AMD Smoother than Intel?

[guskline]

I now own a 3930k, 3770k and FX8350. All are OC'd. I think smoothness has much more to do with the HDD (all three have ssds').

 

[ShintaiDK]

We all know VIA got the most smooth CPUs. Just dump your incredible unsmooth Intel and AMD chips and get VIA. Despite all data showing otherwise.

The smoothness part is the biggest BS ever to make up an excuse for a slower product.

 

[coercitiv]

I think smoothness has much more to do with the HDD (all three have ssds').

And SSD performance is affected by OS power management & CPU power management features.

I don't know how it played out in the desktop field, but mobile platforms certainly did need tweaking under Win 7 in order to get expected SSD random op performance.

 

[ShintaiDK]

A clean install will also feel much better.

 

[SlowSpyder]

I think back when AMD had an integrated memory controller and Intel didn't there seemed to be a subjective difference in responsiveness. These days I generally cannot tell a difference between them (assuming both have SSD's, too).

 

[podspi]

I'd imagine it depends on your use-case. I just switched from a 1055t to a 4670K, and the 4670K is noticeably faster in every case EXCEPT when I am doing a specific operation that uses exactly 4 threads. On the 4670K it saturates the CPU, and things get kinda slow. On the X6, I still had 2 extra cores (and multiple SSDs) so things went more smoothly.

That being said, the 4670K is just so much faster that it was more than worth it. The thing I gotta get done gets done much, much quicker, and everything else is noticeably faster.

 

[SPBHM]

I think back when AMD had an integrated memory controller and Intel didn't there seemed to be a subjective difference in responsiveness. These days I generally cannot tell a difference between them (assuming both have SSD's, too).

I don't believe in that, Core 2 memory performance was lower, but nothing absurd, the FSB speed was pretty high, the memory controller good...

http://www.bit-tech.net/hardware/cpus/2008/04/07/intel_core_2_duo_e8500_e8400_and_e8200/3

doesn't look to me to be to far off from AMD CPUs of the time in memory benchmarks, so I really doubt it would be a cause for a "less smooth" basic experience.

 

[Chiropteran]

All else the same, my 3770k system feels just as smooth or snappy as my older FX-8120 did.

However, interestingly thing I noticed when I first got the 3770. It came with a cheap 1TB hard drive. With that hard drive drive, it felt vastly slower than my 8120 which was running with an SSD. Once I upgraded the intel system to an SSD, it felt smooth again.

Despite the vast difference in CPU performance in benchmarks the FX-8120 and 3770 felt about the same "smoothness" in general real-world usage, only certain heavily taxing applications showed a difference. But swap out an SSD and use a regular spindle based drive, and the difference is huge and very noticeable in almost every application.

Moral of the story: if buying AMD saves you enough money to switch to an SSD, it's probably the better option.

 

[sm625]

It's all about drivers, motherboard utilites, and IRQ/DPC latencies. If I'm hammering my cpu hard enough, I can actually get the audio output to occasionally pop. This clearly means that the buffer isnt being filled before it runs out of data to send to the DACs. It's only for a few milliseconds, but you can tell that its getting hung up while trying to service multiple hardware interrupts at once. There is a good chance that I could switch to a different sound card, one with a different driver or at least a larger buffer, and the problem would go away. Or it could get worse even if the new card was newer and had larger buffers and all better specs. Sometimes it is windows doing it. Windows can place an audio device on the same IRQ as another device, and if they both interrupt often then windows has to interrogate both cards every time one of them interrupts. In such cases, the solution is as simple as moving the card to another slot. Or if they both share the same DMA channel....

 

[Enigmoid]

It comes down to drivers, motherboard/HW settings, and utilities.

CPU is pretty irrelevant.

 

[CakeMonster]

This reminds me of how people react to new or leaked graphics drivers. Already 15+ years ago people would go "It feels smoother than the last one" or "Image quality is better" when a new incremental beta driver for their Geforce or Radeon card leaked. If that had actually been the case, we would have orders of magnitude better looking games from that era now. Yet GLQuake, Diablo2, and HL1 look the same to me.

TLDR: Placebo is nothing new.

 

[pyjujiop]

An i7 2600K should always be faster than an X4 950, in any application or purpose. It's just a superior CPU across the board. The only way I can think of that the Phenom II might give a better experience is if you were running some sort of software that messes up on Hyperthreading. And in that case, just switch HT off and the problem would be solved. There's just nothing at all that a Phenom II X4 offers that could truly make it outperform a quad-core Sandy. The best Intel CPU's that it can beat without overclocking are Core 2 Quads. Even an i5-760 Lynnfield can keep up with a X4 965 that's clocked 600 MHz higher.

 

[BSim500]

Nobody is talking about games , we are talking about general computer use

What, like typing in Word, filling in a few cells in a small 50-row Excel budget spreadsheet, changing the font size of a slide in Powerpoint, listening to an MP3 in Winamp, watching a hardware-accelerated 1080p video in VLC / MPC-HC / Youtube, scrolling down the page in Anandtech or clicking "Check Mailbox" in Outlook / Thunderbird?... OMG, the difference is horrendous! LOL. Most of these don't even show up as 1-20% CPU usage of 1 core of even on a 3GHz Celeron/Pentium or 1-30% of an AMD A4, and are entirely subjective & prone to bias / placebo. In my experience unless you've got an iffy driver / virus / malware / trojan installed somewhere (or haven't bothered to remove the 40x odd "free cr*pware" apps preloaded onto many Branded laptops), most real-world "general usage responsiveness" difference is down to SSD vs HDD and massively variable ISP latency (for web browsers).

Placebo is nothing new.

One of the most enlightening jobs I've ever had when it comes to seeing inside the human mind was working in a high-end hi-fi store dealing with those funny "superman" audiophiles every day. I remember one came in loudly and proudly dictating a several hundred dollar budget for a simple 1m twin-phono interconnect. I took two cables out - a premium $499 one, and a cheap $5 one. "Oooh! that's MUCH better" he proclaimed after I told him I swapped them over and out came the usual buzzwords ("dynamic soundstage", "warmer tonality", "tighter timing", "superior transient response", "diminshed jitter", etc). Thing is I didn't swap them - I just unplugged and plugged the same $5 cable back in again (out of his line of sight)... :whiste:

In another double-blind test of a group of 20 "golden eared listeners" owning setups from $2k-$10k, it turned out that none could ABX a 96khz SACD / DVD-audio from a 44.1KHz CD, only 2 could ABX a $500 cable vs a $5 cable (and they both failed with the same $500 vs a better $30 cable). 1/3rd of them however, couldn't even tell the difference between a $500 cable and telephone wire, and 2 amusingly even failed to ABX a $500 cable vs a wire coathanger. Seriously. The oldest aged 58 couldn't hear over 14khz (due to natural Presbycusis). Man did they sulk like little children when they found out. :biggrin:

Trust me, if something cannot be objectively and repeatedly measured, you can take any surrounding emotional claims of "feelings" of an expensive piece of kit you've just bought with an almighty grain of salt. Arguing over "responsiveness" at 1-20% typical CPU load on either brand or 8 vs 4 cores for general office / internet apps which use half of one core is to me, a bit like arguing over those little $50 a piece plastic pyramids that lift the cable off the carpet so you don't get "carpet interference" with your already shielded and insulated digital cable. There might be some argument for running 4x x264 encoding threads whilst applying 30x PS CS6 filters, whilst trying to play Crysis 3 all at the same time, but no-one does that in real life. That's why I included the Skyrim benchmark - it's one thing that can objectively measure "smoothness" (which is not average fps, but rather a consistent minimum fps) via fraps without bias. I suspect the reason "it doesn't count for general computer use" is simply because it gives the "wrong" answer...

 

[AnandThenMan]

You can only judge when you own the different CPU's.

Exactly this.

 

[coffeejunkee]

Exactly this.

Maybe, but what if one person owning an Intel and AMD cpu told you the Intel cpu was smoother and another person owning the same cpu's told you the AMD was smoother?

For the record, I sometimes forget switching back from Windows power saver plan and usually only notice it when playing a game or doing video encoding. So with regular tasks I can't even tell the difference between 1.6 and 3.6 GHz.

 

[Pheesh]

If you're getting hitches,microstutters or sound popping, it's generally not the CPU, it's surrounding drivers/peripherals/HDD

You likely have a DPC latency issue. There is one CPU related aspect that can make the symptoms of DPC latency worse and that is certain power saving settings. You can always disable those in the bios/OS.
http://www.resplendence.com/latencymon

 

[TuxDave]

Well I moved from an AMD 64 x2 3800+ to an Intel i7 965. The Intel processor was waaaay smoother.

 

[blackened23]

lolz!

That was my first thought at reading this thread as well. Sometimes you can only sit back and laugh at this stuff.

 

[SecurityTheatre]

I would describe it like this:

A muscle car goes 0-60 in 4 seconds but its a choppy raw acceleration while still fast.
A luxury sports car may do 0-60 in the same 4 seconds with a smoother acceleration.

Speed can be quick and still have "microstutter" as you refer to it. Microstutter seems
like a good term for what I am thinking.

This is completely bullcrap. A CPU is nothing like a car.

In fact, general desktop operations don't tax the CPU very much, and those that do, the x86 instruction set does not vary. The "time to process" difference is on the scale of nanoseconds, which is 4-5 orders of magnitude (10,000 times) too fast to perceive.

This is either in your head, or it is the result of drivers not functioning properly on your Intel system.

Let me repeat. There is NO WAY a CPU can cause this sort of variation, based simply on the brand. Period.

 

[Gikaseixas]

there's no way to tell because a lot is going on inside a PC, every piece of hardware contributes for the experience, not just the cpu. I have both brands and both are smooth but i like a bit better the i7 experience in the end. The sata performance is the difference IMO. AMD needs to update their ancient mobo chip for the FX series.

 

[el etro]

Game smoothness and minimums performance are related to the per-thread processor performance.

 

[norseamd]

did you change hard drives or ssds

 

[KingFatty]

Intel beats AMD on benchmarks. No dispute from me. But 2 years ago when I switched from an AMD X4 950 to an i7 2600k, I noticed the system wasn't as smooth or fluid feeling. Has anyone else suspected that while Intel is faster, the AMD seems to run more fluidly and smoother? I am thinking about switching to an FX-6350 or FX-8350.

What was your usage when you noticed a change in smoothness? Opening/closing programs, browsing the internet, playing video files, gaming, booting up? Was it more noticeable in one scenario or another?

The huge problem here is it's impossible to change *ONLY* the CPU, because you are *FORCED* to also change the motherboard, and perhaps even the RAM. Also, you risk damaging electrical parts when you swap them between computers. Maybe when you upgraded your motherboard and CPU, you accidentally shocked the RAM with static electricity and damaged the RAM a bit for the Intel computer, causing the issue?

Anyway the point is that there are so many other reasons for the lack of smoothness, that are unrelated to the type of CPU, that are far more likely to be the cause.

 

[witeken]

The fastest CPU is by definition the smoothest, right?

 

[Homeles]

The fastest CPU is by definition the smoothest, right?

Unless you've got some sort of brand bias, yeah.