Do AMD cpus at least give a smoother desktop experience w/more cores?

[TeknoBug]

Technically a Phenom II X4 955BE should still be sufficient today, I have one but can't use it anymore since the pins are bent and it won't work even when pins are straightened out- I do have a Phenom II X6 1090T laying around which works 100% but no motherboard to put it in right now since mobo the FX 4350 is in doesn't play nice with it due to being 4+1 phase. And speaking of the FX 4350, it works just fine for desktop use and OC's easily to 4.7GHz without any extra tweaks. For gaming is a different story.

And yup SSD changes everything. It makes my A4 5000 laptop bearable to use now.

 

[Phynaz]

To answer the original question, no they don't.

 

[frozentundra123456]

I find it hilarious that your entire post was a distortion of facts and you turned around accusing someone else of doing it.

Listen, AMD processors suck. I know it hurts your feelings, but I'm assuming you're an adult and it's something you simply a fact of life you need to get cozy with.

Saying they "suck" is a bit harsh. Like I said many posts ago, pretty much any big core cpu except for maybe single module APUs and the lowest end celerons are fine for daily light desktop use. That said, i would agree that intel generally offers better, more efficient solutions except for a few niche scenarios.

And to answer the original question, I dont think AMD offers smoother performance unless you construct artificial scenarios to load up lower end intel chips with unrealistic workloads.

 

[BurnItDwn]

I guess you should use a Macbook Pro.

No thanks.

Laptop offers no advantage over desktop for me.
The only real competitive advantage could be that Laptops are cheap to buy, and they use less power so the electric bill is cheaper. A macbook is expensive and inferior to a desktop, a lose lose proposition.

 

[MiddleOfTheRoad]

OSX is with 9,47% on the statistics I'm using, far above the Linux rounding error, and it operates in a very profitable segment of the market, so they are far more relevant than Linux.

The Chrome OS is just 0,5%, and not grouped with Linux. Unknown OS *may* be based on Linux, but not only Linux, at least his is what the results tell us. They can be anything, Android laptops and BSD based systems for example, or even a Linux modified enough to not look like what any of the distros would on the web.

This is just sad backpeddling on your part. Acting as if Android isn't also linux..... The bottom line is Vista and the original Windows 8 combined is roughly the same marketshare as desktop linux. Spin it anyway you want. People still use Vista and 8 to this day, so they are also relevant.

 

[MiddleOfTheRoad]

I find it hilarious that your entire post was a distortion of facts and you turned around accusing someone else of doing it.

Listen, AMD processors suck. I know it hurts your feelings, but I'm assuming you're an adult and it's something you simply a fact of life you need to get cozy with.

It isn't a distortion. I'm using the same charts he's quoting. If you actually paid attention to the forums, you'd know that I run about 75% Intel CPU's for the most part -- but there is very little need for it. Modern games rely on GPU's a lot more than CPU's. There isn't a single game out there that my retired FX-8320 couldn't play smoothly.

 

[jhu]

This is just sad backpeddling on your part. Acting as if Android isn't also linux..... The bottom line is Vista and the original Windows 8 combined is roughly the same marketshare as desktop linux. Spin it anyway you want. People still use Vista and 8 to this day, so they are also relevant.

Android is Linux, but not GNU/Linux (eg Debian, Arch, etc.). The userland is completely different. BTW, who even uses Android on the desktop???

 

[TheELF]

Android is Linux, but not GNU/Linux (eg Debian, Arch, etc.). The userland is completely different. BTW, who even uses Android on the desktop???

AMD users when they want a smooth desktop experience....



(just kidding)

 

[MiddleOfTheRoad]

Android is Linux, but not GNU/Linux (eg Debian, Arch, etc.). The userland is completely different. BTW, who even uses Android on the desktop???

To be honest, Remix OS is a pretty useful version of Android for the desktop. I'd actually rather run Remix OS than Chrome OS from what I've seen of Remix.

http://www.theverge.com/2016/1/6/10726986/remix-os-android-desktop-ces-2016

 

[TheELF]

Modern games rely of GPU's a lot more than CPU's. There isn't a single game out there that my retired FX-8320 couldn't play smoothly.

Smoothly isn't the problem,smoothly with high FPS is the problem.

 

[MiddleOfTheRoad]

Smoothly isn't the problem,smoothly with high FPS is the problem.

Considering the majority of recent game releases are bottlenecked at the GPU -- Intel and AMD give you the exact same FPS.

 

[Phynaz]

Considering the majority of recent game releases are bottlenecked at the GPU -- Intel and AMD give you the exact same FPS.

Not sure if serious or delusional.

OH, wait, you're comparing an i3 to an FX-8350 aren't you...in that case I'll agree with your statement.

 

[frozentundra123456]

Not sure if serious or delusional.

They aren't mutually exclusive are they?

 

[Phynaz]

True

 

[mrmt]

This is just sad backpeddling on your part. Acting as if Android isn't also linux..... The bottom line is Vista and the original Windows 8 combined is roughly the same marketshare as desktop linux. Spin it anyway you want. People still use Vista and 8 to this day, so they are also relevant.

Android isn't Linux. You can't get an APK and run on Linux, and you cannot run Linux apps on Android. Just because Android uses some element of the Linux kernel doesn't make it Linux. It would be akin to call OSX FreeBSD because they share the same origin.

And no, I'm not backpedaling on anything. It's you equating "Unknown" with Linux without any basis to do that. 1.47% is all we know *for sure* of Linux users, and that's just irrelevant.

Considering the majority of recent game releases are bottlenecked at the GPU -- Intel and AMD give you the exact same FPS.

Oh, please, stop. You are not making any easier for me to pick my new signature.

 

[Madpacket]

OP: your question is hard to answer as "smoother" is not really quantifiable without more context. However I'll try to answer you anyways in what I consider smoother. (less hitching, load times of applications, multi-tasking, gaming, etc).

If I'm interpreting what you mean by smoother then yes of course on modern desktop operating systems more cores will give you a smoother desktop experience than less cores. AMD FX smoother than Intel? No, not unless you're comparing a Core i3 or maybe an i5 in rare cases to an overclocked FX 8320E while having a boatload of applications fighting for CPU resources.

I currently own an AMD FX-8350 and I honestly can't tell the difference between it and my much more expensive Core i7-5930K or 4790K systems for regular tasks. Regular tasks for me being (40'ish tabs in chrome, 20'ish more in Firefox all with CPU intensive adblockers and privacy tools installed, VPN client running, Outlook Mail, Spotify streaming music, BOINC running POEM++, VMWare with a few VM's running, several large PDF's, Excel and Word files open, Anti Malware / Anti Virus, IRC client and usually a photo editor all open and running at the same time, etc).

I suspect it's because I'm never really maxing out any of these processors fully on the desktop and all systems are maxed out RAM wise and have extremely fast SSD's. I do notice a difference while playing some CPU intensive games but not to the point where I feel the 8350 is holding back my Geforce 970 or Radeon 390 from a decent gaming experience. I haven't done extensive testing with Crossfire Fury X's on anything but the 5930K but I suspect the 8350 would handle the cards just fine at 1440P/4K. Freesync and Gsync have helped mask some of the tremendous framerate swings Digital Foundry always likes to harp on so these "lower end" CPU's don't really affect the gaming experience if you have a decent monitor with these capabilities.

All that being said why is there so much hate towards the AMD FX processors around here? or actually AMD in general?

I understand AMD marketing made the FX chips out to be the best thing ever before release (that's what marketing teams are hired for, that's their job, they lie for a living, just like lawyers do) so were you an early adopter and bought one at an inflated price and feel betrayed that you were lied to?

Since release the market have decidedly put these FX chips at the prices they belong and currently they offer a great value if you want a PC that's not excellent at anything specific but pretty good handling most computing tasks provided you have an SSD and a decent amount of RAM (but that's true of almost any processor released in the last five years). If you happen to be lucky enough to live near a Microcenter they are stupid cheap (8320E + FX99 board) for ~$160.00 but even at NewEgg prices they're still a good value. DDR3 is still cheaper than DDR4 and the even some the cheaper FX boards have kept up feature wise (M.2 PCIe, USB 3.1, 1150 audio etc).

Anyways, it's natural to have brand preferences and bias (I fully admit I've written some stupid posts in this area) but I really question some of people on this forum who here who are super pro or vehemently anti (vendor of choice) and feel the need to constantly defend or attack other members for choosing a specific product. Maybe I'm just getting old but try being a little more objective?

My 2 cents.

 

[Exophase]

In multitasking the FXs trounce anything that is below the i7 and even this latter is no better, the 2 modules designs should be vastly superior to i3s :

http://www.computerbase.de/2015-10/intel-core-i5-6500-5675c-4690-test/6/

In some instances when FPU is used concurently to Integer the ST and MT perf of the FPU task collapse completely for CPUs below the i7..

Your own link with exactly two highly arbitrary multitasking benchmarks (that make no real sense for real world use) shows i5-2500K beating FX-8350 in one of the tests.

This is an especially bad set of tests because it doesn't even factor in how the secondary benchmark (WinRAR) performs. A change in OS scheduling policy like increasing priority for the primary benchmark could greatly skew the results, and that would be in the user's control. But what if they wanted WinRAR to complete as fast as possible in these scenarios?

 

[VirtualLarry]

Android isn't Linux. You can't get an APK and run on Linux, and you cannot run Linux apps on Android. Just because Android uses some element of the Linux kernel doesn't make it Linux.

Linux IS a kernel. GNU/Linux, or a Linux distro (kernel + userland) is different.

Android uses the Linux kernel, so therefore, Android IS a flavor of Linux.

 

[TeknoBug]

Android isn't Linux.

You have some lessons to learn.

 

[MiddleOfTheRoad]

Not sure if serious or delusional.

OH, wait, you're comparing an i3 to an FX-8350 aren't you...in that case I'll agree with your statement.

Well, yeah -- a 2012 era 8 core Vishera is only about as fast a dual core 2015 Skylake.
But hey, at least the Vishera can run Prime95 without crashing. I'd say that would qualify as the OP's "smoother desktop experience."

 

[MiddleOfTheRoad]

Android isn't Linux.

Thank you for my new signature.

 

[superstition]

No thanks.

Laptop offers no advantage over desktop for me.
The only real competitive advantage could be that Laptops are cheap to buy, and they use less power so the electric bill is cheaper. A macbook is expensive and inferior to a desktop, a lose lose proposition.

Goal posts moved.

My response was to the person who characterized laptops in a way that doesn't apply well to a Macbook Pro. The only exception is the heat management which is not an issue if you don't game and if you sit the machine atop a large cast iron pan if you're doing something very CPU heavy. In fact, it has quite good heat management given that it is completely silent for most tasks.

Otherwise it's a lot more convenient than a desktop (especially ergonomically), has a display that is plenty high-quality, etc.

 

[Abwx]

Your own link with exactly two highly arbitrary multitasking benchmarks (that make no real sense for real world use) shows i5-2500K beating FX-8350 in one of the tests.

You didnt catch the essential, it s an Integer based task and a FP task is added, in wich case the Integer task execution time increase by 50% for a i5 while the FP task throughput is more than halved, even a single thread FP task see IPC collapsing if an Integer task is loading the CPU concurrently.

That s it s Winrar + CB is of no importance, essential is the bahaviour when there s a multithreaded intgeger task + a single threaded or multitjreaded FP task..

And no, the i5 beat the 8350 in no task set apart in Cinebench ST in the Winrar + CB ST test, look at it s behaviour in Winrar + CB MT test, the CB score collapse by 76%.

This is an especially bad set of tests because it doesn't even factor in how the secondary benchmark (WinRAR) performs. A change in OS scheduling policy like increasing priority for the primary benchmark could greatly skew the results, and that would be in the user's control. But what if they wanted WinRAR to complete as fast as possible in these scenarios?

By changing the OS scheduling you will eventualy execute a task with priority while the other is waiting for free exe ressource but it wont change the picture, that i5s are very mediocre at multitasking.

 

[Yuriman]

^ Did they measure the time to complete both Cinebench and WinRAR, when they were run at the same time? I was under the impression they only recorded Cinebench, with "WinRAR running in the background". It may very well be the case that the FX chip maintained the same Cinebench and WinRAR throughput, and that the i5 dropped on both, but without the other have of the story (WinRAR), we can't know for certain.

 

[Exophase]

You didnt catch the essential, it s an Integer based task and a FP task is added, in wich case the Integer task execution time increase by 50% for a i5 while the FP task throughput is more than halved, even a single thread FP task see IPC collapsing if an Integer task is loading the CPU concurrently.

That s it s Winrar + CB is of no importance, essential is the bahaviour when there s a multithreaded intgeger task + a single threaded or multitjreaded FP task..

So you have a sample of basically one combination of benchmarks, there are many variables and a major part of the result is obscured.. you can't possibly conclude what you're concluding from this.

Rather than the very odd case of HT increasing performance by well over 100% that you'd never see anywhere else, it makes much more sense that for the duration that performance was measured WinRAR was simply scheduled on the four cores for more time than Witcher 3 was. And it's a fairly sensible policy that a new heavy task should take precedence over a long running one.

Having cores which are otherwise idle will force the scheduling policy to be more round-robin because they're nothing else you can really do until you run out of hardware threads. That'll look like a big advantage if you're looking at average performance over some limited interval for a task that runs indefinitely, where it's being de-prioritized, but if you look at combined execution time for two complete finite tasks the story will be different.

It's not even clear what they're actually doing though, nothing's really explained.

And no, the i5 beat the 8350 in no task set apart in Cinebench ST in the Winrar + CB ST test, look at it s behaviour in Winrar + CB MT test, the CB score collapse by 76%.

Except the first test on the page, also a "multitasking" test, where i5-2500K is ranked over FX-8350.

By changing the OS scheduling you will eventualy execute a task with priority while the other is waiting for free exe ressource but it wont change the picture, that i5s are very mediocre at multitasking.

Realistically speaking, in terms of total throughput over time they will be at worst around 75-80% of their hyper threading i7 counterpart; you know, the gains from HT that are actual realistic over pretty much any test ever. They won't be some miserable fraction that you think they will and these exotic mixes of threads probably doesn't change that.