Intel's Future Chips: News, Rumours & Reviews

[jdwii]

Coffee-lake

https://newsroom.intel.com/newsroom/wp-content/uploads/sites/11/2017/09/8th-gen-intel-core-overview.pdf

 

[MDXX]

Man 360$, i was sooo hoping for 320-340$ :/. Oh well, ill still get it, should be a beast of a chip

 

[jdwii]

After these prices Ryzen price cuts must happen if not Amd will be selling CPU's with worse value to their Intel counterparts.

We are all going to get some massive good deals on CPU's to bad everything else is so expensive haha

 

[-Fran-]

I don't think Intel is trying to replace it's own line up nor AMDs, but bumping it in price artificially. I really don't think the cost per waffer increases with an additional (or two) CPU core in the die, but this is only to protect the still relatively new Kaby-Lake family (less than a year) while they phase them out (this reasoning was in the Tom's article and I agree with it). Not the only reason, but sounds the most "Intel-like" one to me. Why charge less when you can charge more?

That being said, I expect the i7-8700K to be quite the performer and I hope Intel does not disappoint with the whole platform, since it's forcing everyone to upgrade, even though it uses the same socket (obviously, not the same pin-layout).

And in the same Intel comments, I like the i9 7960X, even at it's USD$1700. It's not a good "value" proposition, but it's a strong offering if you want the best. Too bad Toms couldn't make the big brother work, but it doesn't surprise me after all the problems the X299 platform has been showing.

Cheers!

 

[YoAndy]

Spoiler

Spoiler

We have talked about this multiple times, of course AMD want us to use their CPU's at higher resolutions so they can hide their weak single core performance, because at low resolutions, the Intel advantage is significant, and it trails off as you run games at higher resolutions and detail settings only (because the graphics card becomes the performance bottleneck) duh. The intel core i7 7700K outpaced the 1700 by around 40fps in the CPU-intensive game Ashes of the Singularity test.
The 7700K’s good performance continued to single-threaded tests: its 472-point result in POVRay was easily ahead of the Ryzen chip’s 315-point result, and it is nearly 60 points better in Cinebench. The Core i7 is a better overclocker than AMD, too, and its power consumption isn't much higher than the Ryzen 7 1700.The only area where the Core i7-7700K falls behind is in multi-threaded benchmarks, don't let fake reviews fool you.

Do you really want me to believe that when we ran a test and my i7 990X past 4GHz released Q1'11 gets equal and sometimes 20+ more FPS than my Friend's new Ryzen 7 each one using one of my 2 GTX 980 Ti's graphics cards while playing online together!!!??? Same games, same levels walking side by side while talking through a headset. !!!??? I just can't.

As the graph shows depending on the game results can vary depending on the title, and there were compatibility issues with RAM and optimization issues with games. Try comparing frame rates now. agesa 1.0.0.6 update have made things much better now for RAM compatibility, and tightened the internal latency problem with Ryzen a bit. You do know that the difference in using 2133MHz RAM vs 3200MHz RAM offers a 15% better FPS performance increase vs. Intel using the same RAM. I'm sure your testing conditions were likely less than ideal.

Lol I'm still using triple channel DDR3 and he is using DDR4

You processor doesn't suffer Internal latency problems. Faster RAM reduces internal latency associated with infinity fabric. Also, as mentioned before,"agesa 1.0.0.6 update have made things much better now for RAM compatibility, and tightened the internal latency problem with Ryzen a bit." Like I also said, "I'm sure your testing conditions were likely less than ideal."

The problem here is Ryzen, They have way to many issues, like ram and not been able to run any faster than 3200mhz or most processors not been able to overclock past 4GHz without liquid nitrogen. In the gaming world speed is king and below 4GHz is just not good.

 

[juanrga]

I don't think Intel is trying to replace it's own line up nor AMDs, but bumping it in price artificially. I really don't think the cost per waffer increases with an additional (or two) CPU core in the die, but this is only to protect the still relatively new Kaby-Lake family (less than a year) while they phase them out (this reasoning was in the Tom's article and I agree with it). Not the only reason, but sounds the most "Intel-like" one to me. Why charge less when you can charge more?

That being said, I expect the i7-8700K to be quite the performer and I hope Intel does not disappoint with the whole platform, since it's forcing everyone to upgrade, even though it uses the same socket (obviously, not the same pin-layout).

And in the same Intel comments, I like the i9 7960X, even at it's USD$1700. It's not a good "value" proposition, but it's a strong offering if you want the best. Too bad Toms couldn't make the big brother work, but it doesn't surprise me after all the problems the X299 platform has been showing.

The die is about 25% larger, so it will cost more to fabricate.

The new Cofee Lake line replaces the KabyLake line, so I don't understand what "protect the still relatively new Kaby-Lake family" means. Intel almost always price the new chips in the same price bracket that the chips replaced. This was demonstrated before using the HEDT line: 6 core chips for same price than 4-core chips then 8 core chips for same price than 6-core chips then 10 core chips for same price than 8-core chips.

So what is this all this conspiracy again?

 

[adamsleath]

intel are pretty clever with their products and marketing. cfl/8th gen is a complete range with obvious and significant performance advantages over 7th gen and at a similiar pricepoint to the existing lineup.

the nuclear reactor motherboard melting 18 core x chip on the other hand. 4.4ghz is basically its limit. ..ive seen the benches. 10-20% tops over TR and it doesnt trump 1950x in all cases (but most of them) for 70-100% price premium. but for those who can pay. why not.
.

 

[-Fran-]

I don't think Intel is trying to replace it's own line up nor AMDs, but bumping it in price artificially. I really don't think the cost per waffer increases with an additional (or two) CPU core in the die, but this is only to protect the still relatively new Kaby-Lake family (less than a year) while they phase them out (this reasoning was in the Tom's article and I agree with it). Not the only reason, but sounds the most "Intel-like" one to me. Why charge less when you can charge more?

That being said, I expect the i7-8700K to be quite the performer and I hope Intel does not disappoint with the whole platform, since it's forcing everyone to upgrade, even though it uses the same socket (obviously, not the same pin-layout).

And in the same Intel comments, I like the i9 7960X, even at it's USD$1700. It's not a good "value" proposition, but it's a strong offering if you want the best. Too bad Toms couldn't make the big brother work, but it doesn't surprise me after all the problems the X299 platform has been showing.

The die is about 25% larger, so it will cost more to fabricate.

The new Cofee Lake line replaces the KabyLake line, so I don't understand what "protect the still relatively new Kaby-Lake family" means. Intel almost always price the new chips in the same price bracket that the chips replaced. This was demonstrated before using the HEDT line: 6 core chips for same price than 4-core chips then 8 core chips for same price than 6-core chips then 10 core chips for same price than 8-core chips.

So what is this all this conspiracy again?

Hey, you learned a new word: "conspiracy"! Now you just need to learn how to use it properly. Reason being, I don't know why you think what I said is a "conspiracy". It's just an observation from the reviews circulating (I think I made it explicit I read it in Toms article).

But hey! New fancy word none the less!

Now, to the interesting bits: "The die is about 25% larger, so it will cost more to fabricate." <- I'm not an Intel fab worker, but I'm pretty sure it can't be a linear cost, otherwise they would have priced them *way* higher. It might be a few extra cents per working die (QA included)?

And this: "he new Cofee Lake line replaces the KabyLake line" <- I'll argue semantics, since it's not replacing it yet. They have a higher price range and different core count (among other less important and obvious differences). Plus, it seems marketing is making a very good point of making sure everyone gets a new MoBo for them.

I think your point boils down to "why question Intel's higher price tag for a new CPU?". My answer, in very good etiquette is to answer with another question: "why not?". Value is relative, but I'll keep it short and not delve into it too much.

EDIT: Forgot a word.

 

[juanrga]

I don't think Intel is trying to replace it's own line up nor AMDs, but bumping it in price artificially. I really don't think the cost per waffer increases with an additional (or two) CPU core in the die, but this is only to protect the still relatively new Kaby-Lake family (less than a year) while they phase them out (this reasoning was in the Tom's article and I agree with it). Not the only reason, but sounds the most "Intel-like" one to me. Why charge less when you can charge more?

That being said, I expect the i7-8700K to be quite the performer and I hope Intel does not disappoint with the whole platform, since it's forcing everyone to upgrade, even though it uses the same socket (obviously, not the same pin-layout).

And in the same Intel comments, I like the i9 7960X, even at it's USD$1700. It's not a good "value" proposition, but it's a strong offering if you want the best. Too bad Toms couldn't make the big brother work, but it doesn't surprise me after all the problems the X299 platform has been showing.

The die is about 25% larger, so it will cost more to fabricate.

The new Cofee Lake line replaces the KabyLake line, so I don't understand what "protect the still relatively new Kaby-Lake family" means. Intel almost always price the new chips in the same price bracket that the chips replaced. This was demonstrated before using the HEDT line: 6 core chips for same price than 4-core chips then 8 core chips for same price than 6-core chips then 10 core chips for same price than 8-core chips.

So what is this all this conspiracy again?

Hey, you learned a new word: "conspiracy"! Now you just need to learn how to use it properly. Reason being, I don't know why you think what I said is a "conspiracy". It's just an observation from the reviews circulating (I think I made it explicit I read it in Toms article).

But hey! New fancy word none the less!

Now, to the interesting bits: "The die is about 25% larger, so it will cost more to fabricate." <- I'm not an Intel fab worker, but I'm pretty sure it can't be a linear cost, otherwise they would have priced them *way* higher. It might be a few extra cents per working die (QA included)?

And this: "he new Cofee Lake line replaces the KabyLake line" <- I'll argue semantics, since it's not replacing it yet. They have a higher price range and different core count (among other less important and obvious differences). Plus, it seems marketing is making a very good point of making sure everyone gets a new MoBo for them.

I think your point boils down to "why question Intel's higher price tag for a new CPU?". My answer, in very good etiquette is to answer with another question: "why not?". Value is relative, but I'll keep it short and not delve into it too much.

EDIT: Forgot a word.

The new die cost more than "few cents", but it doesn't cost that nonsense that Charlie wrote in an article some weeks ago...

Yes, CoffeLake (aka 8th gen) replaces KabyLake, just as Kabylake replaced Skylake. I don't even know why you claim that saying that CoffeLake replaces Kabylake is semantics.

A change on mobos was expected. Old mobos are designed for up to four cores and 2400MHz, and new mobos are designed for up to six-cores and 2666MHz. Again, what are those conspiracy accusations?

 

[YoAndy]

I don't think Intel is trying to replace it's own line up nor AMDs, but bumping it in price artificially. I really don't think the cost per waffer increases with an additional (or two) CPU core in the die, but this is only to protect the still relatively new Kaby-Lake family (less than a year) while they phase them out (this reasoning was in the Tom's article and I agree with it). Not the only reason, but sounds the most "Intel-like" one to me. Why charge less when you can charge more?

That being said, I expect the i7-8700K to be quite the performer and I hope Intel does not disappoint with the whole platform, since it's forcing everyone to upgrade, even though it uses the same socket (obviously, not the same pin-layout).

And in the same Intel comments, I like the i9 7960X, even at it's USD$1700. It's not a good "value" proposition, but it's a strong offering if you want the best. Too bad Toms couldn't make the big brother work, but it doesn't surprise me after all the problems the X299 platform has been showing.

The die is about 25% larger, so it will cost more to fabricate.

The new Cofee Lake line replaces the KabyLake line, so I don't understand what "protect the still relatively new Kaby-Lake family" means. Intel almost always price the new chips in the same price bracket that the chips replaced. This was demonstrated before using the HEDT line: 6 core chips for same price than 4-core chips then 8 core chips for same price than 6-core chips then 10 core chips for same price than 8-core chips.

So what is this all this conspiracy again?

Hey, you learned a new word: "conspiracy"! Now you just need to learn how to use it properly. Reason being, I don't know why you think what I said is a "conspiracy". It's just an observation from the reviews circulating (I think I made it explicit I read it in Toms article).

But hey! New fancy word none the less!

Now, to the interesting bits: "The die is about 25% larger, so it will cost more to fabricate." <- I'm not an Intel fab worker, but I'm pretty sure it can't be a linear cost, otherwise they would have priced them *way* higher. It might be a few extra cents per working die (QA included)?

And this: "he new Cofee Lake line replaces the KabyLake line" <- I'll argue semantics, since it's not replacing it yet. They have a higher price range and different core count (among other less important and obvious differences). Plus, it seems marketing is making a very good point of making sure everyone gets a new MoBo for them.

I think your point boils down to "why question Intel's higher price tag for a new CPU?". My answer, in very good etiquette is to answer with another question: "why not?". Value is relative, but I'll keep it short and not delve into it too much.

EDIT: Forgot a word.

The new die cost more than "few cents", but it doesn't cost that nonsense that Charlie wrote in an article some weeks ago...

Yes, CoffeLake (aka 8th gen) replaces KabyLake, just as Kabylake replaced Skylake. I don't even know why you claim that saying that CoffeLake replaces Kabylake is semantics.

A change on mobos was expected. Old mobos are designed for up to four cores and 2400MHz, and new mobos are designed for up to six-cores and 2666MHz. Again, what are those conspiracy accusations?

You haven't heard the latest news?!?, Intel's new facility will be underground in area 51😀

 

[juanrga]

Some leaked benches on CoffeLake

 

[-Fran-]

The new die cost more than "few cents", but it doesn't cost that nonsense that Charlie wrote in an article some weeks ago...

Yes, CoffeLake (aka 8th gen) replaces KabyLake, just as Kabylake replaced Skylake. I don't even know why you claim that saying that CoffeLake replaces Kabylake is semantics.

A change on mobos was expected. Old mobos are designed for up to four cores and 2400MHz, and new mobos are designed for up to six-cores and 2666MHz. Again, what are those conspiracy accusations?

So, for the costs. Would you like to share a breakdown of the Intel manufacturing costs? I'll just allow us to assume packaging and distribution will not be affected.

As for the replacement... Has Intel announced the phase out for KabyLake yet? Any official word on that, or you're just assuming since there's a new product that targets similar verticals it's an instant replace?

And as for the MoBo, I don't care why they need to change it technically. This might as well be a business decision and not a technical decision. I do understand the pin-layout change, but all else equals, why adding two additional cores to the CPU package and bumping the speeds would necessitate a new *platform* in your opinion?

 

[Isaiah4110]

Can we please stop the ridiculous back and forth? Many people are still excited about the new AMD processor line because of the potential it brings for real competition again among those who actually live on a budget and seek to buy the best bang-for-buck parts. Looking at what's available RIGHT NOW, today, if I were to place a parts order (I'm not), I would be comparing the following:

AMD Option:
PCPartPicker part list / Price breakdown by merchant

CPU: AMD - Ryzen 5 1600X 3.6GHz 6-Core Processor ($234.68 @ Amazon)
CPU Cooler: Noctua - NH-L9x65 SE-AM4 CPU Cooler ($44.90 @ Amazon)
Motherboard: MSI - X370 GAMING PLUS ATX AM4 Motherboard ($112.97 @ Amazon)
Memory: G.Skill - Ripjaws V Series 16GB (2 x 8GB) DDR4-2400 Memory ($133.99 @ Newegg)
Total: $526.54
Prices include shipping, taxes, and discounts when available
Generated by PCPartPicker 2017-09-25 14:09 EDT-0400


Intel Option:
PCPartPicker part list / Price breakdown by merchant

CPU: Intel - Core i5-7600K 3.8GHz Quad-Core Processor ($209.89 @ Amazon)
CPU Cooler: Noctua - NH-L9x65 33.8 CFM CPU Cooler ($48.95 @ Amazon)
Motherboard: ASRock - Z270 Extreme4 ATX LGA1151 Motherboard ($121.98 @ Newegg)
Memory: G.Skill - Ripjaws V Series 16GB (2 x 8GB) DDR4-2400 Memory ($133.99 @ Newegg)
Total: $514.81
Prices include shipping, taxes, and discounts when available
Generated by PCPartPicker 2017-09-25 14:07 EDT-0400


That would be my starting point, I between those two options, Intel is slightly cheaper and gives similar or slightly better performance out of the box. (Personally I would probably start with a ~$20 cooler because I don't like to overclock until I've had the system a while and want/need more performance, but for price comparison it was easier to grab a cooler compatible with both platforms without extra costs.) Right now? I would buy the Intel option. But here's the thing: Intel is about to release the 8000 series CPU line, and what is the price point going to be? If they can only be purchased for $240+ then that swings the price point a bit and, for those who do not want to buy a non-current generation and are on a tight budget, might shift the decision in AMD's favor.


WHY is this relevant? Because when was the last time AMD was this close to Intel in the price/performance comparison at stock settings? 3 years ago, when I built my last rig, AMD wasn't even a serious consideration because it couldn't come close from a price/performance standpoint. Now AMD is close enough to at least be in the discussion. For the VAST majority of builders and PCs, that's all that matters. Those who are considering dropping $2k on JUST the CPU are not usually as concerned with price/performance.




Now, can we PLEASE get off the AMD-Intel fanboy fight and get back to JUST discussing Intel news and updates on this thread?

 

[Phaaze88]

^Intel is NOT cheaper... you picked the wrong processor to compare with that 1600.

This is some hardcore back-and-forth, even if the real world application improvements aren't as cut and dry as some would like.

 

[adamsleath]

around october 5th im expecting so many reviews of the cfl / z370 systems. it will be a flood of info to sort through.....im very interested to see the lower cost 8th gen reviews aswell.

i guess at some point the production of 7th gen will stop and there will be plenty of second hand 7th gen chips down the line which are still very good performance. but the 6-core standard is a game changer in the market imo.

 

[8350rocks]

nobody plays games on HEDT @ 720p though...

The 720p testing has a different reason. The reason has been explained a dozen of times. Ignoring the reason will not stop reviews from performing 720p and 1080p tests.

It does have a different reason, to create an artificial benchmark that proves nothing. performance there does not accurately indicate performance in anything.

Believe it or not some people play 1080p at high refresh rates. Mind blown.

Testing in 720p allows you to see the framerate at which the CPU can push to at maximum. Let's say in 1 year 200hz@1080p monitors become standard for 1080p. People upgrade GPUs more often than CPUs. While Ryzen mght only be able to push 110fps, An Intel chip might push 160fps or more.

This is a fallacy.

720p does not create a situation where you can learn any useful real world performance. Show me one example where 720p performance is indicative of a real world performance result in anything.

1080p is also ridiculously lauded as the best resolution for gaming, the only people who say that either want a big number in the corner of their monitor (because they cannot truthfully tell the difference between 90 FPS and 190 FPS, despite what they claim), or people who have never played on 4K or ultrawide.

1080p looks great if you came from a laptop. If you have 4K and see the clarity and fidelity routinely, 1080p looks like an etch-a-sketch version of what you could be seeing.

There is nothing useful coming out of 720p benchmarks, and as 1440p and greater become more ubiquitous, then 1080p will become the new 720p benchmark creating an artificial scenario that does not justifiably show performance that has any meaning.

 

[ibjeepr]

nobody plays games on HEDT @ 720p though...

The 720p testing has a different reason. The reason has been explained a dozen of times. Ignoring the reason will not stop reviews from performing 720p and 1080p tests.

It does have a different reason, to create an artificial benchmark that proves nothing. performance there does not accurately indicate performance in anything.

Believe it or not some people play 1080p at high refresh rates. Mind blown.

Testing in 720p allows you to see the framerate at which the CPU can push to at maximum. Let's say in 1 year 200hz@1080p monitors become standard for 1080p. People upgrade GPUs more often than CPUs. While Ryzen mght only be able to push 110fps, An Intel chip might push 160fps or more.

This is a fallacy.

720p does not create a situation where you can learn any useful real world performance. Show me one example where 720p performance is indicative of a real world performance result in anything.

1080p is also ridiculously lauded as the best resolution for gaming, the only people who say that either want a big number in the corner of their monitor (because they cannot truthfully tell the difference between 90 FPS and 190 FPS, despite what they claim), or people who have never played on 4K or ultrawide.

1080p looks great if you came from a laptop. If you have 4K and see the clarity and fidelity routinely, 1080p looks like an etch-a-sketch version of what you could be seeing.

There is nothing useful coming out of 720p benchmarks, and as 1440p and greater become more ubiquitous, then 1080p will become the new 720p benchmark creating an artificial scenario that does not justifiably show performance that has any meaning.

Testing at a lower resolution allows you to remove the GPU from the equation (you'd hope) so that you are getting the max FPS the CPU can push which should be well below the max FPS the GPU could push at that resolution. So for CPU testing a low resolution is used. Most use 1080 as their test resolution though.

E-sports gamers play at 1080 because yes, they want max FPS and they don't care about all the distracting eye candy.

 

[juanrga]

The new die cost more than "few cents", but it doesn't cost that nonsense that Charlie wrote in an article some weeks ago...

Yes, CoffeLake (aka 8th gen) replaces KabyLake, just as Kabylake replaced Skylake. I don't even know why you claim that saying that CoffeLake replaces Kabylake is semantics.

A change on mobos was expected. Old mobos are designed for up to four cores and 2400MHz, and new mobos are designed for up to six-cores and 2666MHz. Again, what are those conspiracy accusations?

So, for the costs. Would you like to share a breakdown of the Intel manufacturing costs? I'll just allow us to assume packaging and distribution will not be affected.

As for the replacement... Has Intel announced the phase out for KabyLake yet? Any official word on that, or you're just assuming since there's a new product that targets similar verticals it's an instant replace?

LOL.

And as for the MoBo, I don't care why they need to change it technically. This might as well be a business decision and not a technical decision. I do understand the pin-layout change, but all else equals, why adding two additional cores to the CPU package and bumping the speeds would necessitate a new *platform* in your opinion?

Why would I share a mere opinion when I can give tech facts? Slide #7 in the set given by jdwii explains in a plain language why a new chipset is needed for the new chips.

 

[gamerk316]

nobody plays games on HEDT @ 720p though...

The 720p testing has a different reason. The reason has been explained a dozen of times. Ignoring the reason will not stop reviews from performing 720p and 1080p tests.

It does have a different reason, to create an artificial benchmark that proves nothing. performance there does not accurately indicate performance in anything.

Believe it or not some people play 1080p at high refresh rates. Mind blown.

Testing in 720p allows you to see the framerate at which the CPU can push to at maximum. Let's say in 1 year 200hz@1080p monitors become standard for 1080p. People upgrade GPUs more often than CPUs. While Ryzen mght only be able to push 110fps, An Intel chip might push 160fps or more.

This is a fallacy.

720p does not create a situation where you can learn any useful real world performance. Show me one example where 720p performance is indicative of a real world performance result in anything.

1080p is also ridiculously lauded as the best resolution for gaming, the only people who say that either want a big number in the corner of their monitor (because they cannot truthfully tell the difference between 90 FPS and 190 FPS, despite what they claim), or people who have never played on 4K or ultrawide.

1080p looks great if you came from a laptop. If you have 4K and see the clarity and fidelity routinely, 1080p looks like an etch-a-sketch version of what you could be seeing.

There is nothing useful coming out of 720p benchmarks, and as 1440p and greater become more ubiquitous, then 1080p will become the new 720p benchmark creating an artificial scenario that does not justifiably show performance that has any meaning.

Testing at a lower resolution allows you to remove the GPU from the equation (you'd hope) so that you are getting the max FPS the CPU can push which should be well below the max FPS the GPU could push at that resolution. So for CPU testing a low resolution is used. Most use 1080 as their test resolution though.

E-sports gamers play at 1080 because yes, they want max FPS and they don't care about all the distracting eye candy.

It's more then that, it's about knowing if your CPU becomes a problem when paired with future GPU upgrades. Its about performance headroom.

Yes, at 1440p and GPU X, two CPUs might be equal. But what about a more powerful GPU? Can each CPU drive it? How much performance headroom is on the table? That's the point of low resolution tests. It's not about "real world, today", it's about "real world, tomorrow".

 

[juanrga]

Another leaked benchmark shows the $360 six-core CoffeeLake dangerously close to the $500 8-core RyZen even in benchmarks that scale above six-cores

 

[juanrga]

nobody plays games on HEDT @ 720p though...

The 720p testing has a different reason. The reason has been explained a dozen of times. Ignoring the reason will not stop reviews from performing 720p and 1080p tests.

It does have a different reason, to create an artificial benchmark that proves nothing. performance there does not accurately indicate performance in anything.

Believe it or not some people play 1080p at high refresh rates. Mind blown.

Testing in 720p allows you to see the framerate at which the CPU can push to at maximum. Let's say in 1 year 200hz@1080p monitors become standard for 1080p. People upgrade GPUs more often than CPUs. While Ryzen mght only be able to push 110fps, An Intel chip might push 160fps or more.

This is a fallacy.

720p does not create a situation where you can learn any useful real world performance. Show me one example where 720p performance is indicative of a real world performance result in anything.

1080p is also ridiculously lauded as the best resolution for gaming, the only people who say that either want a big number in the corner of their monitor (because they cannot truthfully tell the difference between 90 FPS and 190 FPS, despite what they claim), or people who have never played on 4K or ultrawide.

1080p looks great if you came from a laptop. If you have 4K and see the clarity and fidelity routinely, 1080p looks like an etch-a-sketch version of what you could be seeing.

There is nothing useful coming out of 720p benchmarks, and as 1440p and greater become more ubiquitous, then 1080p will become the new 720p benchmark creating an artificial scenario that does not justifiably show performance that has any meaning.

Testing at a lower resolution allows you to remove the GPU from the equation (you'd hope) so that you are getting the max FPS the CPU can push which should be well below the max FPS the GPU could push at that resolution. So for CPU testing a low resolution is used. Most use 1080 as their test resolution though.

E-sports gamers play at 1080 because yes, they want max FPS and they don't care about all the distracting eye candy.

It's more then that, it's about knowing if your CPU becomes a problem when paired with future GPU upgrades. Its about performance headroom.

Yes, at 1440p and GPU X, two CPUs might be equal. But what about a more powerful GPU? Can each CPU drive it? How much performance headroom is on the table? That's the point of low resolution tests. It's not about "real world, today", it's about "real world, tomorrow".

Exactly!

In this example all CPUs seem to give the same performance

In reality this is fictitious, because the GPU is bottlenecking the fastest CPUs. Then we lower the resolution to to reduce the HPU bottleneck and test the CPUs and we get a different picture

The above 4K slide is what we obtain today with a current GPU. In one or two years when we upgrade to a faster GPU, the GPU-bottleneck will be removed and future 4K benches will look as current 1080p benches.

People that chose RyZen now and claim that "it is enough for gaming" will be bottlenecking its system in future GPU upgrades, whereas people on fastest Intel CPUs will continue to play games well in future.

 

[KirbysHammer]

I don't think Intel is trying to replace it's own line up nor AMDs, but bumping it in price artificially. I really don't think the cost per waffer increases with an additional (or two) CPU core in the die, but this is only to protect the still relatively new Kaby-Lake family (less than a year) while they phase them out (this reasoning was in the Tom's article and I agree with it). Not the only reason, but sounds the most "Intel-like" one to me. Why charge less when you can charge more?

That being said, I expect the i7-8700K to be quite the performer and I hope Intel does not disappoint with the whole platform, since it's forcing everyone to upgrade, even though it uses the same socket (obviously, not the same pin-layout).

And in the same Intel comments, I like the i9 7960X, even at it's USD$1700. It's not a good "value" proposition, but it's a strong offering if you want the best. Too bad Toms couldn't make the big brother work, but it doesn't surprise me after all the problems the X299 platform has been showing.

The die is about 25% larger, so it will cost more to fabricate.

The new Cofee Lake line replaces the KabyLake line, so I don't understand what "protect the still relatively new Kaby-Lake family" means. Intel almost always price the new chips in the same price bracket that the chips replaced. This was demonstrated before using the HEDT line: 6 core chips for same price than 4-core chips then 8 core chips for same price than 6-core chips then 10 core chips for same price than 8-core chips.

So what is this all this conspiracy again?

Hey, you learned a new word: "conspiracy"! Now you just need to learn how to use it properly. Reason being, I don't know why you think what I said is a "conspiracy". It's just an observation from the reviews circulating (I think I made it explicit I read it in Toms article).

But hey! New fancy word none the less!

Now, to the interesting bits: "The die is about 25% larger, so it will cost more to fabricate." <- I'm not an Intel fab worker, but I'm pretty sure it can't be a linear cost, otherwise they would have priced them *way* higher. It might be a few extra cents per working die (QA included)?

And this: "he new Cofee Lake line replaces the KabyLake line" <- I'll argue semantics, since it's not replacing it yet. They have a higher price range and different core count (among other less important and obvious differences). Plus, it seems marketing is making a very good point of making sure everyone gets a new MoBo for them.

I think your point boils down to "why question Intel's higher price tag for a new CPU?". My answer, in very good etiquette is to answer with another question: "why not?". Value is relative, but I'll keep it short and not delve into it too much.

EDIT: Forgot a word.

The fact that you think a CPU die only costs a few cents only proves that you have little to no understanding of CPUs.
Going up to 6 cores, even if it did only increase manufacturing cost by a few cents would still mean more chips fail quality control because there are more cores to fail.

Intel is releasing a chip that is likely a fair bit more expensive to produce and selling it for the same price as their old line. For the first time, they're taking a cut into their profits to compete. That's what we all wanted, isn't it?

But of course you've been posting nothing but hate for Intel since we started here so what could I expect.

 

[-Fran-]

So, for the costs. Would you like to share a breakdown of the Intel manufacturing costs? I'll just allow us to assume packaging and distribution will not be affected.

As for the replacement... Has Intel announced the phase out for KabyLake yet? Any official word on that, or you're just assuming since there's a new product that targets similar verticals it's an instant replace?

LOL.

And as for the MoBo, I don't care why they need to change it technically. This might as well be a business decision and not a technical decision. I do understand the pin-layout change, but all else equals, why adding two additional cores to the CPU package and bumping the speeds would necessitate a new *platform* in your opinion?

Why would I share a mere opinion when I can give tech facts? Slide #7 in the set given by jdwii explains in a plain language why a new chipset is needed for the new chips.

So you don't have a breakdown of costs to refute (either way) any claims?

And I don't remember the slides. I'll try and find them, but I bet it's the same dumb reasons AMD has given from FM1 to FM2 or AM3 to AM3+.

EDIT: Misplaced quote tags.
EDIT2:

The fact that you think a CPU die only costs a few cents only proves that you have little to no understanding of CPUs.
Going up to 6 cores, even if it did only increase manufacturing cost by a few cents would still mean more chips fail quality control because there are more cores to fail.

Intel is releasing a chip that is likely a fair bit more expensive to produce and selling it for the same price as their old line. For the first time, they're taking a cut into their profits to compete. That's what we all wanted, isn't it?

But of course you've been posting nothing but hate for Intel since we started here so what could I expect.

So you do have a breakdown of the costs? You're saying I don't understand the costs, but since you do, I would imagine you do have a breakdown?

 

[-Fran-]

Here are the reasons for the Platform change according to the slides (caps taken as is, lol):

-> IMPROVED POWER DELIVERY FOR 6-CORE PROCESSORS
-> ENHANCED PACKAGE POWER DELIVERY FOR OVERCLOCKING
-> MEMORY ROUTING SUPPORT FOR DDR4-2666

They read funny to me, but yes, layman terms indeed.

Cheers!

 

[goldstone77]

Manufacturing cost have decline with every node! I've already displayed this information here in this forum.
Click here for link

Click here for Intel's Annual Income Statement.
Don't forget to add 3 zero's (billions)

Intel has been able to make significantly more chips per waffer, and only recently has cost increased from R&D, which only amounts to a drop in 5 gallon bucket! How do you think the $1800 the 10 core 6950X a year ago can be sold as a 10 core 7900X for $962 this year, and still make a profit! The last 10 years Intel has not passed on the cost savings from manufacturing to the consumer like with of consumer electronics markets where you have 10+ manufactures like TVs! You are being sold melted, purified sand! Think about that for a min!