41 replies to this topic

[Narcissistic Martyr]

I've got an fx8320 set at 4GHz aka fx8350 settings with a CM212+. It runs nice and cool, isn't too loud, and gets the job done. What more do I need?

[Az0r]

I gained around 10min fps (40fps to 50fps) from my OC

[Catamount]

TheFlayedman, on 14 January 2013 - 12:46 PM, said:

A stock 3750k is a bottleneck to MW:O now. Just look at the performance charts that wete lifted from the russian site

It's not that simple. At what settings and resolution is this true? When paired with what GPU?

Whether a computer component bottlenecks a system isn't simply dichotomous; system's are bottlenecked by whatever the slowest component is that's involved in what you're doing, which, in gaming, depends highly on how you're doing it. Play MWO on a middling GPU at a good eyefinity resolution and the highest settings, and almost no CPU would bottleneck it since any GPU would already be giving miserable framerates (I know someone who has a hard time keeping MWO smooth with 3 monitors on a GTX 670); play at 1024x768 on low, with a very powerful GPU, and no CPU, at any clockspeed, would be able to keep up.

Narcissistic Martyr, on 14 January 2013 - 05:05 PM, said:

I've got an fx8320 set at 4GHz aka fx8350 settings with a CM212+. It runs nice and cool, isn't too loud, and gets the job done. What more do I need?

Cupcakes.

Edited by Catamount, 15 January 2013 - 06:51 AM.

[TheFlayedman]

At 1080p which is the resolution most gamers use.

[Honey Badger]

i5-3570k
ASUS Maximus V Gene
ThermalTake Water 2.0 Performer

46x multiplier, 1.24v, 38* C

Run decent OC at a pretty conservative temp. With a MSI GTX 660 TF/OC running stock settings, max settings except medium shadows @ 1080p, I'm getting consistent 50-80fps.

The fps for this machine has really helped me to excel as a speed scout and brawler. Even with ECM, I'm able to be c-bill and kill productive with my founders jenner.

[Catamount]

TheFlayedman, on 15 January 2013 - 09:08 AM, said:

At 1080p which is the resolution most gamers use.

Perhaps a link would help?

That still doesn't tell enough about how a 3570k is "bottlenecking" (neither the exact hardware, usage, nor even the definition of 'bottlenecking' being used).

When paired with my 5850 at 1920x1200 and a mix of medium and high settings, I get 60fps most of the time, and there's no evidence the weak point is the CPU when I'm not (say, when I drop down to 50fps instead). The fact that OCing to 4.3 or 4.4 ghz doesn't make a discernible difference, even benchmarking, would suggest it's not the 3570k.

If I were using a 690, and was unhappy because the CPU wasn't churning out more than 90fps, then it might be different.

[Flapdrol]

Overclocking is absolutely worth it for MWO.

Higher cpu speed will mean a higher minimum fps. In a big brawl the fps drops enormously and the only thing that helps is a faster cpu.

My minimum fps in a big upclose brawl is about 35 with a core i5-750 on 4.2 Ghz. If you run a sandy bridge quad or maybe dualcore at stock frequency you should have something similar. If you then overclock that sandy or ivy 40% you'll have a minimum fps of about 50, a very noticable difference.



This is a typical game for me, game starts with ~120 fps, then gradually gets worse when the mechs get close to eachother, when mechs die fps goes up agian. Notice the gpu load gets lower as the fps does, cpu doesn't cut it.

[pantiehampster]

Oh let me play!

Xeon E5649 @ 3.8ghz (19x200), 1.3v core.
Also have 1.399v on the QPI which is dangerously high to keep it stable.
Ram is at 1200mhz, 9-9-9-25 timings, which is dissapointing because theyre Corsair Dominator GT's and should run 2000mhz @ 9-10-9-27, but i've never ever gotten them stable above 1600mhz, even with cas 10 latency and the cpu at stock clocks.

Edited by pepto biscuit, 15 January 2013 - 10:43 AM.

[Az0r]

Catamount, on 15 January 2013 - 09:21 AM, said:

Perhaps a link would help?

That still doesn't tell enough about how a 3570k is "bottlenecking" (neither the exact hardware, usage, nor even the definition of 'bottlenecking' being used).

When paired with my 5850 at 1920x1200 and a mix of medium and high settings, I get 60fps most of the time, and there's no evidence the weak point is the CPU when I'm not (say, when I drop down to 50fps instead). The fact that OCing to 4.3 or 4.4 ghz doesn't make a discernible difference, even benchmarking, would suggest it's not the 3570k.

If I were using a 690, and was unhappy because the CPU wasn't churning out more than 90fps, then it might be different.

Going from your 5850 to a 7970 wont increase your FPS any. This indicates a CPU bottleneck. That's the reason that people running current gen AMD chips at 4.0ghz with 670s or 7950s or better cards are still seeing 40-50fps. Also going from stock to 4.3 or 4.4 will make a marked difference in your minimum fps (a stock 3570k is powerful enough to produce 60fps out of combat)

[Aznpersuasion89]

Az0r, on 15 January 2013 - 10:44 AM, said:

Going from your 5850 to a 7970 wont increase your FPS any. This indicates a CPU bottleneck. That's the reason that people running current gen AMD chips at 4.0ghz with 670s or 7950s or better cards are still seeing 40-50fps. Also going from stock to 4.3 or 4.4 will make a marked difference in your minimum fps (a stock 3570k is powerful enough to produce 60fps out of combat)

im running amd 3.6ghz and a 7850 and seeing those frame rates. silly FX chips

Edited by Aznpersuasion89, 15 January 2013 - 11:24 AM.

[Catamount]

The above two posts would seem to indicate two very contradictory things. I guess I'll have to just go and test the effects of different clockspeeds myself.

[Oderint dum Metuant]

This is the problem it differs machine to machine.

Like i have shown before running a 2500k@4.5ghz and the GTX 670 FTW 4GB i can get 30FPS + over Three 1080P screens, but it does fluctuate.

On a single 1080P screen, the game never ever dips lower than 57FPS.

The problem with the russian site benchmarks mentioned above, and the links are in stickies here, was that for the CPU testing, they used a GTX 690, which was going to be crippled by no SLI, so it was effectively running a down clocked 680.



The second part that makes these results off, is the top end of the chart, both sandybridge chips perform better at stock clocks than the Ivybridge I7, knowing that Ivybridge has at least 10% better IPC that's just wrong.

Aside from 10 year old Core 2 Duos, the people with the biggest problem running this game, are people using Bulldozer based chips.

Edited by DV McKenna, 15 January 2013 - 02:45 PM.

[Cid F]

i7 3770K@4.2GHz

This high enough for 24/7 crunching WCG and gaming.

[Flapdrol]

DV McKenna, on 15 January 2013 - 02:44 PM, said:

On a single 1080P screen, the game never ever dips lower than 57FPS.

The second part that makes these results off, is the top end of the chart, both sandybridge chips perform better at stock clocks than the Ivybridge I7, knowing that Ivybridge has at least 10% better IPC that's just wrong.

Do you have a screenshot of msi afterburners graph? if you set polling every 5 seconds you should be able to fit 5 games in a row or something, if you make the afterburner window bigger. I'd like to know how the game performs on different systems.

on the screenshot of the russian site there's no ivy brigde cpu by the way, the 3xxx chip is a 6 core sandy bridge E, lower clockspeed means the sandy quads outperform it.

Edited by Flapdrol, 15 January 2013 - 03:38 PM.

[Youngblood]

Alright, my new motherboard is more or less in place with its current OC. I got RAM and the board itself to clock really well, 2800 MHz northbridge with 1.3v CPUNB, and 2600 MHz HT, 8-8-8-24-1T 1600 MHz RAM. However, for some reason the CPU doesn't want to cooperate with stability at 4.0 GHz. Have it at 3.9 for now. This is all with multipliers only, by the way.

AMD Phenom II X4 980 BE
Gigabyte GA-78LMT-USB3 (has a VRM heatsink)
Samsung MV-3V4G3D/US 8GB (2x4GB) DDR3 1600

[darkfall13]

AMD Phenom II X4 965BE OC 4.07GHz @ 1.424V Core (49C watercooled)
HIS Radeon 5870HD OC @ 1.00GHz Core 1270MHz Memory

24hrs Prime95 stable, 24hrs Memtest86+ stable, 12hrs FurMark stable
At least for me the CPU OC helped me gain ~10 fps
Recently dropped in another 5870 for Crossfire; don't do it if you want to see fps in MWO...

[Tabrias07]

i5-2500k at 4.4 GHz with Coolermaster Hyper 212+
XFX DoubleD HD6870

Game runs at a constant 40 fps in 1080p at Very High. Probably going to buy another 6870 and start using crossfire now that RadeonPro has pretty much fixed microstuttering. I'll need a better power supply too though...

Edited by Tabrias07, 17 January 2013 - 09:50 AM.

[Flapdrol]

might want to hold that off, no crossfire support yet.

[Youngblood]

Just wanted to post an update on my own machine. Dropped my RAM to stock timings to ease up on my CPU's IMC, allowing a 4.0 GHz overclock on 2.8 Ghz NB clock.

AMD Phenom II X4 980 BE
Gigabyte GA-78LMT-USB3
Samsung MV-3V4G3D/US 8GB (2x4GB) DDR3 1600 11-11-11-28-2T

[Bennz]

@ DV McKenna

the i7 39xx is not an ivy bridge, its an Sandy Bridge E.