Intel Coffee Lake Review Thread

[Majestic]

Just now, DocSwag said:

Don't you love it when I provide a legitimate source and then instead of refuting it, you think it's funny?

Fam I delid myself, don't even. Even when you reapply similar w/mk paste, being 5.6-ish, it's still has a shitty heat transfer rate at high voltages and you require heavy cooling.

5.6 w/mk vs. >80 w/mk of solder. Yeah no, it's shit dude.

[TidaLWaveZ]

30 minutes ago, DocSwag said:

Why is this boring?

Because it's not the company they show their undivided loyalty to.

[JoeCoke]

Aaaaand Intel has beaten Ryzen. 8600k beats the 1600x and even the 1700 in most cases, so dayum. I only wonder if the Ryzen refresh coming in February with slight clock bumps and maybe slight ipc bumps will be able to level the playing field.

 

Now I have to really think about my upgrade... Intel is better now no question, but we don't know what the Ryzen refresh is going to do and AM4 will be supported until 2020. Sooooo I'm just gonna wait and see what happens next before I upgrade.

[Majestic]

1 minute ago, Cookybiscuit said:

This is the shit kicker.

 

Bad enough are the sky high temps, even worse is the fact that you can't even fucking get one.

The only reason they're doing this is because AMD is selling chips. The chips Intel wants to sell. They're hoping people will now wait for them to become available.

Given the amount of zealots, i'd say Intel's mindshare games are working.

[handymanshandle]

6 minutes ago, Od1sseas said:

On Geekbench Ryzen 3 1200

Single-Core: 3400

Multi-Core: 9300

 

while i3 8100 scores

Single-Core: 4600

Multi-Core: 13600

 

The i3 8100 has the same performance as the i5 7400

 

Overclocking the Ryzen 3 1200 to 3.9 Ghz we getting kinda the same scores as the i3 8100

 

Intel destroyed AMD

4 minutes ago, PCGuy_5960 said:

Really? Use a real benchmark, like Cinebench. Geekbench sucks.

We aren't taking into account value/performance.

 

Yes Intel has always been able to beat AMD on performance when AMD can make something that performs well.

 

Reasonably Intel made a response, the performance we're getting to me at least is just something to be expected.

 

 

 

[thorhammerz]

17 minutes ago, DocSwag said:

As for "shitty tim," the TIM Intel uses is actually very high quality stuff. 

Performance-wise, it's nothing special at 8.0 w/m.k. (that's slightly worse than the Arctic MX-4, which is itself fairly lacklustre).

 

Compared to something like CLU (at about 38w/m.k.) or Thermal Grizzly Conductonaut (at about ~73w/m.k.), it falls rather flat (and thermal conductivity of the paste between the die and IHS matters a LOT more in comparison to the paste you might put between the IHS and the HSF) at about 10-20% of liquid metal's (pure indium's w/m.k. is 86, so props to TG for making a commercial product that can provide similar results!) performance.

 

Intel uses it because of its high durability and reliability (if you don't break the material, of course) over the course of Intel's 3 year warranty, not for its raw thermal conductive performance.

[TidaLWaveZ]

7 minutes ago, Majestic said:

The only reason they're doing this is because AMD is selling chips. The chips Intel wants to sell. They're hoping people will now wait for them to become available.

Given the amount of zealots, i'd say Intel's mindshare games are working.

Not quite.

 

Ryzen isn't making the large scale sales(pre configured PC's and laptops for vendors like HP,Lenovo,Dell,Asus,Apple,etc.) The sales they are making are limited to budget techies looking to build in the last three months or so.

[PCGuy_5960]

2 minutes ago, Majestic said:

5.6 w/mk vs. >80 w/mk of solder. Yeah no, it's shit dude.

He never said that it's better than solder though. He simply said that the TIM is good enough (unless you are using 1.45+V) 

 

And let's ask @done12many2 and @MageTank. Do you guys think that the paste is the issue or do you think that the adhesive is the issue? 

[Lathlaer]

14 minutes ago, Princess Cadence said:

Linus videos sucks, when I watched it I feared that the i7 8700k could be disappointing somehow as their gaming benchmarks showed the i7 7700k was a tiny bit ahead, then every single other video shows the Coffee Lake easily taking the lead and now I feel much better with my soon to become true upgrade [:

Yeah, I have been wondering how they managed to have virtually the only review where 7700k is ahead in games.

[Majestic]

Just now, PCGuy_5960 said:

He never said that it's better than solder though. He simply said that the TIM is good enough (unless you are using 1.45+V) 

 

And let's ask @done12many2 and @MageTank. Do you guys think that the paste is the issue or do you think that the adhesive is the issue? 

No he said it's high quality, it's not. It's only very durable. But it has a very mediocre heat conduction. 

[goodtofufriday]

As an AMD Fanboy, im more willing to attribute intels new processers to AMDs success. Since without it Intel likely would have kept the status quo. 

 

That being said its good to see the tict for tact. It means both intel and amd with try to out do each other. Just hopefully this time intel doesnt play so dirty that every country starts to sue them again. 

[PlayStation 2]

I'm watching the DigitalFoundry video and I'm impressed at how it's neck and neck with the R7 1800X in the Handbrake h.264 encode. Also seems to handle minimum frame times far better than the 7700K.

Also, Richard's hand movements.

Edited October 5, 2017 by Dan Castellaneta
Whoops, not HEVC; was h.264

[Dylanc1500]

1 minute ago, Majestic said:

No he said it's high quality, it's not. It's only very durable. But it has a very mediocre heat conduction. 

It doesn't necessarily have to have high heat conductivity in order to be considered high quality. He is correct from a quality standpoint in that it is good. As has been stated, Intel uses it for a good enough thermal purpose, for the compromise of having a higher level of durability. Solder has its use cases, and if you NEED the upmost in thermal conductivity then there are options out there for you, but it isn't the end all be all.

[Majestic]

3 minutes ago, Dylanc1500 said:

It doesn't necessarily have to have high heat conductivity in order to be considered high quality. He is correct from a quality standpoint in that it is good. As has been stated, Intel uses it for a good enough thermal purpose, for the compromise of having a higher level of durability. Solder has its use cases, and if you NEED the upmost in thermal conductivity then there are options out there for you, but it isn't the end all be all.

For locked consumer-grade chips, MAYBE. Not for their K SKU's or their HEDT chips. Gamers Nexus has expressed their dismay about it aswell, using the same source.

[done12many2]

16 minutes ago, PCGuy_5960 said:

He never said that it's better than solder though. He simply said that the TIM is good enough (unless you are using 1.45+V) 

 

And let's ask @done12many2 and @MageTank. Do you guys think that the paste is the issue or do you think that the adhesive is the issue? 

 

Man, just let the guy continue on with his pubic service announcement.  It's just easier that way.  

[Ezio Auditore]

37 minutes ago, Majestic said:

And i'm sure that was by complete accident.

Yes because i'm running SFF. And interesting how this used to be the big point people brought up when it was about Intel/Nvidia.

Even if you are running SFF I see no reason why you would buy according to the perf/watt graph. You would check the power that can be provided by your psu (that you can afford) and buy the processor accordingly.

[Notional]

Meanwhile in Norway:

 

 

 

Seriously, I get there are not a whole lot of chips, and we do see some price gouging, but this is just an outright joke. Yet again Intel acts like they are still a monopoly. No one in their right mind would buy these chips at these prices.

[Majestic]

Just now, Ezio Auditore said:

Even if you are running SFF I see no reason why you would buy according to the perf/watt graph. You would check the power that can be provided by your psu (that you can afford) and buy the processor accordingly.

Thermals? Delta-T? Had I not delidded my 4670K, it would be running 85-90C. Now it's 65-70C. And lower perf./watt means more heatoutput into the case for the same performance. Hence I'm also using pascal and not polaris. Sure SFF is a niche, but I never generalised. 

 

And hidden costs like powerusage add up when you're living in western europe at 22-23 eurocent per KHW.

[mynameisjuan]

5 minutes ago, Notional said:

Seriously, I get there are not a whole lot of chips, and we do see some price gouging, but this is just an outright joke. Yet again Intel acts like they are still a monopoly. No one in their right mind would buy these chips at these prices.

MSRP is intels price, what stores mark them up to has nothing to do with intel so dont point your fingers at them. The 8700k can be picked up on newegg for $379 which is a good price. 

[DocSwag]

2 minutes ago, Majestic said:

Fam I delid myself, don't even. Even when you reapply similar w/mk paste, being 5.6-ish, it's still has a shitty heat transfer rate at high voltages and you require heavy cooling.

5.6 w/mk vs. >80 w/mk of solder. Yeah no, it's shit dude.

You can't just base measurements purely off of thermal conductivity. For example, the difference between liquid metal and normal thermal paste when it comes to thermal paste on the CPU doesn't make much difference at all.

 

As well, there are legitimate reasons why you can't just use solder on any CPU, such as the solder cracking.

[i_build_nanosuits]

look at that, 179$ for a 4.0ghz quad-core CPU on intel's latest architecture...for a lower cost gaming system, this is an amazing offering IMHO:

 

https://www.newegg.com/Product/Product.aspx?Item=N82E16819117823&ignorebbr=1&nm_mc=AFC-C8Junction&cm_mmc=AFC-C8Junction-PCPartPicker, LLC-_-na-_-na-_-na&cm_sp=&AID=10446076&PID=3938566&SID=

 

...and did i mentionned it's unlocked and can most likely run at 5.0ghz+...and shit all over Ryzen in games

 

[Ezio Auditore]

8 minutes ago, Majestic said:

Thermals? Delta-T? Had I not delidded my 4670K, it would be running 85-90C. Now it's 65-70C. And lower perf./watt means more heatoutput into the case for the same performance. Hence I'm also using pascal and not polaris. Sure SFF is a niche, but I never generalised. 

 

And hidden costs like powerusage add up when you're living in western europe at 22-23 eurocent per KHW.

A mobile cpu has higher performance/watt than an avg. Desktop cpu but you didn't buy the mobile one, did you?

[TheOriginalHero]

35 minutes ago, TidaLWaveZ said:

Not quite.

 

Ryzen isn't making the large scale sales(pre configured PC's and laptops for vendors like HP,Lenovo,Dell,Asus,Apple,etc.) The sales they are making are limited to budget techies looking to build in the last three months or so.

Thats mostly due to the lack of an igpu and it being a completely  new architecture for amd. I believe many will start using the apus in their systems once released.

[Majestic]

4 minutes ago, DocSwag said:

You can't just base measurements purely off of thermal conductivity. For example, the difference between liquid metal and normal thermal paste when it comes to thermal paste on the CPU doesn't make much difference at all.

 

As well, there are legitimate reasons why you can't just use solder on any CPU, such as the solder cracking.

What? It's an easy -15/20 degrees when you apply it on the DIE->IHS. If you're talking about between the heatspreader and heatsink, that's because you're using a much larger surface area compared to the DIE->IHS. Ofcourse it's going to be more effective at smaller surface area's.

 

And if you're talking about applying it to the LID-> Heatsink on a stock chip, the bottleneck is not elleviated, so ofcourse you won't notice a difference. Honestly, I feel like you've watched Linus' video on this and bought into his horseshit. No experience, whatsoever.

 

2 minutes ago, Ezio Auditore said:

A mobile cpu has higher performance/watt than an avg. Desktop cpu but you didn't buy the mobile one, did you?

Yeah silly me for not cracking open a laptop, and frankensteining the mobile chip into an ITX case. This makes no sense, sorry.

[Misanthrope]

48 minutes ago, TVwazhere said:

Gotta include cost of system too

 

Personally I'll shell out more for the higher IPC and single core speeds for my applications

And that would matter mostly only on AAA games you want to run at 144hz or higher. Slower games or day to day stop you wouldn't be able to pick them apart on a Pepsi challenge.

 

Again this "better for gaming" is mostly for silly shit like 1080p gaming on a 1080ti just so you can get 180FPS instead of 130 on the most extreme cases.