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

[frozentundra123456]

Nah, it's just the ADF trying to make their ~5 year old processors, that are still sold as new, seem relevant when they really aren't.

Agreed. Seems we go through this every few weeks. Although I have to admit I haven't heard the "it loses all the benchmarks but feels faster" for a while. Usually it has been constructing artificial scenarios running several intensive programs at once to try to make FX seem more competitive.

 

[DrMrLordX]

AMD CPUs = inferior in every respect to intel. :thumbsdown:

AMD cpu tech is stuck in 2008.

Um, no. Especially not taken in context of the OP's question. AMD chips are fine for general desktop use. My A10-7700k does the job nicely, handling CPU and GPU tasks with aplomb. You won't notice a difference between it and an Intel system costing twice as much until you start doing something intensive. Hell you don't even notice a difference at the desktop between 3.4 GHz and 4.7 GHz on my chip, despite benchmark and application performance showing marked differences.

Bottom line: desktop usage with a quad (or better) at any speed beyond ~2-3GHz is going to be about the same for everyone, especially if you have plenty of RAM and an SSD. I have 16Gb DDR3-2400 and an MX200 SSD, and it works nicely.

Its laughable really. Some people seem desperate to defend them and Im not really sure why.

What's laughable is how desperate people are to run them down, even when discussing a trivial task such as simple desktop usage (not referring to whatever multitasking scenario Abwx is discussing).

I would think any CPU with a decent clock these days would feel smooth for basic use, as long as you avoid AM1 you should be good.

Yes! Heck you can even do "okay" with AM1, if you go with the 5350, though those Kabini cores do make the OS (well, Win8/Win10) lock up pretty hard if you load them up 100%.

A turd is a turd no matter how much you polish it.

Again, we're just talking general desktop usage. AMD hardware dating back to 2009 will suffice, as will most/all of their modern processors (obvious exceptions: E1/E2, etc).

 

[2is]

Um, no. Especially not taken in context of the OP's question. AMD chips are fine for general desktop use.

Again, we're just talking general desktop usage. AMD hardware dating back to 2009 will suffice, as will most/all of their modern processors (obvious exceptions: E1/E2, etc).

The debate stems from the fact that the OP did not ask if they were fine for general desktop use. He asked if they were SMOOTHER then Intel's offerings. It's right there in the thread title.

 

[desura]

FWIW, OP here.

I did consider AMD CPU's when building my computer, but looking at the math, both the extra power draw and the expense of the PSU unit, I went with Intel. The extra $50 or $60 I paid for the Intel would be made up in lower power consumption over two years.

But, eh, just curious if AMD's approach had any real life merit. Then again, my experience with phones has been that Apple's dual core solutions run circles around Android's 8 core solutions.

 

[Gikaseixas]

He asked if they were SMOOTHER then Intel's offerings

Not top Intel chips but i5's and below the FX can challenge.

 

[MongGrel]

Agreed. Seems we go through this every few weeks. Although I have to admit I haven't heard the "it loses all the benchmarks but feels faster" for a while. Usually it has been constructing artificial scenarios running several intensive programs at once to try to make FX seem more competitive.

Hardly a weekend goes by without some kind of AMD vs Intel dog fight in here I guess.

I used to use AMDs a decade+ back, but I imagine I'll be using my second hand OC'd X5680/X5650 hexacores for awhile yet.

 

[SlowSpyder]

I have a couple AMD systems at the moment. The A8 5600k and FM2 board I picked up in FS/FT for $100 years back. With a SSD and 16GB DDR3 it is more than enough for my wife to play farmville 2 with her friends. Log in and do work from home, HD vids, normal stuff. System is snappy and responsive at all times. BTW, FV2 eats up way too much ram and CPU cores in Chrome for what it is.

I recently picked up a FX-8320e for $90 here in FS too. Added a MSI 970 Gaming for $80, and it has so far proved as good a 1080p gaming experience as the i5 4570 rig I have been using for months. Of course I had to overclock it to 4.5 GHz and buy a $25 cooler for it. AMD shipped coolers are worthless even running stock because they are so loud. Where as the i5 required zero extra time or effort to enjoy. I am able to use the exact same settings for Fallout 4 that I had already settled in on with the i5. Same system specs right down to the Zotac GTX970 except the FX has 16GB 1866 ram instead of 24GB 1600.

They both coast along at the 60FPS cap I leave on, then dip in the same places, like looking down into diamond city. The AMD dips lower, but not to unplayable levels. Playing on the FX stock with turbo enabled, however, had weird hitches. I could just be walking along and it would suddenly hitch and drop down into the low 20s for a moment. Then go back to 60fps with nothing changed on POV. Turn around and walk the same path and no hitch in basically identical circumstances. Perhaps due to the turbo implementation? Or the board/chipset and its drivers are responsible? As the i5 with turbo does not experience the hitches at all. Both have the steam folder on a 1TB black label. And the hitches vanished with the manual overclock.

That's more or less what I've said before, if you're ok with tweaking your CPU's settings, the FX can be a good alternative. These days the available AM3+ boards are starting to look a bit aged, though. But still, I think the FX have aged pretty well. I always buy an aftermarket cooler anyway, just picked up a 240mm AIO I would have gotten whether I got an Intel or AMD system, so that's a non-issue for me.

This

My FX 8350 is a beast for desktop, good for gaming but it does suck a bit of power. Good compromise IMO.

Using both an i5 3570k and an 8350 daily doing similar tasks I see differences. Basically anytime where I would max out the 3570k the 8350 does much better than it.

Oh no you di'ent! I just bought another FX, gonna hand-me-down my 7970 and stick one of the FX's (I'm keeping the one that clocks higher ) in a system for my nine year old. Should run Scrap Mechanic and Minecraft well!

 

[2is]

Not top Intel chips but i5's and below the FX can challenge.

FX hasn't been able to challenge i5's for the past few years in like 90% of tasks, what makes you think now that i5's are faster and FX is still the same chip they can suddenly challenge? They cannot. in all but some fringe obsecure test, and the magical "feels smoother" voodoo that no benchmark can seem to replicate, they are slower, produce more heat, cosume more power and are limited to an ancient chipset. The case for an FX was weak years ago, it's practically non-existant today.

 

[Gikaseixas]

FX 8 series have been trading blows with 2500k, 3570k and 4670k. Intel does have the upper hand in gaming but for multi threaded apps, it's a wash or a slight win for AMD.

Now if you include i7's, then i'll agree, no match period.

 

[aigomorla]

They can't match Intel's single threaded performance, but how do they fare for desktop?

this question has so much salt i dont know what to say.

First off, who told you they dont fare well in single threaded performance?
Just because they are not the fastest speed demons out there like a 6700k, it does not mean they are absolute junk.

They do very good in single threaded performance, they just take a little bit more time then an intel, and that time factor is not as great as you think.

How do they fare for a desktop?
See this is why i am asking are you being salty here?
AMD's make great desktops if your budget allocates to that.
Sure you can spend more and get an Intel, but when it boils down to it, did you require that power?

More cores does not translate into faster machine, or smoother performance as well.

 

[2is]

FX 8 series have been trading blows with 2500k, 3570k and 4670k. Intel does have the upper hand in gaming but for multi threaded apps, it's a wash or a slight win for AMD.

Now if you include i7's, then i'll agree, no match period.

Shall we trade benchmarks and count up the i5 vs FX8 wins? I've been down this road before and I can assure you, the FX8 will not win this contest. It traded blows with the 2500k and pretty much limited it's glory to that. It lost more often then not with a 3570k and that trend continued with it losing even more ground to the 4670k. Time has not been on AMD's side.

 

[SlowSpyder]

FX 8 series have been trading blows with 2500k, 3570k and 4670k. Intel does have the upper hand in gaming but for multi threaded apps, it's a wash or a slight win for AMD.

Now if you include i7's, then i'll agree, no match period.

Told ya.

I am in no way saying an FX is better than an i5, certainly not an i7, I think Intel CPU's are more balanced and at the top they have more absolute performance. But my actual experience using an FX is much more enjoyable than internet comments would make you believe. Just started Fallout 4, I'm a few hours in, smooth sailing with every single setting maxed (46-60FPS so far what I'm seeing), GPU generally at 99% usage. My FX isn't really holding me back yet.

 

[Phynaz]

The op should have gone to a tech forum like SA or Toms.

If a flaw was found on an intel chip here, 3/4 of the population would try to bury it.

FX 8xxx provides an excellent desktop experience. It only starts to suffer when you compare pure single app performance.

On heavy multitasking of light apps, the ssd and preloading common apps matters more.

I thought you bailed on this forum?

 

[mrmt]

Told ya.

I am in no way saying an FX is better than an i5, certainly not an i7, I think Intel CPU's are more balanced and at the top they have more absolute performance. But my actual experience using an FX is much more enjoyable than internet comments would make you believe. Just started Fallout 4, I'm a few hours in, smooth sailing with every single setting maxed (46-60FPS so far what I'm seeing), GPU generally at 99% usage. My FX isn't really holding me back yet.

Given that you had to overclock the hell out of your FX just to reach parity with a stock Intel processor, what's the point of it? Doesn't sound like a good deal to me.

 

[Vortex6700]

I thought you bailed on this forum?

I read much more than I write.

FWIW those crushing defeats on benchmarks are usually not even noticeable to the naked eye.

I would like to see what would happen in a double blind test.

 

[Timur Born]

https://translate.google.com/transl...r-Photoshop-und-Lightroom-1109093/&edit-text=

Thats with 30 RAW files from a D800 which are around 1.23GB in size.

Yes, but that is only for conversion, aka Export. This can be done unattended and is the one task where LR scales really well. Most of the daily editing and viewing tasks have you sitting and waiting. And these are the tasks where Lightroom does not scale so well (or in the case of face recognition, not at all).

I was mostly surprised to see how all the cloud synchronization tools are single core bottlenecked. Dropbox is even the positive surprise that uses a second core partly during heavy indexing (likely for stuff happening around the pure indexing task), the others don't even do that.

 

[myocardia]

FWIW those crushing defeats on benchmarks are usually not even noticeable to the naked eye.

Can you give an example of this? For instance, do you mean in gaming, where both are above 60 FPS? Or some other type of scenario, whether it's related to gaming, or not?

 

[desura]

Has anyone done a test to see how much of the professor is actually utilized for most tasks? With reports that most software isn't very multithreaded optimized I'd reckon that most of the processor is actually idle much of the time.

 

[superstition]

Playing on the FX stock with turbo enabled, however, had weird hitches. I could just be walking along and it would suddenly hitch and drop down into the low 20s for a moment. Then go back to 60fps with nothing changed on POV. Turn around and walk the same path and no hitch in basically identical circumstances. Perhaps due to the turbo implementation? Or the board/chipset and its drivers are responsible?

8320E has a TDP of 95 watts. Turbo mode uses APM which will throttle to keep the chip within its TDP.

If you run an FX chip for gaming you want to turn off APM. If you're overclocking your FX chip you want to turn off APM.

 

[Vortex6700]

Can you give an example of this? For instance, do you mean in gaming, where both are above 60 FPS? Or some other type of scenario, whether it's related to gaming, or not?

In gaming I have learned to ignore most benchmarks. They can be useful, but I haven't found a benchmark I can't beat without overclocking (ignoring stock overclock [fx9590]). My rig doesnt go under 60 on an unmodded game.

I have been told by multiple sites my cpu can't go over 60 on Arma 3 or DCS with "ultra" settings, and yet I do every week at 5760 x 1080. I don't think this makes me special. Instead I think it is the fact that no competent pc gamer is going to play a game before configuring it for thier pc.

The benchmarks I am specifically referring to in the earlier post are light apps (productivity, browsers, 7zip) because that was the op's topic.

Also, the performance of windows and windows apps on particular processors is up to Microsoft. There is a good 50% margin on os and app performance that comes down to how they developed and tested thier program.

I do not like Piledriver, and I have concerns about zen. But that is not enough to override my ethics and go intel for the interim.

 

[podspi]

Any source on that?
Cause Intel is claiming the complete opposite.
Foreground task will run normally as if alone and the second task will only get the idle cycles/commands.
If by low priority background task you mean it being set to Idle,after all that is in the hands of the user he can put it to low or below normal if he wants it to take away for the main task or he may choose not to.
https://software.intel.com/en-us/articles/performance-insights-to-intel-hyper-threading-technology

Actually, your link basically confirms what he is saying. Similar to AMD's suggestions for scheduling bulldozer modules, Intel suggests one active thread per core until all cores are utilized. They also note that all major OSes support HT, so this is typically not something the developer has to worry about.

 

[MiddleOfTheRoad]

Shall we trade benchmarks and count up the i5 vs FX8 wins? I've been down this road before and I can assure you, the FX8 will not win this contest. It traded blows with the 2500k and pretty much limited it's glory to that. It lost more often then not with a 3570k and that trend continued with it losing even more ground to the 4670k. Time has not been on AMD's side.

Oh give the bias a break for once. "Desktop experience" would likely include running Linux -- which is where an FX-8350 manages to outrun an i7 3770k for a lot of tasks.

Source:
http://www.phoronix.com/scan.php?page=article&item=amd_fx8350_visherabdver2&num=1

 

[2is]

Oh give the bias a break for once. "Desktop experience" would likely include running Linux -- which is where an FX-8350 manages to outrun an i7 3770k for a lot of tasks.

Source:
http://www.phoronix.com/scan.php?page=article&item=amd_fx8350_visherabdver2&num=1

Yea that's relevant...

https://www.netmarketshare.com/operating-system-market-share.aspx?qprid=10&qpcustomd=0

Give the hopes and dreams a rest. FX was never a "good" CPU and it's junk now.

Source:
Nearly every CPU review

 

[USER8000]

Yes, but that is only for conversion, aka Export. This can be done unattended and is the one task where LR scales really well. Most of the daily editing and viewing tasks have you sitting and waiting. And these are the tasks where Lightroom does not scale so well (or in the case of face recognition, not at all).

I was mostly surprised to see how all the cloud synchronization tools are single core bottlenecked. Dropbox is even the positive surprise that uses a second core partly during heavy indexing (likely for stuff happening around the pure indexing task), the others don't even do that.

A lot of tasks can be more limited by I/O - basically the speed of the disks you are using,and starting with Lightroom 6,Adobe are trying to offload some of the processing functions to the graphics card.

Oh give the bias a break for once. "Desktop experience" would likely include running Linux -- which is where an FX-8350 manages to outrun an i7 3770k for a lot of tasks.

Source:
http://www.phoronix.com/scan.php?page=article&item=amd_fx8350_visherabdver2&num=1

Shall we trade benchmarks and count up the i5 vs FX8 wins? I've been down this road before and I can assure you, the FX8 will not win this contest. It traded blows with the 2500k and pretty much limited it's glory to that. It lost more often then not with a 3570k and that trend continued with it losing even more ground to the 4670k. Time has not been on AMD's side.

Yet,when enthusiasts on forums are worrying on silly E-PENIS wrangling when one basic task is XYZ % "faster" or "slower" in some meaningless benchmark,every CPU company including Intel is concentrating on tablet chips,etc which are worse than most desktop chips,have limited I/O,crap graphics,etc.

Cause' 0.02 seconds doing one task has got to be "better" than 0.04 seconds doing another,right??

Look at all those Atom based tablets,Apple A9 based tablets and phones,etc and more people are doing their general computing on these kind of devices. Even the consoles use tablet chips.

Even look at some of the mission computers for the latest Mars Rovers and mapping satellites with their radiation hardened low power chips - the latest iPad probably has more processing power than one of them.

Most enthusiasts desktops are supercomputers in comparison.

 

[2is]

Oh give the bias a break for once. "Desktop experience" would likely include running Linux -- which is where an FX-8350 manages to outrun an i7 3770k for a lot of tasks.

Source:
http://www.phoronix.com/scan.php?page=article&item=amd_fx8350_visherabdver2&num=1

I have to LOL at this since I just actually took a look at the charts... I had assumed in my previous reply you were telling the truth and/or knew what you were talking about, but after further review one (or more) of two things must be true.

1) You are comparing and OC'd FX 8 to a stock 3770k
2) You failed to read that "lower is better" on multiple charts.
3) All of the above

Keep googling, and try to find a source that actually says what you think it says.