Intel Reports Second-Quarter 2018 Financial Results

spectre/meltdown: was it planned? Like all script put into a webpage, someone knows its there because they put it there ... just saying

So with Intel profits soaring (after all, toothpaste is cheaper than thermal paste) are we all going to start buying just Intel for gaming again? Or no?
 
I did say "consumer PC". I'm not ignorant to the diversification of Intel's business model, I just had hopes that a legitimate performance threat from AMD would make Intel do something other than heavy marketing on marginal increases to sell their products in the consumer space.

If you are saying that "nothing really happend for consumers from Intel Core 2 Extreme QX9770 -> i7-8700k" (about 10 years difference for consumer PC's) I think you are posting crap, sorry to say.
 
I did say "consumer PC". I'm not ignorant to the diversification of Intel's business model, I just had hopes that a legitimate performance threat from AMD would make Intel do something other than heavy marketing on marginal increases to sell their products in the consumer space.
AMD hasn’t changed the landscape, not by a long shot. What they did was get back into the game and gave us great price/performance. Nothing they did was revolutionary other than price bracket. That alone is not enough to sway an entire consumer market to switch brands.
 
In terms of IPC you are looking at a couple percent at best. While I do appreciate more cores, you aren't getting any real improvement per thread.

CB15 single thread score

Skylake i7: 182

Kabylake i7: 195

CoffeLake i7: 206

cb15-1.png


And compared to a similarly priced AMD Ryzen, you get less performance per dollar.

Price is not a linear function of performance. So the chip that targets inferior performance is going to provide more performance per dollar. Not a mystery here.
 
CB15 single thread score

Skylake i7: 182

Kabylake i7: 195

CoffeLake i7: 206

View attachment 92134

Huh. You failed to add these figures for single thread to argue a point of IPC. Wonder why?

Skylake i7: 182 Max Turbo Frequency 4.2Ghz

Kabylake i7: 195 Max Turbo Frequency 4.50 GHz

CoffeLake i7: 206 Max Turbo Frequency 4.70 GHz

Price is not a linear function of performance. So the chip that targets inferior performance is going to provide more performance per dollar. Not a mystery here.

Price may not be linear but the performance relative to turbo sure is.
 
Huh. You failed to add these figures for single thread to argue a point of IPC. Wonder why?

Skylake i7: 182 Max Turbo Frequency 4.2Ghz

Kabylake i7: 195 Max Turbo Frequency 4.50 GHz

CoffeLake i7: 206 Max Turbo Frequency 4.70 GHz

I am discussing performance, not IPC. I was discussing multithreading performance when I wrote

More than 50% performance increase from Kabylake to CoffeeLake is marginal increases?

and I was discussing single thread performance when I wrote

CB15 single thread score
 
We have one as a build machine. And our throughput has shoot through the roof compared to our 3 year old Xeon that cost an arm and leg. That AMD builds night and day and never goes down.

Definitely, memories... absolutely true. AMD had dual core, Intel did not. It was amazing..... until it wasn't.
 
Funny when you answer a question that starts with "In terms of IPC" and it ends up with "I am discussing performance, not IPC"

LOL

The original poster spugm1r3 whom I was replying was discussing performance, and I replied him with a post about how more than 50% performance increase from Kabylake to CoffeeLake cannot be labeled as "heavy marketing on marginal increases".

Then DigitalGriffin replied me with a post that begin mentioning IPC, but he continues with claims about performance "While I do appreciate more cores, you aren't getting any real improvement per thread. Not that many apps thread well across multiple cores. And you pay through the nose for those extra cores. So the only benefits to more cores is if you do heavy multi tasking."

Then I replied him, noticing that single thread performance has also increased from Skylake to CoffeeLake. Adding moar cores wasn't the only improvement. As I demonstrated single thread performance has increased by 13%.

Then you did enter the discussion making funny accusations about me.
 
They are funny because you specifically omit information to further whatever cause you have. Intel has ~2-28 cores in their process for the past few years, yet you throw out a 50% improvement number touting only one segment of their profile. Then when someone points out that in terms of IPC things really haven't improved tremendously over the period. You do it again.

Then you redeem yourself with a great post over here.

IPC is independent of node size...

It is what it is.
 
They are funny because you specifically omit information to further whatever cause you have. Intel has ~2-28 cores in their process for the past few years, yet you throw out a 50% improvement number touting only one segment of their profile. Then when someone points out that in terms of IPC things really haven't improved tremendously over the period. You do it again.

Then you redeem yourself with a great post over here.

So I am discussing multithreading performance. Someone moves the discussion to single thread performance (and he confounds IPC with single thread performance). And you first accuse me of moving the discussion away from IPC and now accuse me of omitting information

reverse-1275389857_naked-gun-facepalm.gif
 
No. We're discussing performance. His/their point was valid. You try and invalidate his point by showing something different. I point out the misgivings of that post. You post funny gif. (repeat)

If someone makes a point that encompasses more than just generic terms. AKA IPC. Accept it refute it, but don't convolute it.
 
Back
Top