My wife had an issue with her RX470 failing to start games. Turns out, it had nothing to do with AMD - it was Win10 had updated the Intel IGP driver update and that was causing the problem.
No IPC gain over Kaby - just clock bump and core count improvements. Does appear to have more overclock potential though.
I think it will be an interesting 12 months ahead.
Intel damage limitation: 1. Reduce prices, get rid of some of the artificial restrictions on Pentiums, i3s etc. 2. More price cuts and unusually early announcements about Coffee Lake. 3. Get Coffee Lake with all the features we can include by Q4. Position as Ryzen equal.
Intel warp drive part - so long tic-tac-toe! 1. Coffee Lake - obviously but this was pretty much happening. 2. Cannon Lake - a 10nm Coffee Lake test bed? 8th gen tag suggests this. 3. Ice Lake. This is the jump to warp and what Intel hope (?) to position as the "Ryzen killer".
Also expect AMD to maybe accelerate the introduction of Raven Ridge (integrated graphics) as part of their response.
At the end of the day Intel are a very big company and can - if they choose - "throw money" at things. Clearly - having moved from tick-tock to tic-tac-toe they would prefer not to but Kaby, i9 Slylake, Coffee, Cannon and Ice Lake .... if the latter two release as planned .... is a significant increase in what we had 12 months ago and should herald an exciting 12 months. Maybe not for motherboard m/fs - reports surfacing that they cannot cope which is why we may not be getting backward compatibility.
No IPC gain over Kaby - just clock bump and core count improvements. Does appear to have more overclock potential though.
I think it will be an interesting 12 months ahead.
Intel damage limitation: 1. Reduce prices, get rid of some of the artificial restrictions on Pentiums, i3s etc. 2. More price cuts and unusually early announcements about Coffee Lake. 3. Get Coffee Lake with all the features we can include by Q4. Position as Ryzen equal.
Intel warp drive part - so long tic-tac-toe! 1. Coffee Lake - obviously but this was pretty much happening. 2. Cannon Lake - a 10nm Coffee Lake test bed? 8th gen tag suggests this. 3. Ice Lake. This is the jump to warp and what Intel hope (?) to position as the "Ryzen killer".
Also expect AMD to maybe accelerate the introduction of Raven Ridge (integrated graphics) as part of their response.
At the end of the day Intel are a very big company and can - if they choose - "throw money" at things. Clearly - having moved from tick-tock to tic-tac-toe they would prefer not to but Kaby, i9 Slylake, Coffee, Cannon and Ice Lake .... if the latter two release as planned .... is a significant increase in what we had 12 months ago and should herald an exciting 12 months. Maybe not for motherboard m/fs - reports surfacing that they cannot cope which is why we may not be getting backward compatibility.
If Cannon Lake released as planned, then neither Kaby Lake nor Coffee Lake would have existed. Intel only released Sky Lake Refresh and Sky Lake Refresh Refresh because Cannon Lake was so delayed. Some rumors now put Cannon Lake as a late 2018 product, and even then, only for low power dual cores that are completely uninteresting in desktops.
They can't keep on doing die shrinks forever. There have been delays before, but two entire intermediate generations on an old process node that weren't originally supposed to exist is a new thing. I'm not predicting that this will be the end of die shrinks. EUV will help, but that's 2019 if you want to be optimistic. But at this point, Moore's Law is on life support if not dead outright, as we're not getting die shrinks on the accustomed schedule anymore.
If Cannon Lake released as planned, then neither Kaby Lake nor Coffee Lake would have existed. Intel only released Sky Lake Refresh and Sky Lake Refresh Refresh because Cannon Lake was so delayed. Some rumors now put Cannon Lake as a late 2018 product, and even then, only for low power dual cores that are completely uninteresting in desktops.
They can't keep on doing die shrinks forever. There have been delays before, but two entire intermediate generations on an old process node that weren't originally supposed to exist is a new thing. I'm not predicting that this will be the end of die shrinks. EUV will help, but that's 2019 if you want to be optimistic. But at this point, Moore's Law is on life support if not dead outright, as we're not getting die shrinks on the accustomed schedule anymore.
Would it be a surprise if Intel doesn't deliver 10nm until - as you say - late 2018 or beyond? No. Which would mean no release of Ice Lake next summer.
However both Samsung and TSMC are shipping (what they call) 10nm products. And talking about (what they call) 7nm next year. All the same caveats of course - fully agree with your comments on die shrinkage.
Now whilst a Snapdragon 835 is not an e.g. i7 these products do mean that there are two companies using a 10nm process. For Intel "manufacturing world leader" this must both grate and be a concern. (Some debate on when is Xnm manufacturing process actually Xnm of course - hence the + and ++ etc).
As you say EUV could help - GlobalFoundries clearly hope so in their plans to skip 10nm.
The general assumption regarding EUV has been that its benefits will come at a cost. Have Intel decided to accept these cost? They seem to have loosened the purse strings so to speak and seem to be looking to - for want of a phrase - skip 8th gen cpus (not quite obviously). Just speculation but announcing 9th gen before 8th gen was released .... its not business as normal at the moment.
If Cannon Lake released as planned, then neither Kaby Lake nor Coffee Lake would have existed. Intel only released Sky Lake Refresh and Sky Lake Refresh Refresh because Cannon Lake was so delayed. Some rumors now put Cannon Lake as a late 2018 product, and even then, only for low power dual cores that are completely uninteresting in desktops.
They can't keep on doing die shrinks forever. There have been delays before, but two entire intermediate generations on an old process node that weren't originally supposed to exist is a new thing. I'm not predicting that this will be the end of die shrinks. EUV will help, but that's 2019 if you want to be optimistic. But at this point, Moore's Law is on life support if not dead outright, as we're not getting die shrinks on the accustomed schedule anymore.
Would it be a surprise if Intel doesn't deliver 10nm until - as you say - late 2018 or beyond? No. Which would mean no release of Ice Lake next summer.
However both Samsung and TSMC are shipping (what they call) 10nm products. And talking about (what they call) 7nm next year. All the same caveats of course - fully agree with your comments on die shrinkage.
Now whilst a Snapdragon 835 is not an e.g. i7 these products do mean that there are two companies using a 10nm process. For Intel "manufacturing world leader" this must both grate and be a concern. (Some debate on when is Xnm manufacturing process actually Xnm of course - hence the + and ++ etc).
As you say EUV could help - GlobalFoundries clearly hope so in their plans to skip 10nm.
The general assumption regarding EUV has been that its benefits will come at a cost. Have Intel decided to accept these cost? They seem to have loosened the purse strings so to speak and seem to be looking to - for want of a phrase - skip 8th gen cpus (not quite obviously). Just speculation but announcing 9th gen before 8th gen was released .... its not business as normal at the moment.
As you say though it may be 2019 before we know!
From various other measurements besides the headline process node number, it sounds like what Intel is calling 10 nm is closer to what other foundries are calling 7 nm it is to what other foundries are calling 10 nm.
As I understand it, the big hold-up on EUV thus far has been that it can't process wafers fast enough for a variety of technical reasons. If a foundry had decided that was acceptable and decided several years ago that they'd push ahead with EUV and just accept the reduced throughput, they could be shipping products using EUV today. The problem is that they'd build all that infrastructure for a new process node, and only be able to produce 10% or 20% or some such as many wafers as they could if they hadn't insisted on EUV. That makes it far too expensive to use for any high volume products.
There has been a lot of progress on EUV over the years, but progress has been much, much slower than anticipated. EUV was originally supposed to be used somewhere around 45 nm process nodes, and we've gone far beyond that.
AmazingAveryAge of Conan AdvocateMemberUncommonPosts: 7,188
On Cannon Lake - here is confirmation on shipping times. Most likely mobile product with Ice Lake the one to watch just in time for Zen2 showdown.
After running this story, Intel reached out to us with a statement on where it's at with Cannon Lake. "We'll be shipping our first 10nm products near the end of the year beginning with a lower volume SKU followed by a volume ramp in the first half of 2018," an Intel spokesperson said. It looks like Cannon Lake may be on track, after all. + here
On Cannon Lake - here is confirmation on shipping times. Most likely mobile product with Ice Lake the one to watch just in time for Zen2 showdown.
After running this story, Intel reached out to us with a statement on where it's at with Cannon Lake. "We'll be shipping our first 10nm products near the end of the year beginning with a lower volume SKU followed by a volume ramp in the first half of 2018," an Intel spokesperson said. It looks like Cannon Lake may be on track, after all. + here
Personally not sold on the 8700, it runs far too hot, prefer a Ryzen solution at the moment. Heat kills electronics. The only solution to the heat is to delid and hell if that is a solution for anyone. As to Cannon Lake, it is a not designed to be a chip for the masses, only a limited release. Going to have to wait until Icey Lake for that.
From what I see, Intel is in trouble. Intel is barely holding on in the desktop arena while AMD is owning the workstation market at the moment and I don't see anything from Intel to counter. Have to wait and see if their engineers can remedy that.
The motherboard manufacturers are not happy with Intel either. The constant change of new chipsets makes it hard for them to make a profit. I can get loaded AMD motherboards for $150 less than Intel boards with the same features.
For the thread count absolutely great if not spending above 2k USD
But the PCIE Lanes are only 12 for those running SLI or high 1080TI SLI systems not good socket 2066 still better.
Although I think since my CPU isn't hitting above 46 % usage when playing two games in ULTRA HD Graphics I can skip the upgrade as it costs around $600 just to upgrade the board unless I don't care and wait for something better the upgrade cost doesn't feel worthit unless I just don't care.
For someone who doesn't have a i7 absolutely likely a good CPU.
The PCI-e situation on Coffee Lake is a little confusing. You really need to read individual board specifications as they go into what you get with a certain chip. The Z370 chipset offers 24 pci-e 3.0 slots itself which should allow 2 GPU and a single m2 drive. Other peripherals would share the remaining 4 pci-e slots. Ryzen by comparison pushes almost everything onto the CPU where they are slowly phasing out the chipset entirely. The benefit of Intel's solution is that it is not bound by the memory frequency as much as Ryzen is.
The PCI-e situation on Coffee Lake is a little confusing. You really need to read individual board specifications as they go into what you get with a certain chip. The Z370 chipset offers 24 pci-e 3.0 slots itself which should allow 2 GPU and a single m2 drive. Other peripherals would share the remaining 4 pci-e slots. Ryzen by comparison pushes almost everything onto the CPU where they are slowly phasing out the chipset entirely. The benefit of Intel's solution is that it is not bound by the memory frequency as much as Ryzen is.
This is something I find confusing when thinking of a new build with VR in mind.
Brenics ~ Just to point out I do believe Chris Roberts is going down as the man who cheated backers and took down crowdfunding for gaming.
I am not all that wound up about the PCI lane debate.
Sure, more is “gooder” than less, and there really is no reason to artificially limit them except to screw with those few people who do. I think it’s one of those marketing points that AMD is latching on to that doesn’t mean much with regard to performance or actual use, but because it’s a great illustration of typical Intel douchbagery.
I guess some (very few) people still run multi-GPU, or raid a lot of M.2’s. But that isn’t a typical use case by far, and the standard recommendations (for better or worse) tend to run against actually using any of that — UNLESS you know what you are doing and why you are doing it (in which case you probably aren’t asking for advice anyway).
If it were an important use consideration for a build, I would probably be looking at server class motherboards and Epyc/Xeon Solutions rather than consumer gamer gear anyway.
That's the sort of thing that companies do when they believe that they're behind and need to catch up. And since Coffee Lake is really just Sky Lake Refresh Refresh, Intel could have launched it about two years ago if so inclined, though perhaps clocked a few hundred MHz lower.
AmazingAveryAge of Conan AdvocateMemberUncommonPosts: 7,188
In other words, while Z270 motherboards might be able to handle the jump to 6-core, things could get tricky when and if Intel introduces 8-core SKUs. If that is the case, then Intel erred on the side of caution by restricting Coffee Lake on Z270 motherboards altogether, rather than supporting Coffee Lake up to 6 cores, and requiring Z370 for 8-core CPUs. Of course, this is all speculation at the moment. However, with AMD pushing out 8-core CPUs to the mainstream market with Ryzen, it wouldn't surprise us if Intel followed suit in the not-too-distant future. Either way, Wu doesn't see the situation as a bad one for customers. "Usually, the kind of customers that already have a Z270 board will probably not upgrade for this generation. For the majority of Z370 users, maybe they have been using their PC for three or five years, so these are still good products for them to upgrade to," Wu said.
I'd agree, with the historical cadence of hardware releases, that most people would have felt the need to upgrade something every 2-4 years. And people who built Ivy-based rigs probably have upgraded a video card or storage between then and now.
But CPU/Motherboards - not all that much has changed. We have USB 3.1 (and maybe even 3.2) now, but even today's motherboards don't support that terribly well. M.2/NVMe is starting to take off, maybe. Intel is certainly pushing Optane now. But all of that can be added via PCI to older systems if you needed it.
So I don't know. I can understand the itch to just build something new. But I still don't see a strong value case to upgrade a gaming rig from anything Sandy Bridge or later just yet.
The article sounds like an excuse for Intel to push the 370 chipset - people who would be buying 8000 series CPUs wouldn't be coming from 170 or 270 chipsets yet, it's only the older systems. That isn't entirely true - a lot of people do like the option of a drop-in upgrade. The issue here is, it's not really a drop-in upgrade, it's more or less the same thing that has been available since Skylake first came out, plus two cores. Everyone is mad because they can't get those two extra cores without also getting the new motherboard and everything else that goes with it.
Shoot, a couple of years ago AMD released a 220W 8 Core CPU on legacy AM3+ socket. They didn't need to "upgrade the power delivery" for that - you could (at your own peril) drop it into the same motherboard that 4100 ran on, and run the 9590 on it without AMD requiring a new motherboard -- just because there is no reason except the charge additional money for a new chipset to do that.
AmazingAveryAge of Conan AdvocateMemberUncommonPosts: 7,188
The problem with the z370 is the motherboard prices. Where you can get a decent z270 for around $150, the z370's go for around $450. Add in the higher CPU costs and it really bumps up the system cost. Unless you have no money restraints, the z270 is a much better bargain, and the same goes for Ryzen.
AmazingAveryAge of Conan AdvocateMemberUncommonPosts: 7,188
The problem with the z370 is the motherboard prices. Where you can get a decent z270 for around $150, the z370's go for around $450. Add in the higher CPU costs and it really bumps up the system cost. Unless you have no money restraints, the z270 is a much better bargain, and the same goes for Ryzen.
Not sure where you are seeing $450 unless you are talking "Premium". You can get decent z370 for $150 too
The problem with the z370 is the motherboard prices. Where you can get a decent z270 for around $150, the z370's go for around $450. Add in the higher CPU costs and it really bumps up the system cost. Unless you have no money restraints, the z270 is a much better bargain, and the same goes for Ryzen.
Not sure where you are seeing $450 unless you are talking "Premium". You can get decent z370 for $150 too
The problem with the z370 is the motherboard prices. Where you can get a decent z270 for around $150, the z370's go for around $450. Add in the higher CPU costs and it really bumps up the system cost. Unless you have no money restraints, the z270 is a much better bargain, and the same goes for Ryzen.
Not sure where you are seeing $450 unless you are talking "Premium". You can get decent z370 for $150 too
So you show low end boards for the z370 and then try to compare them with z270 with much better features? ROFL!
Not really mate, you came in spouting $450 FUD. They do not go for $450...
For example MSI's mid range AM4 and z370 are priced more or less the same. In the links above if you read it, it says there is minimal build cost differences.
The Ryzen where supposed to be the kings of 3D rendering... well, I feel sorry for that 1800x:
If you compare your overclocked processor against other processors running at stock speeds, then pick a test where your processor does as well as possible, it's a good way to boost your own ego.
The Ryzen where supposed to be the kings of 3D rendering... well, I feel sorry for that 1800x:
If you compare your overclocked processor against other processors running at stock speeds, then pick a test where your processor does as well as possible, it's a good way to boost your own ego.
More like Intel's Ego, maybe? Lemme guess... Ryzen owner? Do I need to point out that the non boosted 8700k also beats all the other processors?
It's not an overclock btw, it's one parameter in the Z370 bios, "optimize muti core performance". It simply puts all cores at the maximum official turbo speed. And as said, it only gives it an edge on the non optimized 8700k, which still beats the crap out the competition at stock speed.
My ego is ok, but my customer satisfaction is definitely boosted. This is a top chipset with a top processor. My best upgrade since my 4790k.
Comments
No IPC gain over Kaby - just clock bump and core count improvements. Does appear to have more overclock potential though.
Intel damage limitation:
1. Reduce prices, get rid of some of the artificial restrictions on Pentiums, i3s etc.
2. More price cuts and unusually early announcements about Coffee Lake.
3. Get Coffee Lake with all the features we can include by Q4. Position as Ryzen equal.
Intel warp drive part - so long tic-tac-toe!
1. Coffee Lake - obviously but this was pretty much happening.
2. Cannon Lake - a 10nm Coffee Lake test bed? 8th gen tag suggests this.
3. Ice Lake. This is the jump to warp and what Intel hope (?) to position as the "Ryzen killer".
Also expect AMD to maybe accelerate the introduction of Raven Ridge (integrated graphics) as part of their response.
At the end of the day Intel are a very big company and can - if they choose - "throw money" at things. Clearly - having moved from tick-tock to tic-tac-toe they would prefer not to but Kaby, i9 Slylake, Coffee, Cannon and Ice Lake .... if the latter two release as planned .... is a significant increase in what we had 12 months ago and should herald an exciting 12 months. Maybe not for motherboard m/fs - reports surfacing that they cannot cope which is why we may not be getting backward compatibility.
They can't keep on doing die shrinks forever. There have been delays before, but two entire intermediate generations on an old process node that weren't originally supposed to exist is a new thing. I'm not predicting that this will be the end of die shrinks. EUV will help, but that's 2019 if you want to be optimistic. But at this point, Moore's Law is on life support if not dead outright, as we're not getting die shrinks on the accustomed schedule anymore.
Would it be a surprise if Intel doesn't deliver 10nm until - as you say - late 2018 or beyond? No. Which would mean no release of Ice Lake next summer.
However both Samsung and TSMC are shipping (what they call) 10nm products. And talking about (what they call) 7nm next year. All the same caveats of course - fully agree with your comments on die shrinkage.
Now whilst a Snapdragon 835 is not an e.g. i7 these products do mean that there are two companies using a 10nm process. For Intel "manufacturing world leader" this must both grate and be a concern. (Some debate on when is Xnm manufacturing process actually Xnm of course - hence the + and ++ etc).
As you say EUV could help - GlobalFoundries clearly hope so in their plans to skip 10nm.
The general assumption regarding EUV has been that its benefits will come at a cost. Have Intel decided to accept these cost? They seem to have loosened the purse strings so to speak and seem to be looking to - for want of a phrase - skip 8th gen cpus (not quite obviously). Just speculation but announcing 9th gen before 8th gen was released .... its not business as normal at the moment.
As you say though it may be 2019 before we know!
As I understand it, the big hold-up on EUV thus far has been that it can't process wafers fast enough for a variety of technical reasons. If a foundry had decided that was acceptable and decided several years ago that they'd push ahead with EUV and just accept the reduced throughput, they could be shipping products using EUV today. The problem is that they'd build all that infrastructure for a new process node, and only be able to produce 10% or 20% or some such as many wafers as they could if they hadn't insisted on EUV. That makes it far too expensive to use for any high volume products.
There has been a lot of progress on EUV over the years, but progress has been much, much slower than anticipated. EUV was originally supposed to be used somewhere around 45 nm process nodes, and we've gone far beyond that.
Ridelynn said: On Coffee Lake there are ton of reviews listed here: https://videocardz.com/73089/intel-8th-gen-core-coffee-lake-s-review-roundup
IMO the i7 8700K is an absolute beast. Out of all of the reviews I think the final words summary here and in general is one of the most honest reviews out there > http://www.guru3d.com/articles_pages/intel_core_i7_8700k_processor_review,28.html
On Cannon Lake - here is confirmation on shipping times.
Most likely mobile product with Ice Lake the one to watch just in time for Zen2 showdown.
"We'll be shipping our first 10nm products near the end of the year beginning with a lower volume SKU followed by a volume ramp in the first half of 2018," an Intel spokesperson said.
It looks like Cannon Lake may be on track, after all. + here
From what I see, Intel is in trouble. Intel is barely holding on in the desktop arena while AMD is owning the workstation market at the moment and I don't see anything from Intel to counter. Have to wait and see if their engineers can remedy that.
The motherboard manufacturers are not happy with Intel either. The constant change of new chipsets makes it hard for them to make a profit. I can get loaded AMD motherboards for $150 less than Intel boards with the same features.
https://www.newegg.com/Product/Product.aspx?Item=N82E16819117827
For the thread count absolutely great if not spending above 2k USD
But the PCIE Lanes are only 12 for those running SLI or high 1080TI SLI systems not good socket 2066 still better.
Although I think since my CPU isn't hitting above 46 % usage when playing two games in ULTRA HD Graphics I can skip the upgrade as it costs around $600 just to upgrade the board unless I don't care and wait for something better the upgrade cost doesn't feel worthit unless I just don't care.
For someone who doesn't have a i7 absolutely likely a good CPU.
The Z370 chipset offers 24 pci-e 3.0 slots itself which should allow 2 GPU and a single m2 drive. Other peripherals would share the remaining 4 pci-e slots.
Ryzen by comparison pushes almost everything onto the CPU where they are slowly phasing out the chipset entirely.
The benefit of Intel's solution is that it is not bound by the memory frequency as much as Ryzen is.
This is something I find confusing when thinking of a new build with VR in mind.
Brenics ~ Just to point out I do believe Chris Roberts is going down as the man who cheated backers and took down crowdfunding for gaming.
Sure, more is “gooder” than less, and there really is no reason to artificially limit them except to screw with those few people who do. I think it’s one of those marketing points that AMD is latching on to that doesn’t mean much with regard to performance or actual use, but because it’s a great illustration of typical Intel douchbagery.
I guess some (very few) people still run multi-GPU, or raid a lot of M.2’s. But that isn’t a typical use case by far, and the standard recommendations (for better or worse) tend to run against actually using any of that — UNLESS you know what you are doing and why you are doing it (in which case you probably aren’t asking for advice anyway).
If it were an important use consideration for a build, I would probably be looking at server class motherboards and Epyc/Xeon Solutions rather than consumer gamer gear anyway.
https://www.hardocp.com/news/2017/10/09/newegg_confirms_no_stock_on_coffee_lake_any_time_soon
That's the sort of thing that companies do when they believe that they're behind and need to catch up. And since Coffee Lake is really just Sky Lake Refresh Refresh, Intel could have launched it about two years ago if so inclined, though perhaps clocked a few hundred MHz lower.
Of course, this is all speculation at the moment. However, with AMD pushing out 8-core CPUs to the mainstream market with Ryzen, it wouldn't surprise us if Intel followed suit in the not-too-distant future.
Either way, Wu doesn't see the situation as a bad one for customers.
"Usually, the kind of customers that already have a Z270 board will probably not upgrade for this generation. For the majority of Z370 users, maybe they have been using their PC for three or five years, so these are still good products for them to upgrade to," Wu said.
Ivy Bridge is just a hair over 5 years old now.
I'd agree, with the historical cadence of hardware releases, that most people would have felt the need to upgrade something every 2-4 years. And people who built Ivy-based rigs probably have upgraded a video card or storage between then and now.
But CPU/Motherboards - not all that much has changed. We have USB 3.1 (and maybe even 3.2) now, but even today's motherboards don't support that terribly well. M.2/NVMe is starting to take off, maybe. Intel is certainly pushing Optane now. But all of that can be added via PCI to older systems if you needed it.
So I don't know. I can understand the itch to just build something new. But I still don't see a strong value case to upgrade a gaming rig from anything Sandy Bridge or later just yet.
The article sounds like an excuse for Intel to push the 370 chipset - people who would be buying 8000 series CPUs wouldn't be coming from 170 or 270 chipsets yet, it's only the older systems. That isn't entirely true - a lot of people do like the option of a drop-in upgrade. The issue here is, it's not really a drop-in upgrade, it's more or less the same thing that has been available since Skylake first came out, plus two cores. Everyone is mad because they can't get those two extra cores without also getting the new motherboard and everything else that goes with it.
Shoot, a couple of years ago AMD released a 220W 8 Core CPU on legacy AM3+ socket. They didn't need to "upgrade the power delivery" for that - you could (at your own peril) drop it into the same motherboard that 4100 ran on, and run the 9590 on it without AMD requiring a new motherboard -- just because there is no reason except the charge additional money for a new chipset to do that.
Intel i7-8700K vs i7-2600K - FINALLY Time to Upgrade Sandy Bridge?
https://www.newegg.com/Product/ProductList.aspx?Submit=ENE&N=-1&IsNodeId=1&Description=z370&bop=And&PageSize=96&order=PRICE
Sorry but differences in system cost build is marginal
https://www.gamersnexus.net/hwreviews/3076-intel-i7-8700k-review-vs-ryzen-streaming-gaming-overclocking
http://cpu.userbenchmark.com/Compare/Intel-Core-i7-8700K-vs-AMD-Ryzen-7-1700X/3937vs3915
The bigger issue is just that you need 370 at all.
For example MSI's mid range AM4 and z370 are priced more or less the same. In the links above if you read it, it says there is minimal build cost differences.
Here something for you to look at, there is just 1 in that list that is +$450 lol - Analyzing z370 for Intel's 8th generation Coffee Lake: A quick look at 50+ motherboards
Core i5-8400 gaming performance
Core i5-8400 application performance
----------------------------------------------------------
Core i7-8700K gaming performance
Core i7-8700K application performance