Monster Hunter Official Benchmark Download

Spets

2015-01-09 09:53

It runs yes but uses PhysX which is another layer ontop which is CUDA, accelerated on nVidia GPU's only, none nVidia GPU's fallsback to CPU accelerated which is 20x slower. Even if you have a nVidia GPU, e.g i have a GTX470 but can't use it unless i hack the drivers and modify the game system file checks it may work i have seen some things on it but it's a lot of hassle to try to get it to run.

There's no layers, it uses DirectCompute which runs on the GPU. AMD usually does poorly until they have a driver profile available, don't think this scenario is any different.

#4990051

Ryu5uzaku

2015-01-09 12:28

The fur effect places and whatnot run perfectly fine the slowdowns are in really weird places with nothing happening on my 290x but avg fps on 1080p SMAA 63 fps and avg on 720p 4xMSAA 64 fps both scored around 14k. So I would still play this even if the performance is random cause monster hunter is awesome. So it ain't the gameworks effects doing the dirty work here at all it is just not optimized on amd hardware simple as that.

#4990129

GeniusPr0

2015-01-09 15:51

Maybe drivers are an issue. I'm using 344.80.

#4990132

-Tj-

2015-01-09 15:55

Well idk where this min fps happens, but when I run in ranked 720p I never see minimum 50fps drop, could be at very begining when it starts benchmarking. Also there are 4-5 spots where cpu starts doing some crazy stuff and heats a little - I think its calculating with avx2 commands by changing scenes, guess those that dont have super duper avx2 it can cause extra pauses/stutters/fps drops there.. E.g. right before it shows that furry cat, later right before those smaller raptors attack that big dino and a little later when one bigger dragon spies from the hill on top looking down, and one by chasing scene.

#4990343

UZ7

2015-01-09 21:56

I made a listing in the benchmark thread if you guys want to add yourselves in! http://forums.guru3d.com/showthread.php?t=396046 nVidia: https://docs.google.com/spreadsheets/d/1d0EVh60ICP45moG_4AZ8MC6DNQT2GRzCFmcLm0yzT-g/edit?usp=sharing AMD: https://docs.google.com/spreadsheets/d/1-WqmKKj8g9vM1ngwgg0FwpSpvGvBC6qpf059MkfJcsk/edit?usp=sharing No competition, just a listing to see what kind of hardware gets what, all setups are welcome :P

#4990346

ruiner13

2015-01-09 22:00

Here's my cards running their usual OC (see AB pic). Not a bad score in competition mode. First run forgot to turn off frame limiter at 120fps! Score: 31467 Min 50.1, Avg 143.7. Noticed that there was a point where my GPU Bus was at 45% usage, that's quite high and I'm running PCIe x16 3.0. Sorry for folks only running slower busses! http://s13.postimg.org/mignyvulv/mh2.jpg http://s13.postimg.org/hxuhjyawj/gpu_oc.jpg

#4990348

Ryu5uzaku

2015-01-09 22:02

There are framerate drops on all hardware. The only one with 60fps minimum at 1280x720 is using 3 980s,which is only 120fps higher than my minimum framerate.

But drop to 13 is rather high on 3770k @ 4.8ghz.

#4990383

ruiner13

2015-01-09 23:08

Those are around the same scores I got, some of the 980 sli high oc scores are pretty impressive.

Doesn't seem that much higher than my 2x780s running at 1175-1200.

#4990394

GeniusPr0

2015-01-09 23:39

Doesn't seem that much higher than my 2x780s running at 1175-1200.

It won't be. CryEngine3.

#4990398

lmimmfn

2015-01-09 23:46

On my oldie rig: http://s27.postimg.org/4zoagewkz/mho_benchmark_2015_01_09_22_35_48_438.jpg

#4990401

BLEH!

2015-01-09 23:50

This is blatantly overoptimised for nVidia cards... i7-980X @ 4GHz 7970 Xfire @ 1050/1500. 12029, Average 54.7, Min 12.4.

#4990414

Wault

2015-01-10 00:32

Hello, Here is my contributionfor the 720p bench. FX8350@4,675GHz with a Sapphire 290 Tri-X GPU@957/1250MHz => 58.4/12.3 => 12786 GPU@1094/1414MHz => 63.2/13.9 => 13901 If the game is not optimized for AMD before it's launch, it will be very disappointing, but overall it was a rather enjoyable bench. 🙂 I was very found of the fur in the GameCube game StarFox Adventures and I think the airs of the New Dawn démo were fabulous, but I didn't like the cat's fur in this bench. It felt... not natural at all. :frown:

#4990439

-Tj-

2015-01-10 01:43

edit.. Min fps happens in last scene by that giant scorpion, one particular angle has this 51fps drop for a split second, by hairworks cat its 56fps 🤓 That scorpion camera scene looked gpu bound - compute perf. http://s4.postimg.org/8e4jadajx/mho_benchmark_2015_01_10_01_49_16_659.jpg upload image online

#4990457

GeniusPr0

2015-01-10 02:17

And by that you mean that Maxwell architecture doesnt like cryengine 3, or?

Means Kepler isn't far behind Maxwell like in DA:I and The Crew. In Crysis 3 they are literally a frame per sec apart on average. GTX 980 Vs. 780 Ti that is.

#4990715

Mineria

2015-01-10 19:02

Nothing overclocked http://i.imgur.com/F5VNfxg.jpg http://i.imgur.com/on0JUan.jpg

#4990819

GeniusPr0

2015-01-10 23:48

Its not like they are far appart in DAI either http://gamegpu.ru/rpg/rollevye/dragon-age-inkvizitsiya-test-gpu.html But sure, they are even closer in this benchmark 🙂

Not sure what's going with Kepler, but in the higher resolutions the difference starts to become even more clear. 37 vs. 33 at 1440P. That Russian site always seems to post the best performance of Kepler though, which is really strange. Anyway, wow, it loads up 10+ threads on the CPU.

#4990826

GeniusPr0

2015-01-11 00:04

37 vs 34* But yes, it does look like Maxwell scales better at higher resolutions - i suspect that it has something to do with the compression technique Maxwell took into use, to overcome the 256 bit bus bottleneck. And yeah, frostbite is one of the best engines when it comes to multithreading, along with cryengine 🙂

Hilbert reported 33 :P

#4990830

-Tj-

2015-01-11 00:21

37 vs 34* But yes, it does look like Maxwell scales better at higher resolutions - i suspect that it has something to do with the compression technique Maxwell took into use, to overcome the 256 bit bus bottleneck. And yeah, frostbite is one of the best engines when it comes to multithreading, along with cryengine 🙂

Actually its because of more ROPS and higher pixel fillrate, pixel fillrate is very resolution depended. 50gpixel vs 80gpixel will help for sure, especially in cryengine 🤓 I remember playing Crysis1 with each new gpu that had at least 80% higher pixel fillrate, texel fillrate didnt matter much after +50gtexel/s at least at lower resolution..

#4991137

PinguX

2015-01-11 20:18

CPU 4GHz GPU 1000/1250MHz [IMG]http://i61.tinypic.com/ve6hiu.jpg[/IMG]

#4991217

Dazz

2015-01-11 22:46

Ok done another test, the frame rates at WFHD & 4xFSAA average 41fps, min 12.6 (give or take .2fps) next test i did was at 640x480 & No FSAA and got 98fps with a min of 10.4fps Tell me how come when running at a significantly lower resolution did my minimum frame rates drop by 20%!!!!!!!!!!!!!!!!!!!!!!!!!!!!! I tell you why cause the GPU bottleneck has been completely removed and pushed onto the CPU which in turn is more taxing when it has to calculate psychics on top that's why there is a big drop! Now nvidia physx can run in software (CPU with any GPU) or hardware mode (GPU only on nVidia GPU) Look in the game directory and you can see the nvidia physx files right here.

I made a listing in the benchmark thread if you guys want to add yourselves in! http://forums.guru3d.com/showthread.php?t=396046 nVidia: https://docs.google.com/spreadsheets/d/1d0EVh60ICP45moG_4AZ8MC6DNQT2GRzCFmcLm0yzT-g/edit?usp=sharing AMD: https://docs.google.com/spreadsheets/d/1-WqmKKj8g9vM1ngwgg0FwpSpvGvBC6qpf059MkfJcsk/edit?usp=sharing No competition, just a listing to see what kind of hardware gets what, all setups are welcome :P

Interesting looking at them, that nVidia GPU's gain a massive increase when FSAA has been turned off in some cases 50%, and AMD's gets next to zero increase in min fps. Think that completely answers my Question. I have been real time monitoring it and the frame rates drops with the water effects, CPU usage maxes out on all cores.