Jump to content

(Update: 12/26/14)Lets Put To Bed The Amd Fx Performance Rumors In Mwo.


171 replies to this topic

#41 Zuesacoatl

    Member

  • PipPipPipPipPipPipPip
  • Elite Founder
  • Elite Founder
  • 614 posts
  • LocationColorado

Posted 26 September 2014 - 01:33 AM

View PostFlapdrol, on 22 September 2014 - 12:10 PM, said:

Resolution only stresses the gpu, completely cpu independant.

Don't bring that logic in here, this is an intel fanboi thread. Logic has no place here lol. Down wit da AMD brah

#42 POWR

    Member

  • PipPipPipPipPipPipPip
  • Bridesmaid
  • 553 posts
  • LocationAarhus, Denmark

Posted 26 September 2014 - 01:41 AM

MWO is terrible on the technical front anyway. There's a constant amount of very serious pop-in and awful LOD management being done. My system, 4.2ghz i5 3570 + 680gtx, is being stomped down to 30fps at 2560x1440 when there are a few things going on. Zooming in on stuff with particles in the way will stomp it down in the sub 20 for brief periods, causing major choppiness.

Contrast this with Battlefield 4, which I can run at high/ultra at the same 1440p resolution, where I might get framedips in to the low 50s, occasionally spiking in to the 40s, but generally it runs at 60-80 fps. And that's on 64 player games. MWO needs some serious work done on its client optimisation and presentation. My experience with AMD FX series CPU in other games tells me they will do just fine if the circumstances are right (a.e. hardware balance) and the game is well made.

#43 Oderint dum Metuant

    Member

  • PipPipPipPipPipPipPipPipPip
  • Ace Of Spades
  • Ace Of Spades
  • 4,758 posts
  • LocationUnited Kingdom

Posted 26 September 2014 - 02:47 AM

The last two posts in this thread make my brain hurt.

#44 steverogers

    Member

  • Pip
  • 11 posts
  • LocationToronto

Posted 26 September 2014 - 03:51 AM

View PostDV McKenna, on 26 September 2014 - 02:47 AM, said:

The last two posts in this thread make my brain hurt.


Lol :) True that.

#45 Lordred

    Member

  • PipPipPipPipPipPipPipPip
  • The 1 Percent
  • The 1 Percent
  • 1,474 posts

Posted 26 September 2014 - 05:25 AM

View PostZuesacoatl, on 26 September 2014 - 01:33 AM, said:

Don't bring that logic in here, this is an intel fanboi thread. Logic has no place here lol. Down wit da AMD brah


What do you mean? Am putting up the facts about it only...

#46 MercJ

    Member

  • PipPipPipPipPip
  • Galaxy Commander III
  • Galaxy Commander III
  • 184 posts

Posted 26 September 2014 - 05:54 AM

Lordred, if you get a chance and you're willing to share, let me know what utility you're using to capture CPU/GPU load and activity. I've got a Kaveri APU (7850K) that I'm interested in getting some results on, to see if Steamroller cores end up making any difference :) I know they'd be different systems, but if I can match your settings hopefully our results will be somewhat comparable...?

And thanks again for taking the time to put these numbers up. Data takes time to gather, I appreciate seeing the numbers!

#47 xWiredx

    Member

  • PipPipPipPipPipPipPipPip
  • Elite Founder
  • Elite Founder
  • 1,805 posts

Posted 26 September 2014 - 06:04 AM

View PostZuesacoatl, on 26 September 2014 - 01:33 AM, said:

Don't bring that logic in here, this is an intel fanboi thread. Logic has no place here lol. Down wit da AMD brah


I've seen a decent amount of threads with actual proof of Intel's performance dominance in MWO, but every AMD person that has claimed to have just as good or better performance has never offered any evidence to back up their claims. If anything, you should call MWO itself an Intel fanboy. It has nothing to do with this thread, and if I had to guess a lot of people going Intel these days are AMD fanboys that are simply tired of not getting the performance they want (myself included, went from a Phenom II to Core i7, and now I'm on my second Core i7 since AMD still hasn't gotten their asses in gear... waiting for 2016).

#48 Catamount

    Member

  • PipPipPipPipPipPipPipPipPip
  • LIEUTENANT, JUNIOR GRADE
  • 3,305 posts
  • LocationBoone, NC

Posted 26 September 2014 - 06:14 AM

I ran AMD in MWO just prior to 12 mans, so sadly I never got to test my Phenom II X4 (capped on overclock from 3.4... to 3.6 :( ) against my 3570k, but even in 8 mans it was no contest. The Phenom II just didn't deliver, and the FX chips are generally worse. You need an FX-8XXX to even get a real quad core and feed the game four FPUs, and the FX series seems to have worse IPC across the board, especially per-core, than the old Deneb/Thuban chips managed.

When a Phenom II X6 is the peak of AMD performance in MWO, it gives some idea of how sorely lacking AMD is right now, and how big a mistake CMT is (though it might well pay off later since I bet they learned an awful lot about architecture design). To say Intel dominates is frankly an understatement. Intel dominates in all gaming, and where it doesn't make a difference, it's because they're titles where CPUs don't matter.

This is a foaming at the mouth rabid AMD fanboy speaking here, and I still went Intel for CPUs. I will on my notebook, too, when the new mobile hardware drops.

Edited by Catamount, 26 September 2014 - 06:15 AM.


#49 Goose

    Member

  • PipPipPipPipPipPipPipPipPip
  • Civil Servant
  • Civil Servant
  • 3,463 posts
  • Twitch: Link
  • LocationThat flattop, up the well, overhead

Posted 26 September 2014 - 06:25 AM

View PostLordred, on 22 September 2014 - 11:23 PM, said:

So running the 8350 as a true quad core did in fact have zero effect. Core utilization is up on the four cores because there are not any others to offset the load to.

Gonna call this myth busted pending further evidence.

Posted Image

http://mwomercs.com/...ost__p__3371858

I would like to know why your CPU readings are so agitated: Are you polling faster then 1 second per?

#50 Lordred

    Member

  • PipPipPipPipPipPipPipPip
  • The 1 Percent
  • The 1 Percent
  • 1,474 posts

Posted 26 September 2014 - 08:59 AM

View PostGoose, on 26 September 2014 - 06:25 AM, said:

Posted Image

http://mwomercs.com/...ost__p__3371858

I would like to know why your CPU readings are so agitated: Are you polling faster then 1 second per?



Yes, I was polling at 100ms, Would you recommend a much slower poll speed, which I would be perfectly happy doing.

I will admit I am still new to core logging, I have prior to this always done it by 'feel'


Edit: I am amazed by the 980 thus far, 4k is very run-able with it at reasonable frame rates.

Edited by Lordred, 26 September 2014 - 09:02 AM.


#51 Lordred

    Member

  • PipPipPipPipPipPipPipPip
  • The 1 Percent
  • The 1 Percent
  • 1,474 posts

Posted 26 September 2014 - 11:30 AM

Ok, I switched it up a little, this should be a cleaner presentation. I am now polling data at 2000ms (2 seconds) IF this looks good to you guys I will resume testing with this method / delivery

First up.

River City.

Settings / Config
Very High
Damage Glow on
Motion Blur off
PostAA on
Framerate limit set to 70

FX-8350 @ 4700mhz
Dram @ 2133 10-11-10
GPU GTX 980 (default)

Posted Image

Edited by Lordred, 26 September 2014 - 11:49 AM.


#52 Thorqemada

    Member

  • PipPipPipPipPipPipPipPipPipPip
  • 6,385 posts

Posted 26 September 2014 - 12:30 PM

I use an FX8350 at stock 4 GHz + a R9 280X and Win7 x64.
Resolution is 1920 x 1200 and Setting is Dx11 @ Very High (Cockpit Glass off as i do not like it) Vsync On.
With this settings you get between 60 and 35 FPS (up to 150 FPS with Vsync Off) depending on Screenaction but....

Most Players, even Players using Intel-CPUs tone down their settings and with only 2 Options set to low you get a 66% FPS increase and stay all the time at or above 60 FPS!

Its absolutely not necessary to buy a new CPU when 60 FPS is your target FPS !

I tested this many months ago and if i remember right it was Shadows and Postprocessing but i am not sure - you can test it easily yourself.

For me average 45 FPS is good enough and i kept all settings the highest.

Edited by Thorqemada, 26 September 2014 - 12:32 PM.


#53 Goose

    Member

  • PipPipPipPipPipPipPipPipPip
  • Civil Servant
  • Civil Servant
  • 3,463 posts
  • Twitch: Link
  • LocationThat flattop, up the well, overhead

Posted 26 September 2014 - 01:35 PM

View PostLordred, on 26 September 2014 - 11:30 AM, said:

Ok, I switched it up a little, this should be a cleaner presentation. I am now polling data at 2000ms (2 seconds) IF this looks good to you guys I will resume testing with this method / delivery

First up.

River City.

I'd just assumed 1 second was a/ the standard. :dunno:

I do hold River City (day) as one of the hard maps, but I've never gotten off my rump and made a list of which map is hard why: Frozen Colony is about Particles hate, for example …

#54 Lordred

    Member

  • PipPipPipPipPipPipPipPip
  • The 1 Percent
  • The 1 Percent
  • 1,474 posts

Posted 26 September 2014 - 01:42 PM

Updated with live data from a Forest Colony Snow run.

#55 ninjitsu

    Member

  • PipPipPipPipPipPip
  • FP Veteran - Beta 2
  • FP Veteran - Beta 2
  • 402 posts

Posted 26 September 2014 - 01:48 PM

Are you using msi afterburner to graph your data?

#56 Lordred

    Member

  • PipPipPipPipPipPipPipPip
  • The 1 Percent
  • The 1 Percent
  • 1,474 posts

Posted 26 September 2014 - 01:50 PM

I am using MSI afterburner, and importing the data into Open office spreadsheet, and creating the charts with that.

#57 Lordred

    Member

  • PipPipPipPipPipPipPipPip
  • The 1 Percent
  • The 1 Percent
  • 1,474 posts

Posted 26 September 2014 - 04:14 PM

Updated with Tourmaline run.

#58 Lordred

    Member

  • PipPipPipPipPipPipPipPip
  • The 1 Percent
  • The 1 Percent
  • 1,474 posts

Posted 26 September 2014 - 04:25 PM

View PostMercJ, on 26 September 2014 - 05:54 AM, said:

Lordred, if you get a chance and you're willing to share, let me know what utility you're using to capture CPU/GPU load and activity. I've got a Kaveri APU (7850K) that I'm interested in getting some results on, to see if Steamroller cores end up making any difference :) I know they'd be different systems, but if I can match your settings hopefully our results will be somewhat comparable...?

And thanks again for taking the time to put these numbers up. Data takes time to gather, I appreciate seeing the numbers!



Sorry I missed this MercJ. I am now using MSI Afterburner polling at 2000ms, I log the GPU Utilization, CPU Utilization, FPS, Gram and Dram.

#59 Jetfire

    Member

  • PipPipPipPipPipPipPipPip
  • Legendary Founder
  • Legendary Founder
  • 1,746 posts
  • LocationMinneapolis, MN

Posted 26 September 2014 - 04:50 PM

I run with MSAA, Shadows at LOW (this seems to have an incredibly overstated effect on framerate ~10-15 fps), post processing and environment on HIGH and the rest VERY HIGH, Vsync Off, otherwise the same and get a solid 40+ FPS with a 7970 Ghz and FX 8350 at 4.7 Ghz with 16 gb pc 2400 ram. I have found that almost every patch post DX11 my FPS stability and value have grown as well as the visual quality I observe qualitatively.

What I gather is what others have posited, MWO is a huge mess on the code side so with a ton of brute force you can ignore it and i7 chips are pretty good at that, but you pay for it. However MWO is refining its code iteratively and AMD chips can render a playable enjoyable high quality experience now, just not the best money can buy.

I did notice a 10-15% difference unparking cores though... Windows 8 auto parks 4 of them.

#60 Goose

    Member

  • PipPipPipPipPipPipPipPipPip
  • Civil Servant
  • Civil Servant
  • 3,463 posts
  • Twitch: Link
  • LocationThat flattop, up the well, overhead

Posted 26 September 2014 - 05:02 PM

May I suggest HWiNFO64 and GenericLogViewer? HWiNFO now can read fps from RTSS, which comes with Afterburner.





2 user(s) are reading this topic

0 members, 2 guests, 0 anonymous users