Intel: Chips Will Have to Sacrifice Speed Gains for Energy Savings

Well, it's a faster and more expensive a CPU at the end of the day, so you would expect to gain speed, but it also depends on how apps are designed.

I freaking hate JS, it's the main reason of slow loading websites, because browsers and CPU have difficulty to process all of the scripts. But perhaps we should blame dumb developers who don't seem to understand that less is more.

If I say that "Chrome renders the exact same pages faster than Firefox", you should already know that I'm not talking about network speed. And a fast CPU will of course help browsers to load heavy sites more smoothly, but the browser also has to be designed to take advantage of the CPU.

Older browsers like Opera 12 who can not always process JS and HMTL5 correctly, will make the CPU go berserk, no matter how fast the CPU. This will result in slower loading sites, and this hasn't got anything to do with speed.

 

It's definitely up to the devs. That's one of the reasons I like it. It's a very expressive language which means you can do things a lot of different ways. But that also means you can do things in very bad ways. If it wasn't for Javascript though, the internet would suck

 

Huh? No. The program needs to take advantage of the operating system. The operating system needs to take advantage of the CPU.

 

What's with the nitpicking? You know exactly what I mean, and without the OS apps can't run, that's a no brainer. But it's up to the browser engine to speed up loading and you also need a faster CPU for that.

Yes, there's nothing wrong with JS, but it's being used in the wrong way by dumb developers. Even the fastest browsers on the fastest machines sometimes struggle to load all of that crap. Just take a look at how heavy sites like Facebook and Amazon.com feel, they are clearly designed by idiots. For example, vk.com (the Russian FB) feels a lot smoother.

https://vk.com/dasha_pashkova

 

no I said intel have managed to avoid performance loss whilst increasing power efficiency. I didnt say they the only one's, actually the whole industry is going in that direction.

 

yeah I agree, an example is primagames online guides.

On my laptop if javascript is enabled, it overwhelms my laptop's cpu just to scroll the page, disable javascript and its smooth. Of course javascript is needed for the site to function properly, but they have done some bad coding for the scrolling performance to be affected so badly.

 

This is a technical forum where folks come to learn technical details. And no I don't know exactly what you mean because (1) I don't know you or your level of expertise, and you should not expect people should automatically know you. And (2) people spew BS in forums all the time trying obfuscate the issues. Why? I don't know. Perhaps it is to impress others when clearly, they don't know what they are talking about - or maybe they're just confused themselves and don't want to admit it.

I note on Monday you said, "a faster CPU will not improve web-page loading" and now you say, "it's up to the browser engine to speed up loading and you also need a faster CPU for that."

So sorry, but no. I don't know what you are talking about because it does not appear you do either.

 

You're taking nitpicking to another level, this speaks volumes and it's actually you who seems to be confused. Come on, it's not rocket science to figure out what I meant, but instead you kept babbling about network speed being the main bottleneck, while that wasn't even relevant in this thread. And I said: a fast CPU always comes in handy, but if apps do not take advantage of it, it won't help that much. But this is probably too hard to understand for you.

 

Yes, this is what gets on my nerves also. A faster browser and a faster machine will of course help, but not on all sites. A bit OT, but the first versions of Vivaldi used quite a lot of CPU time when loading pages, but they have improved this a lot.

 

It is when you say a CPU doesn't matter then turn around and say it does. It would be easier if you just said what you really mean instead of expecting others to decipher your flip-flops.

You cannot even get this straight. I did not say network speed is the "main" bottleneck. I said it is "typical" for it to be a bottleneck as not everyone can afford a fast connection. In many areas, a fast connection is not even available. So for many users, even those with lessor systems, network speed is indeed, "most likely a bottleneck". Network latency issues have been a plague since the beginning of networks. Network providers have their marketing teams running overtime with TV ads and mailings promoting how their "faster" networks prevent "buffering".

I am sorry you just don't seem to get that.

Now to prevent this from going OT further, I'm stepping out.

 

Bill,

You may 'step out' but I'll 'step in'.

I assume you live in the USA. I cannot speak for that country as a whole. But here it's fairly typical to have a download speed of 100 MBps. You need a lot for that speed to be a bottleneck. For regular browsing it's not an issue.

 

Good point.

Sadly, parts of the US are actually behind many other countries - in part because of our geographical size. But also, we have been isolated from the infrastructure destruction some countries have had to endure due to wars and natural disasters. As tragic as those events are, they mean these places have been rebuilt from the ground up with newer technologies. This is why some countries have all new super highways, underground utilities and more while ours is, in many cases falling apart - except where new development is occurring.

Many cities here are getting fiber to the home and are getting such speeds. But many are still rural and stuck on slow DSL or even satellite connections. There are still an estimated 2 or 3 million on dial-up!

I note in many poorer countries or poorer regions of many countries (parts of India come to mind), the only network access comes through village provided (if lucky) or private co-ops or company shared wifi "hot-spots". They have horrible bandwidth. In these case, it does not matter if you have a super computer, you are going to get lousy page loads.

 

Since apparently you do have to be a rocket scientist to understand my point, I will try to explain it one more time.

My main point: modern day CPU's (I'm not talking about Intel Pentiums) are at a point that they are fast enough for everyday PC usage (web browsing, video gaming, audio/video streaming), so IMO it's not a big deal if CPU makers will not focus on speed in the coming years.

That's why I made the comparison between the Intel Core i5 and i7 (Haswell), I have used them both and didn't see any true improvement, hence my comment about "a faster CPU won't improve stuff like web browsing". Of course, if you're switching from a 5 to 10 year old CPU, it will most likely improve things.

 

Yes exactly, when I talk about browser rendering in a thread about CPU speed, most people will already know that network speed is not relevant.

Like I said, Chrome outperforms Firefox and Opera 12 (regardless of network speed), leading to my conclusion that it's mostly up to the browser developers to speed things up. But a faster CPU always comes in handy especially if websites might become even more bloated in the future, but let's hope they won't. That's why I always use a script blocker, mostly for speed but also for privacy.

http://www.pcworld.com/article/2966...irefox-chrome-edge-ie-and-opera-compared.html

 

Here we go again.

The flip: "a faster CPU won't improve stuff like web browsing".

And then the flop: "But a faster CPU always comes in handy especially if websites might become even more bloated in the future".

Oh well.

 

Like I said, this comparison doesn't make sense in that context. Intel makes a 18c/36t monster but it'd get smoked by a 6700k in Javascript benchmarks. Heck, I'd bet that an overclocked Pentium G3258 would as well.

 

I really don't know what's so hard to understand. Perhaps you should rename yourself into Bill "Not So" Bright.

I said: For current web technologies, modern browsers (especially with script blockers) and CPU's are fast enough. No extra speed is needed, and I don't see websites becoming any more bloated.

See above, my point is we don't need speedier CPU's. Unless we will get to see new developments like new types of HTML Video + more bloated scripting technologies, and new types of video games that don't rely mostly on the GPU. But I don't see it happening.

 

Your point is all over the place. First here then there. Flip flop, flip flop. You claim something, then deny it.

This is why I said way back in my first posts that there are many "potential" bottlenecks in every computer. With the understanding that "potential" means something that could be, but might not be, these bottlenecks may include the CPU's capabilities, the GPU's capabilities, the ISP's bandwidth, the amount of RAM and more.

Note this distraction started by you saying, "the CPU still has a difficulty to process everything in a smooth way". And you also made the blanket statement that network speed has nothing to do with displaying webpages - that it is all about high CPU usage. Then claim CPU speed isn't important. I understand what you are saying. I am saying you don't make sense! And you don't make sense because EVERY computer is different as is just about EVERY browsing scenario. This means your claims suggesting CPU speeds never matter and Internet speeds never matter, that it is all about CPU usage, and any other absolute blanket statement you've made is just incorrect.

If you want to talk specifics - with a specific computer (OS, CPU, RAM, GPU, drives - and browser) and specific minimum network bandwidth and specific website, then fine.

Now you can follow the link in my sig to see if I might have a clue how network latency and computer hardware works. And you can accept it or dismiss it. I don't care. I see no reason to participate in this OT discussion further, especially with your on-going puerile denigrating personal commentary that adds nothing to the technical debate.

 

Funny enough, one of the big things with DirectX 12 is that it should allow significantly better scaling with multicore systems. I say should because it's ultimately up to the developers to take advantage of that but DX12 makes that way easier.

I think you're getting hung up on the fact that you believe it's up to web developers to make their pages faster and the browser engine devs to make their engines faster. Though they are partly to blame, there have been great strides with this stuff recently, especially between the various engines. But I'll point out that they can only do so much. PC gaming is fairly popular but you need quite a bit more power than a 'facebook machine' offers to play games properly. Are you going to blame that on the devs because their games need to be more optimized (though sometimes this is indeed the case)? Or are you going to blame it on the CPU and GPU being the bottlenecks there? It's the same thing. I do see value in your argument in that CPU's these days are fine for the typical computer user. But, in my opinion, more speed is always good. Excessive, maybe sometimes. But good. That said, do you think that if your CPU was twice as fast (frequency wise, everything else equal), you'd be able to tell?

 

Yes of course, speed always comes in handy, but that doesn't automatically mean we need it. The same goes for network speed, I've got a 150 Mbps connection, but do I need it, not really. But this depends per user. And don't forget what this thread is about, CPU's will still gain speed, but perhaps not as fast like back in the days.

And the reason why browser engines needed to become faster is because of all the bloat cooked up by web developers. And a badly programmed browser will still perform badly on a fast CPU. That was my point. But anyway, what type of CPU do you need for video gaming?

 

Too much text Bill. But seems like you're completely missing my point, even though I have tried to clarify. I'm not even going to respond to your babbling, because that would be a waste of time. LOL, you're talking about network speed being a bottleneck in a thread about CPU speed, and then telling me that I'm not making any sense, you're a funny guy.

 

This is very true too. But at the same time, game developers know most of their users don't have $2,000, $3,000 or larger budgets for $300 motherboards, 16GB of RAM, high-end i7, 2 (or 3) monster graphics cards, etc. So they are pretty clever and code their games to provide good "game play" on much lessor systems to increase their market share. Okay, the backgrounds may not be as rich and detailed and there may be fewer independent "objects" moving about, but a well written and developed game can be engrossing and entertaining just the same.

I agree. And it should be noted that one of the problems here is the type content can be displayed faster, depending on the browser used. For this reason, it is pretty easy to find a comparative review that shows your browser of choice, whatever that may be, is the fastest - because the testing criteria uses this content or that content, and browser A works best with this, and B with that.

Talk about the pot calling the kettle black. As Crusher noted, you are obsessed with webpage and browser development in a thread about CPU speed!!!! Contrary to what you may believe, the CPU is used for many more things than browsing the Internet!

And for the umpteenth time, I simply said network speed is but one of many "potential" bottlenecks. Something you at first flat out denied, but now flipped (or is it flopped - I've lost count) and suggest, as I have all along, it depends on the individual user and his/her circumstances.

What I find amusing is your suggestion that, after all these years, browsers are still "poorly programed" - I guess assuming performance is not a priority for Google, Microsoft, or Mozilla and the others. Same with "web developers" who apparently are intentionally adding performance-robbing bloat to their programs. Now that's funny.

 

Another thing I'd like to bring up. Because consumers PCs have become faster and faster, there has been a shift of where the processing happens with web apps. Instead of it all taking place server side, there's more and more of it happening client side. Maybe this is what you're noticing.

 

Wow... I'm not sure I should wade into this conversation, especially given that I didn't dive into all of the detailed back-and-forth in every case... BUT I think a few people could learn a lot about what actually causes slow webpage renders by playing around with an HTTP sniffer app like HttpWatch or educating themselves with Website page performance tools like those at Pingdom.

No modern browser app should be CPU capped by scrolling given the way developers can optimize scroll algorithms and operations now. Few websites are utilizing such heavy Javascript (or other client side coding) that these effect the vast majority of consumer users. Network bandwidth of your local ISP has almost *zero* direct correlation on your web page renders after you achieve a baseline rate of, say, around more than 5 Mbps (unless, obviously, you are sharing the connection with someone else at home that is saturating your bandwidth and you aren't actually getting 5 Mbps dedicated to your browsing). Rather, it's not bandwidth that is generally determinative, its latency... and users and ISPs generally can't control latency beyond a certain point. Latency is largely geographically driven (i.e., how close are you to the source of the content), as well as how CPU / IO loaded the server itself may be, or a key network component such as a router may be, in the path between you and the server.

Modern webpages are not rendered by one big pull of data from a server which could "clog the pipe" so to speak in terms of bandwidth from your ISP. Instead, many if not most webpages are comprised of anywhere from 15 to 30 (sometimes more! especially with ads) HTTP GET operations for all of the various components... HTML, CSS, JS, JPGs, GIFs, etc. Many of which cannot be done in parallel as they rely upon reading one prior to another. Almost none of this has to do with bandwidth (unless you are pulling very large graphics or streaming video) nor does it really have to do with CPU performance given today's baseline CPU capabilities.

Sure, more is generally better... so, yeah, a better CPU is good, more RAM is good, more ISP bandwidth is good, a faster hard drive / SSD is good, etc... but for the casual home PC user that is not playing a game, I would argue that the CPU is rarely the limiting factor. If a home user pulls up taskmgr.exe on their Windows PC and sees their CPU utilization pegged near 100% then they generally have some sort of malware running on their machine (seriously) or have too little RAM and their machine is page thrashing to virtual memory.

 

You can actually see this using the built in dev tools in Chrome or Firefox.
Using this page as an example, you can see that the images took the longest to load (except for one stylesheet).


I understand that this is usually the biggest factor. But the discussion was specifically about CPU speed with mention of specific examples such as usage spikes when scrolling and such, which is why I never brought it up.