Intel Skylake / Kaby Lake

[dullard]

So do you agree with me that there is a sizeable customer base that affords and desires HEDT but is also price sensitive?

No. The number of people who need HEDT is small.

 

[blue11]

The notion that there is going to be any sort of "desktop revolution" is laughable. The best case scenario for Zen on the desktop is that it manages to steal some market share from Intel and contributes to Intel hastening the shutdown of Client Computing Group. Anyone expecting Intel to "innovate" in a declining market, competition or not, is going to be sorely disappointed.

The future for clients is neither AMD nor Intel, but rather Apple, Samsung, etc.

 

[Eug]

So do you agree with me that there is a sizeable customer base that affords and desires HEDT but is also price sensitive?

I would guess that if I polled 100 of my colleagues and friends about HEDT, including several of whom who buy a lot of computer equipment, not a single one would have any idea whatsoever what HEDT actually is.

Compared to their mainstream chips, this market is tiny.

What remains to be seen though is if Ryzen in general can capture more mainstream customers.

 

[nathanddrews]

My vision of the HEDT market has always been a self-fulfilling market: it's YouTube PC reviewers and Twitch streamers, the very people that make money using the platforms they push. They need all the cores and speed for video encoding or streaming. Everyone else has an i3 or i5.

 

[coercitiv]

No. The number of people who need HEDT is small.

Compared to their mainstream chips, this market is tiny.

I was talking relative to the HEDT addressable market, not the entire desktop market. It was claimed that HEDT customers are price insensitive, because they are either wealthy enough or enthusiast enough to desire only the best. For example, they would automatically choose 6900K over 6800k or 6850k.

Do you also think most people who could use a HEDT system don't care about the cost of their build?

 

[dullard]

I was talking relative to the HEDT addressable market, not the entire desktop market. It was claimed that HEDT customers are price insensitive, because they are either wealthy enough or enthusiast enough to desire only the best. For example, they would automatically choose 6900K over 6800k or 6850k.

Do you also think most people who could use a HEDT system don't care about the cost of their build?

Your post actually referred to the entire world, not just the desktop market. I was just saying that it is not a sizable market. The number of people who want more than a 7700k but less than a professional workstation are small. This small subset generally is often not in it for price as the main consideration. Of course, some of them are concerned with price, but that is often a secondary or tertiary consideration.

I don't have a list of Intel processor sales for each chip (does anyone?). But there are plenty of websites that list which chips have used the website. For example, take this one: http://www.cpubenchmark.net/share30.html

The 6900k is nearly triple the number of users on that website (0.8%) compared to the 6850k (0.3%) or 6800k (0.3%). Thus, yes, most HEDT users of that website did automatically choose the 6900k even though it was far more expensive. That is limited to users of that website, yes. But I bet it is indicative of HEDT users as a whole.

 

[coercitiv]

I don't have a list of Intel processor sales for each chip (does anyone?). But there are plenty of websites that list which chips have used the website. For example, take this one: http://www.cpubenchmark.net/share30.html

The 6900k is nearly triple the number of users on that website (0.8%) compared to the 6850k (0.3%) or 6800k (0.3%). Thus, yes, most HEDT users of that website did automatically choose the 6900k even though it was far more expensive. That is limited to users of that website, yes. But I bet it is indicative of HEDT users as a whole.

On that website the 6900K is just as popular as many of Intel's recent mainstream desktop and mobile processors. Care to comment?

 

[dullard]

On that website the 6900K is just as popular as many of Intel's recent mainstream desktop and mobile processors. Care to comment?

You can't honestly compare HEDT to desktop that way. The best you can do is to compare within a class.

 

[LTC8K6]

You're the one suggesting people don't buy Intel's fastest chips because they can't afford them, whereas many of us have pointed out that CPU speed is just one of many factors to consider. If your contention made any sense, the 1% would only ever buy the fastest CPUs but obviously that is not the case. Most people just don't care as much about raw CPU speed as you seem to.

I didn't suggest that at all...it was wildhorse2k...

 

[coercitiv]

The best you can do is to compare within a class.

Fair enough.
5820K 0.2%
5930K 0.1%

But maybe you're right. Maybe the performance over price bias is so pronounced within this segment that 5930K owners already migrated towards BW-E.

 

[Topweasel]

I think the point was that people who have the money generally buy the fastest parts. They generally don't bargain shop.
Also people who want bragging rights and benchmarks generally don't bargain shop. They willingly pay more for that 10% extra performance.
There are a few exceptions of course.

People in the Porsche/BMW/Audi and up brackets generally don't even look into the Ford and Chevy lots.

For a short time now, RyZen has captured a bit more of the market, with Intel caught between chips.

We will have to wait and see how long that lasts.

True but as people say. For a hobby to be performance oriented PC's are one of the cheapest hobbies to get into. Therefore not everyone buying HEDT are people that can just look at a CPU and say F'it. Most people even the ones that have made it good after spending their highschool days putting PC's together from the scraps that they could pick up, still recognize a bad deal when they see one. There is a reason why the people lining up for the 6950 is so much smaller than the 6900, which is so much larger than the 6800 and 6850. With in the line up of more cores, memory, and IO. The 6900 offer more for the money than a 6850, and tons more value than the 6950.

ThreadRipper at a reasonable cost would shake that up by offering those same resources and more for the money. That's not to say a lot of people won't take wildhorse's train of thought and run with it. People still like value more than hate it. If I can get more computing power for equal or less and it's in the range of what I am looking for I'd purchase that. The better comparison would be that i7 EE stuff is the BMW, ThreadRipper is the Audi, Ryzen is the Chevy, and the base i7 is a Toyota. There will still be Beamer guys, like there will still be Intel only people that will just buy "the best", but I think there are a lot of people that just care about how fast the CPU is in the stuff that they do at home and will jump on the best option for them. ThreadRipper's advantage is that most of these tasks are highly MT'd which gives them a heads up since they have more cores than Intel can supply for a while.

 

[LTC8K6]

So do you agree with me that there is a sizeable customer base that affords and desires HEDT but is also price sensitive?

They are price sensitive, but less so than your average computer user/builder, imo. Not to the point of sacrificing too much performance, imo.
What is the boundary level for deciding? If chip A is $500 and chip B is $750, how much faster does chip B need to be to handily outsell chip A, all else being equal?
5% is clearly not going to get many takers.
10% would definitely get some takers.
15% would probably get a lot of takers.

What if chip B is $1K instead?

How about if both chips move up a notch to $750 and $1K?

Is there a point where the price doesn't matter anymore because both chips are so expensive so you might as well just go with the fastest one?

If you are spending at least $750, does $1K seem all that much?
Or maybe the system is so expensive that the CPU cost isn't as big a factor?

I guess we all have to decide for our own situation.

 

[LTC8K6]

True but as people say. For a hobby to be performance oriented PC's are one of the cheapest hobbies to get into. Therefore not everyone buying HEDT are people that can just look at a CPU and say F'it. Most people even the ones that have made it good after spending their highschool days putting PC's together from the scraps that they could pick up, still recognize a bad deal when they see one. There is a reason why the people lining up for the 6950 is so much smaller than the 6900, which is so much larger than the 6800 and 6850. With in the line up of more cores, memory, and IO. The 6900 offer more for the money than a 6850, and tons more value than the 6950.

ThreadRipper at a reasonable cost would shake that up by offering those same resources and more for the money. That's not to say a lot of people won't take wildhorse's train of thought and run with it. People still like value more than hate it. If I can get more computing power for equal or less and it's in the range of what I am looking for I'd purchase that. The better comparison would be that i7 EE stuff is the BMW, ThreadRipper is the Audi, Ryzen is the Chevy, and the base i7 is a Toyota. There will still be Beamer guys, like there will still be Intel only people that will just buy "the best", but I think there are a lot of people that just care about how fast the CPU is in the stuff that they do at home and will jump on the best option for them. ThreadRipper's advantage is that most of these tasks are highly MT'd which gives them a heads up since they have more cores than Intel can supply for a while.

I think AMD has to make some money back at some point, and threadripper just might be priced for profits.

 

[wildhorse2k]

AMD is ruining its profits bad enough by selling unlocked CPUs that get OCed and reach almost the same frequencies like their top processors, because they actually can't produce bad processors at all. So they are getting peanuts while customers get good value. Now that CPU performance doesn't increase very fast per year, market share may be something not worth more than actual profits as people will not be upgrading from those Ryzens 1700 for a long time.

 

[blue11]

AMD is ruining its profits bad enough by selling unlocked CPUs that get OCed and reach almost the same frequencies like their top processors, because they actually can't produce bad processors at all. So they are getting peanuts while customers get good value. Now that CPU performance doesn't increase very fast per year, market share may be something not worth more than actual profits as people will not be upgrading from those Ryzens 1700 for a long time.

AMD (Intel too) is in a bad spot. Even if they succeed in retaking the client market from Intel, they are still stuck with a declining/terminal market, and unlike Intel, AMD does not have a significant product line outside of PC. Intel is at least trying to survive the PC market extinction by diversifying to non-CPU product lines (e.g. memory and connectivity), although the effectiveness is yet to be seen

 

[LTC8K6]

AMD (Intel too) is in a bad spot. Even if they succeed in retaking the client market from Intel, they are still stuck with a declining/terminal market, and unlike Intel, AMD does not have a significant product line outside of PC. Intel is at least trying to survive the PC market extinction by diversifying to non-CPU product lines (e.g. memory and connectivity), although the effectiveness is yet to be seen

There is always that looming in the background. The PC market is doomed. Just how much money can you afford to put into desktop cpus these days?

Can you even hope to get the investment back?

 

[Topweasel]

I think AMD has to make some money back at some point, and threadripper just might be priced for profits.

All these CPU's are priced for profits. It's going to be a guessing game really. They keep using "Disruptive" pricing. Which can mean anything from being priced a lot less or matching their pricing while offering more. I mean realistically in comparison to the 6950, the 16 core with over 50% the cores and possibly near the same clock speed would be a steal at $1700. It's not about about selling at a loss or forsaking profit, but knowing where you sit with your product and selling at value to increase demand.

I was just talking to a friend about HD-DVD and Blu-Ray today. It was in regards to UHD BD being "Premium" a lot longer than Blu-Ray was. Sony promised $1K+ ASPs like Laser Discs. But when push came to shove they had to win and by bringing out the PS3 with BD at $600, a huge portion of its early buyers weren't gamers but people wanting High Def video system at a value. Blu-Ray movies soared in sales after the PS3 release and even though they system as a gaming system struggled the whole gen, it won the the war by being a value priced player amongst its competition. Not cheap and the movies were still 3 times the price of DVDs. So it wasn't for people with budgets. What people saw was great price in comparison to its less capable competition (other BD units) and jumped on it. The PS3 isn't a great example because Sony was selling it at a near $200 loss at the time. But the point stands AMD can still price it well into high profit margins as long as they provide enough value.

But let's not kid ourselves. AMD needs actual purchasers. Nothing worse than having the most pimpest and extreme 16C ThreadRipper if no one is purchasing it. They have a lot of poor history, lots of consumer hesitation, and lots of Intel buying bias to work past (how big of a difference in Price when performance is equal before a user would buy AMD). AMD needs profits, they need ASP, but they need actual purchases as well. That's why they are always on about Disruptive pricing.

 

[wildhorse2k]

I don't think PC market will go extinct. The problem is a.) we already have more CPU performance than we need b.) CPU performance is increasing very slowly, thus customers are postponing upgrades.

But those systems need to be upgraded periodically. It will be interesting to see if 2 companies can survive on the market. AMD need to become profitable very fast and not sell CPUs for scraps.

 

[Topweasel]

I don't think PC market will go extinct. The problem is a.) we already have more CPU performance than we need b.) CPU performance is increasing very slowly, thus customers are postponing upgrades.

But those systems need to be upgraded periodically. It will be interesting to see if 2 companies can survive on the market. AMD need to become profitable very fast and not sell CPUs for scraps.

This will be the point that makes or breaks pricing. What is scraps? I mean reasonably a Zeppelin die costs what $50? So two for ThreadRipper is $100. Once you are making a profit, demand is just if not more important than selling price. There is a curve to sales and while the profit of $100 chip selling for 2k is pretty tasty and something Intel enjoys right now. It's likely AMD could sell more than twice the amount of those CPU's at 1k. I don't think the 16c parts will sell that cheaply but the point still stands. A R7 still has $250 of margin at it lowest and just 1 million of those sold is still 250 million of profit If they sell this at $900 profit at 10,000k sales vs. 2000 at $1.9k profit. That still 9 Million in profits versus 3.8. AMD needs high revenues, higher cash flow, and yes higher ASP. But the ASP isn't nearly as important.

 

[tamz_msc]

a.) we already have more CPU performance than we need

Who's "we"? The people who are looking forward to Skylake-X? Mind you that among them are people who went from a 6950X to a 7700K.

 

[frozentundra123456]

This will be the point that makes or breaks pricing. What is scraps? I mean reasonably a Zeppelin die costs what $50? So two for ThreadRipper is $100. Once you are making a profit, demand is just if not more important than selling price. There is a curve to sales and while the profit of $100 chip selling for 2k is pretty tasty and something Intel enjoys right now. It's likely AMD could sell more than twice the amount of those CPU's at 1k. I don't think the 16c parts will sell that cheaply but the point still stands. A R7 still has $250 of margin at it lowest and just 1 million of those sold is still 250 million of profit If they sell this at $900 profit at 10,000k sales vs. 2000 at $1.9k profit. That still 9 Million in profits versus 3.8. AMD needs high revenues, higher cash flow, and yes higher ASP. But the ASP isn't nearly as important.

It's not that simple. There is far more to the cost of a cpu than the cost of the chip itself. There is research and development, validation, marketing, distribution, etc. Now I have no idea how these costs are assigned to ryzen, but ultimately they must be paid. So the real profit on a chip is far less than just subtracting the die cost from the selling price.[/QUOTE]

 

[Topweasel]

It's not that simple. There is far more to the cost of a cpu than the cost of the chip itself. There is research and development, validation, marketing, distribution, etc. Now I have no idea how these costs are assigned to ryzen, but ultimately they must be paid. So the real profit on a chip is far less than just subtracting the die cost from the selling price.

[/QUOTE]
I know that and I wasn't trying to make it that simple. But regardless you once you cover the actual production cost it's all about intersecting sales price with demand for maximised profit. It doesn't do AMD any good to price a CPU out of sales for the sake of ASP.

Now the reverse end is if there is a cap to their production they need to make sure they stay above a certain price per chip to be profitable. But they have to sell those chips for that to matter. This is the hell AMD has been in. Can't sell it high enough for the company to stay profitable, but high enough that they were making money on sales. Anything Ryzen 5 and above should be priced enough for AMD to make more then enough profit assuming they sell what they can make.

 

[imported_ats]

My vision of the HEDT market has always been a self-fulfilling market: it's YouTube PC reviewers and Twitch streamers, the very people that make money using the platforms they push. They need all the cores and speed for video encoding or streaming. Everyone else has an i3 or i5.

Most actual streamers end up moving to a dual computer setup: one computer for pure gaming and an additional computer for encoding/broadcasting/etc. The requirements for the streaming PC are pretty low and as long as they can support the capture card, it works fine. What usually ends up happening is that as they replace their main rig, it moves down to being the streaming rig. This setup tends to work much better than doing it all in one rig and is preferred by pretty much all professional streamers. Secondary also tends to double as their editing rig as well.

 

[misuspita]

Yes, but that is happening because one capable system meant going HEDT route for Intel, which was more expensive. Now, when there is Ryzen 7 on the market, you can keep only one system, one monitor, less space occupied, etc, etc.

The need for the second system was an effect, not a cause. Why keep 2 windows installs and edit on the lesser secondary system when you can benefit from the power of the better one?

For me that sounds like excuses. If you have a 6900 or 6950 PC, you don't need another. Same for R7 and Threadripper.

 

[mikk]

i7-7900X (@4.0 Ghz): http://ranker.sisoftware.net/show_r...d5e3daebdfe9dbfd8fb282a4c1a499a98ffcc1f9&l=en

For comparison
i7-6950X (@4.50 Ghz)= http://ranker.sisoftware.net/show_r...d5e3daebdfe9dff98bb686a0c5a09da88efdc0f8&l=en
i7-6950X (@3.50 Ghz)= http://ranker.sisoftware.net/show_r...d5e3d5e7dfe9dcfa88b585a3c6a39eae88fbc6fe&l=en

i7-7900X 10C (@4.0 Ghz)= 1386.94Mpix/s
i7-6950X 10C (@4.50 Ghz)= 897.28Mpix/s
i7-6950X 10C (@3.50 Ghz)= 746.64Mpix/s


7900x is called i7 here, possibly the i9 branding was a last minute change from Intel. Based on the scores there is no full speed AVX 512 involved I would say (as we already know). The difference is still big, +55% to a faster clocked 10C Broadwell and 86% to a lower clocked Broadwell 10C.


edit: Not sure about AVX 512 to be honest, I mean +70% at the same clock and core count seems too high.