Dolphin, the GameCube and Wii emulator - Forums

Peter Njeim Wrote:Coffee Lake and Core i9 was a DIRECT response to Ryzen and Threadripper.

To be clear, CPUs are not fast to produce, and this was in the works years before Ryzen even came out. There were definitely some changes to intel's plans (kaby lake and coffee lake would not be that close otherwise) but Intel didn't just make their chips a higher core count out of thin air because of Ryzen!

From Intel:

Quote:Note: Intel® Turbo Boost Technology 2.0 allows the processor to operate at a power level that is higher than its TDP configuration and data sheet specified power for short durations to maximize performance.

https://www.intel.com/content/www/us/en/architecture-and-technology/turbo-boost/turbo-boost-technology.html

If I'm remembering an AdoredTV video I watched over a week ago, this page didn't say that before the Coffee Lake launch or something similarly sneaky like that. If anyone cares to check, the link is here: https://www.youtube.com/watch?v=O98qP-FsIWo


The Multi-core enhancement thing was a bit weird as it seemed to be some of the biggest tech journalists and YouTubers who used it. They're the kind of people who'd receive their test samples for free, but I'm not sure if they got their motherboards via the board manufacturers or via Intel. If it was Intel supplying those boards, it would be very naughty to have it on by default. According to a Jayz2Cents video, ASUS told him that MCE was off by default on that board when he complained to them, despite it turning on when he reset it to factory defaults.

Do you guys all miss my WoT posts? Because I'm about to do one for the first time in years.

Peter Njeim Wrote:I see you are out of the loop on PC hardware. The jump 14nm+ to 14nm++ is simply for efficiency, the size didn't change.

Actually it did.  Check the gate pitch on both.  The size ratings for transistors that manufacturers post are kind of misleading since these are after all 3D objects.  For example AMDs current 14nm process is actually closer in size to Intel's 22nm than to their current 14nm process.

Please don't go straight to assuming I'm an idiot.

Peter Njeim Wrote:Also, you said it was pretty amazing that it can remain 65W for 6 cores 12 threads load. Have you every heard of Ryzen? The R7 1700 does 8 cores 16 threads 65W full load.

Ok....how does that have anything to do with what I said?  

I actually have a Ryzen 1700.  I still find 6 Intel cores at 4.3GHz at 65w pretty damn impressive.  You definitely cannot do that on kaby lake.

Peter Njeim Wrote:If you have been following hardware, you'd know that Coffee Lake was paper launched very early, and released half a year before intel wanted to release it.

Not sure I would call it a paper launch.  There is definitely a shortage due to them rushing it but a lot of people who want them were able to buy them.  I nearly bought one on black friday myself.

Peter Njeim Wrote:This was because Ryzen was too powerful and taking all of the market share for 2017 CPU's. Coffee Lake and Core i9 was a DIRECT response to Ryzen and Threadripper.

This is a common assumption I have seen paraded around the Internet.  There is simply no evidence to support it.

I would also like to point out that when OEMs are taken into account AMDs market share is still extremely small.  Ryzen isn't quite as much of a market disruptor as the armchair analysts online want you to believe considering most OEMs can't even sell them without integrated graphics and the retail market is only a couple percent of the total market at most.  I would also like to point out that it takes multiple years of project pipelining to get these things on the market.  The common idea that Intel quickly cobbled together coffee lake in 3 months simply isn't possible.  They rushed it a bit sure but implying it wouldn't have happened at all if ryzen hadn't come out is just not supported by the evidence.

Peter Njeim Wrote:I appreciate you trying your best to explain everything, but please don't be an intel shill, do a little research next time.

Calling me a shill and insulting my intelligence for explaining to someone the improvements that were made from kaby lake to coffee lake.  Jesus christ you can't make this shit up.  How do you even begin to think that insulting someone this way for something this trivial is a good idea?  Are you really THAT offended by what I said?

Peter Njeim Wrote:AMD is what made this year great for CPU's. I know you will say "but AMD sucked for the past 10 years" and that's because was being monopolistic and didn't allow AMD's better Opteron CPU's from being bought by Dell and HP.

Yeah it was definitely Intel's fault that they made them a better deal.  How dare public companies try to make money.

Peter Njeim Wrote:You can watch the AdoredTV video on YouTube called "A history of a Monopoly" with intel in the thumbnail, you'll see what I mean. AMD was losing so much money, they had to sell all of their fabs, and even their headquarters.

Yeah because they spent 5 billion dollars that they didn't have to buy a company valued at 1.5 billion.  But I'm sure that's Intel's fault somehow too and not the incompetent executives that were running AMD at that point and were quickly fired afterwards.

Now I remember why I stopped going to the hardware subforum.  I would just like to again point out that all of this was in response to a post where I told someone that coffee lake has more cores than kaby lake, that's it.

Peter Njeim Wrote:They used their GPU profits to fund both CPU and GPU divisions, and that money was less than a tenth of intel's. All analysts say that it was amazing how AMD survived, but they did, because AMD is amazing.

This is accurate.  They survived by downsizing massively and targeting thread level parallelism over instruction level parallelism (thread level is much easier to design).  Which was a smart move.

@AnyOldName3

Yes TB has always worked that way.  It can go past the TDP for short periods only.

(11-29-2017, 06:30 AM)NaturalViolence Wrote: [ -> ]Actually it did.  Check the gate pitch on both.  The size ratings for transistors that manufacturers post are kind of misleading since these are after all 3D objects.  For example AMDs current 14nm process is actually closer in size to Intel's 22nm than to their current 14nm process.

That's not strictly true either - the process is named after the smallest single resolve-able feature. Not any kind of average, or based on what features are commonly used for different tasks. Different processes size different things differently, and comparing "Single transistor type size" can be just as misleading as any other random measurement without context.

True. The point I was trying to make is that they are not exactly the same size.

(11-29-2017, 06:37 AM)NaturalViolence Wrote: [ -> ]True.  The point I was trying to make is that they are not exactly the same size.

Oh no, but one isn't "better" in every situation and the other "worse". Or that there's not big room for improvement despite being the "same" sized process. They're just marketing names for different processes.

I'm not sure I''m following you. If one process yields transistors that are smaller in every measurable way, have lower voltage leakage, and faster V transition time is that process not objectively "better"?

(11-29-2017, 07:31 AM)NaturalViolence Wrote: [ -> ]I'm not sure I''m following you.  If one process yields transistors that are smaller in every measurable way, have lower voltage leakage, and faster V transition time is that process not objectively "better"?

That would be true if all those statements were correct - but the top 3 processes (Intel, TSMC, GF & Samsung) all have strengths and weaknesses - even if it's just yields and costs, or targeting different points on the area/power/performance curves.

But it's not really "transistors" individually that you care about, the "process" includes a library of logic cells that are the building blocks actually used, and within them multiple options that may be focused on different things (optimised for power, speed, area etc.), plus placing and routing all those things takes significant effort (and skill), and can against be designed to maximise particular goals. So even on the "same" process, there's the opportunity to have vastly different characteristics of the end product. Plus it's a massive undertaking to harden a design to an actual silicon implementation on a specific process, especially to do it well.

The simple fact that Intel likely have the "best" (on average) process for (relatively) high-power, high-frequency desktop CPUs, but that's likely as much due to the tight integration with the people working on the silicon implementation and effort put into the design and layout as much as the physical characteristics of the end silicon gates. And they seem to be struggling to get the low power efficiency required for mobile devices (you could argue that's as much design issues, as you can't quite compare apples->apples as an atom isn't the same as an arm cortex - but these days x86 decode is relatively small so I somewhat doubt it's a fundamental issue with the ISA rearing its head).

And they're pretty much nowhere with GPUs - if they did have the 'massive' process advantage some people seem to claim, they'd just be able to power through any design deficit. And in CPUs the ryzen stuff shows the others can at least be competitive in many situations.

I think that was a rather long-winded way of saying "It's not that simple, it's really hard to directly compare processes".

NaturalViolence Wrote:Do you guys all miss my WoT posts? Because I'm about to do one for the first time in years.

Yay!

(11-29-2017, 06:30 AM)NaturalViolence Wrote: [ -> ]Do you guys all miss my WoT posts? Because I'm about to do one for the first time in years.

I missed these, thank you!