

1.41V 5 GHz premium chip lol.

StRWuA9.jpg
Click to expand...

OrangeKhrush said:
And he paid like 800 pounds for it, got to be seriously desperate to do that.
Click to expand...

There are three editions: Advanced, Pro, and Ultra. The Ultra series is aimed at people wanting to enter the world of professional overclockers.

The Ultra Edition is aimed at OC enthusiasts who are on the hunt for records and want to compete with the best overclockers!
Click to expand...

He had be well served with Advanded or Pro editions:

The Pro Edition is aimed at experienced overclockers and offers the largest selection of guaranteed OC clock rates [...] The Pro Edition is perfect for gaming PCs to get the maximum power out of the CPU.
Click to expand...


The user didn't even know what he purchased.
 
Last edited:
Frito11 said:
1.4v 24/7 on a delidded watercooled cpu is perfectly fine, on water with a delid i've gone as high as 1.544v before without any damage or degradation on kaby lake, that certainly isn't something you'd want to run 24/7 but its very doable i've seen pppl push as far as 1.6v on ambient on hwbot for bench runs.
Click to expand...

Also ultra edition has a silver heat spread...
 
Just built my SFF i7-8700 build. Bad news for anyone who uses the L9I. It's simply not a powerful enough cooler to not throttle with this CPU.

I've underclocked to 4.0GHZ ACT and it seems to have stabilized in the 80C range. Still, I need to find a better solution.
 
Revenant_Knight said:
Just built my SFF i7-8700 build. Bad news for anyone who uses the L9I. It's simply not a powerful enough cooler to not throttle with this CPU.

I've underclocked to 4.0GHZ ACT and it seems to have stabilized in the 80C range. Still, I need to find a better solution.
Click to expand...
You do mean the “non” k variant? Just clarifying
 
lmao 5ghz 1.41v

i thought 1.35v for 5ghz was already average..

looking for 1.3v for 5.1ghz if possible, super binned thanks babe
 
It said before he bought it what voltage, speed, and how much it was. It’s even printed on the chip. Same stupid people that buy 8700k’s for $700 off eBay. He got what he ordered.
 
What's even the point, the delid has no benefit when any random 8700K off the shelf will match that or beat it.
I wouldn't even pay MSRP for that chip.
 
You must log in or register to reply here.
Back
Top