Jump to content

Recommended Posts

Posted (edited)

Star Swarm is a real-time demo of Oxide Games’ Nitrous engine, which pits two AI-controlled fleets against each other in a furious space battle. Originally conceived as an internal stress test, Oxide decided to release Star Swarm so that the public can share our vision of what we think the future of gaming can be. The simulation in Star Swarm shows off Nitrous’ ability to have thousands of individual units onscreen at once, each running their own physics, AI, pathfinding, and threat assessments.

 

Note:Star Swarm is not a deterministic simulation -- the AI and everything else is being computed in real time, so you will get slightly different results from multiple executions even on the same hardware. Unfortunately, achieving 100% determinism with the highly threaded nature of the Nitrous engine is an unrealistic goal.

 

http://store.steampowered.com/app/267130/

 

So, lets start. I believe this will be a useful tool to stress our components and compare our system performance. Star Swarm supports mantle so it would be interesting way to compare AMD's API to Direct X..

 

I have a mediocre gaming rig that's a couple years old.

Intel 2500k @ 4.6

8gb Corsair XMP @ 1600mhz

Asus 560ti @ 900mhz

 

Star Swarm settings

Scenario: Follow

Settings: Extreme

Stress Test: Timed Run (6 min)

 

Start D3D

 

Average FPS:   26.99

post-5079-0-60575500-1391258656_thumb.jp

Edited by Bordobel
Posted (edited)

Changed scenario to a more demanding one.

 

Star Swarm settings

Scenario: RTS

Settings: Extreme

Stress Test: Timed Run (6 min)

 

Start D3D

 

Average FPS:   9.91 :/

post-5079-0-06194600-1391260579_thumb.jp

Edited by Bordobel
Posted (edited)

Mantle beta drivers are out http://support.amd.com/en-us/kb-articles/Pages/latest-catalyst-windows-beta.aspx

 

AMD Mantle Beta driver is currently supported on:

  • AMD Radeon™ R9 Series GPUs
  • AMD Radeon™ R7 Series GPUs
  • AMD Radeon™ HD 7000 Series GPUs
  • AMD Radeon™ HD 8000 Series GPUs
  • AMD A10-7000 Series and AMD A8-7000 Series APUs
Edited by Bordobel
Posted (edited)

Mantle. The only reason I went to AMD's new 290x card.

 

Ran one with Dx and one with Mantle

 

Mantle first.

(Ran the RTS too with an average 18.68FPS)

post-4961-0-81794100-1391354390_thumb.jp

post-4961-0-55290700-1391355346_thumb.jp

Edited by Shootter
Posted (edited)

@Borobel - relative to the number of games that require more horsepower, you do not have a mediocre gaming rig. You have a better than average rig. Even with the GTX5XX card.

 

And Mantle is intended to help mediocre CPUs running games requiring Dx9, not GPUs.

 

for the Follow mode, I got:

 

Oxide Games
Star Swarm Stress Test - ©2013
C:\Users\Xerxes\Documents\Star Swarm\Output_14_02_03_1909.txt
Version 1.00
02/03/2014 19:09
===========================================================
 
== Hardware Configuration =================================
GPU: NVIDIA GeForce GTX 760
CPU: AuthenticAMD
AMD FX-8350 Eight-Core Processor           
Physical Cores: 4
Logical Cores: 8
Physical Memory: 17123692544
Allocatable Memory: 8796092891136
===========================================================
 
 
== Configuration ==========================================
API: DirectX
Scenario: ScenarioFollow.csv
User Input: Disabled
Resolution: 1920x1080
Fullscreen: True
GameCore Update: 16.6 ms
Bloom Quality: High
PointLight Quality: High
ToneCurve Quality: High
Glare Overdraw: 16
Shading Samples: 64
Shade Quality: Mid
Deferred Contexts: Disabled
Temporal AA Duration: 16
Temporal AA Time Slice: 2
Detailed Frame Info: Off
===========================================================
 
 
== Results ================================================
Test Duration: 360 Seconds
Total Frames: 9707
 
Average FPS: 26.96
Average Unit Count: 4124
Maximum Unit Count: 5559
Average Batches/MS: 477.35
Maximum Batches/MS: 1158.07
Average Batch Count: 17988
Maximum Batch Count: 134817
===========================================================
Edited by Gremlich
Posted

let us please all use the same Settings as the OP ok? ;)

 

Star Swarm settings

Scenario: Follow

Settings: Extreme

Stress Test: Timed Run (6 min)

 

System Hardware

Intel i7-3370K@3,50Ghz

Nvidia Geforce GTX 680 OC

16GB RAM

 

 

 

iLGMp4h.jpg

 

Hmm...I'm really think about to buy a new GPU... :unsure:

 

 

 

 

Posted

I just ran this myself with DX and Mantle using the new cat 14.1 beta drivers

 

Star Swarm settings

Scenario: Follow

Settings: Extreme

Stress Test: Timed Run (6 min)

 

System Hardware

Intel i7-3770K, OC @4.40Ghz

Asus Radeon HD 7970 GHZ Edition (No OC currently)

16GB RAM

 

With the new Cat 14.1 beta drivers average fps went from 40 in DX11 to 58 using Mantle. Thats quite the increase. :)

 

post-2004-0-13309000-1391381606_thumb.jp

post-2004-0-85045200-1391381610_thumb.jp

Posted

So, lets start. I believe this will be a useful tool to stress our components and compare our system performance. Star Swarm supports mantle so it would be interesting way to compare AMD's API to Direct X..

Star Swarm does not only support Mantle, it was build to show the advantages of this technology, so naturally AMD CPUs with AMD GPUs should perform better than Intel and NVidia components.

Posted (edited)

System Requirements

  • Minimum:
    • OS: Windows 7 64-bit / Windows 8 64-bit 
    • Processor: Quad-Core Processor 
    • Memory: 4 GB RAM 
    • Graphics: 1 GB VRAM (Dedicated) DirectX 11 Compatible Video Card 
    • DirectX: Version 11 
    • Hard Drive: 4 GB available space 
    • Additional Notes: Mantle Benchmark requires an AMD Radeon 7x00 Series Video Card or Better and the AMD Mantle drivers.

 

Star Swarm settings

Scenario: Follow

Settings: Extreme

Stress Test: Timed Run (6 min)

 

System Hardware

Intel i7-930, OC @4.2Ghz

Gigabyte 7970 OC Edition (No OC currently)

24GB RAM

 

 

Chase Cam:

===========================================================
Oxide Games
Star Swarm Stress Test - ©2013
C:\Users\Chris\Documents\Star Swarm\Output_14_02_03_0825.txt
Version 1.00
02/03/2014 08:25
===========================================================


== Hardware Configuration =================================
GPU: AMD Radeon HD 7970 Series
CPU: GenuineIntel
Intel(R) Core(TM) i7 CPU         930  @ 4.2GHz
Physical Cores: 4
Logical Cores: 8
Physical Memory:  25760481280
Allocatable Memory: 140737488224256
===========================================================




== Configuration ==========================================
API: Mantle
Scenario: ScenarioFollow.csv
User Input: Disabled
Resolution: 1920x1080
Fullscreen: True
GameCore Update: 16.6 ms
Bloom Quality: High
PointLight Quality: High
ToneCurve Quality: High
Glare Overdraw: 16
Shading Samples:  64
Shade Quality: Mid
Deferred Contexts: Disabled
Temporal AA Duration: 16
Temporal AA Time Slice: 2
Detailed Frame Info: Off
===========================================================




== Results ================================================
Test Duration: 360 Seconds
Total Frames: 17361


Average FPS: 48.22
Average Unit Count: 4519
Maximum Unit Count: 5811
Average Batches/MS: 941.57
Maximum Batches/MS: 2222.16
Average Batch Count: 21563
Maximum Batch Count: 110513
===========================================================
===========================================================
Oxide Games
Star Swarm Stress Test - ©2013
C:\Users\Chris\Documents\Star Swarm\Output_14_02_03_0833.txt
Version 1.00
02/03/2014 08:33
===========================================================


== Hardware Configuration =================================
GPU: AMD Radeon HD 7970 Series
CPU: GenuineIntel
Intel(R) Core(TM) i7 CPU         930  @ 4.2GHz
Physical Cores: 4
Logical Cores: 8
Physical Memory:  25760481280
Allocatable Memory: 140737488224256
===========================================================




== Configuration ==========================================
API: DirectX
Scenario: ScenarioFollow.csv
User Input: Disabled
Resolution: 1920x1080
Fullscreen: True
GameCore Update: 16.6 ms
Bloom Quality: High
PointLight Quality: High
ToneCurve Quality: High
Glare Overdraw: 16
Shading Samples:  64
Shade Quality: Mid
Deferred Contexts: Disabled
Temporal AA Duration: 16
Temporal AA Time Slice: 2
Detailed Frame Info: Off
===========================================================




== Results ================================================
Test Duration: 360 Seconds
Total Frames: 12637


Average FPS: 35.09
Average Unit Count: 4386
Maximum Unit Count: 5426
Average Batches/MS: 573.89
Maximum Batches/MS: 1308.07
Average Batch Count: 17927
Maximum Batch Count: 101680
===========================================================

 

Star Swarm settings

Scenario: RTS

Settings: Extreme

Stress Test: Timed Run (6 min)

 

System Hardware

Intel i7-930, OC @4.2Ghz

Gigabyte 7970 OC Edition (No OC currently)

24GB RAM

 

 

RTS View:

Mantle(BETA):
===========================================================
Oxide Games
Star Swarm Stress Test - ©2013
C:\Users\Chris\Documents\Star Swarm\Output_14_02_03_0330.txt
Version 1.00
02/03/2014 03:30
===========================================================


== Hardware Configuration =================================
GPU: AMD Radeon HD 7970 Series
CPU: GenuineIntel
Intel(R) Core(TM) i7 CPU         930  @ 4.2GHz
Physical Cores: 4
Logical Cores: 8
Physical Memory: 25760481280
Allocatable Memory: 140737488224256
===========================================================




== Configuration ==========================================
API: Mantle
Scenario: ScenarioRTS.csv
User Input: Disabled
Resolution: 1920x1080
Fullscreen: True
GameCore Update: 16.6 ms
Bloom Quality: High
PointLight Quality: High
ToneCurve Quality: High
Glare Overdraw: 16
Shading Samples: 64
Shade Quality: Mid
Deferred Contexts: Disabled
Temporal AA Duration: 16
Temporal AA Time Slice: 2
Detailed Frame Info: Off
===========================================================




== Results ================================================
Test Duration: 360 Seconds
Total Frames: 9112


Average FPS: 25.31
Average Unit Count: 3848
Maximum Unit Count: 5485
Average Batches/MS: 2030.82
Maximum Batches/MS: 2634.62
Average Batch Count: 81636
Maximum Batch Count: 173857
===========================================================
DirectX:
===========================================================
Oxide Games
Star Swarm Stress Test - ©2013
C:\Users\Chris\Documents\Star Swarm\Output_14_02_03_0338.txt
Version 1.00
02/03/2014 03:38
===========================================================


== Hardware Configuration =================================
GPU: AMD Radeon HD 7970 Series
CPU: GenuineIntel
Intel(R) Core(TM) i7 CPU         930  @ 4.2GHz
Physical Cores: 4
Logical Cores: 8
Physical Memory: 25760481280
Allocatable Memory: 140737488224256
===========================================================




== Configuration ==========================================
API: DirectX
Scenario: ScenarioRTS.csv
User Input: Disabled
Resolution: 1920x1080
Fullscreen: True
GameCore Update: 16.6 ms
Bloom Quality: High
PointLight Quality: High
ToneCurve Quality: High
Glare Overdraw: 16
Shading Samples: 64
Shade Quality: Mid
Deferred Contexts: Disabled
Temporal AA Duration: 16
Temporal AA Time Slice: 2
Detailed Frame Info: Off
===========================================================




== Results ================================================
Test Duration: 360 Seconds
Total Frames: 2996


Average FPS: 8.32
Average Unit Count: 3945
Maximum Unit Count: 5783
Average Batches/MS: 866.80
Maximum Batches/MS: 1167.36
Average Batch Count: 101959
Maximum Batch Count: 164895
===========================================================
Edited by Genesis
Posted

Those are some impressive performance boosts with Mantle API, especially with the mix of components Intel CPU and AMD GPU.

I was definitely happy to see a ride from 40 to 58 fps with mantle. Now to wait a few months for drivers to be official and not beta and more games to offer support.

Posted

let us please all use the same Settings as the OP ok? ;)

 

 

Hmm...I'm really think about to buy a new GPU... :unsure:

Fixed follow on mine, RTS too slow. Time to OC

Posted

Wow, if we go buy these numbers mantle is a pretty big performance increase even with high end cpu's...

 

I might have to change to AMD by the end of the year if Nvidia don't bring out a new API of their own.

 

I am not sure though you will get this kind of performance increase when playing actual games. Only in the test situation.

Posted

I am not sure though you will get this kind of performance increase when playing actual games. Only in the test situation.

 

You surely would, if the developers used Mantle. The sticking point here is if AMD is able to convince enough developers to use the new API.

Posted

Star Swarm settings

Scenario: Follow

Settings: Extreme

Stress Test: Timed Run (6 min)

 

System Hardware

i5-2500k Sandy Bridge @ 4.9 Ghz OC

8 GB RAM

MSI GTX670 2GB OC (No actual OC'ing happening)

 

Now, I watched the ENTIRE thing and noticed something a bit odd. Whenever there was a "huge" unit bit in which the FPS dropped, the actual fire-rate and speed of the lasters RARELY slowed. About 200seconds into it there was a marked slowing of the weapons fire, but before that time and after around 240seconds, there was no change. I am not sure exactly why that would have been, but I've been skeptical of this as a true stress test, especially given that it's supposed to test the CPU/GPU, but my CPU rarely went above 60% load on any of the 4 cores, whereas games like GW2 can put up to 80% sustained load on at least 2 cores.  Anyhoo, here's the results for my rig.

 

Please note: Despite the reading of "@ 3.30 Ghz", I run at 4.9 and confirmed through CPUID and Coretemp just to make sure my OC settings hadn't frakked up during a BSOD or something.

 

===========================================================

Oxide Games

Star Swarm Stress Test - ©2013

C:\Users\user\Documents\Star Swarm\Output_14_02_04_1624.txt

Version 1.00

02/04/2014 16:24

===========================================================

== Hardware Configuration =================================

GPU:        NVIDIA GeForce GTX 670

CPU:        GenuineIntel

               Intel® Core i5-2500K CPU @ 3.30GHz

Physical Cores:            4

Logical Cores:            4

Physical Memory:         8559706112

Allocatable Memory:        8796092891136

===========================================================

== Configuration ==========================================

API:                DirectX

Scenario:            ScenarioFollow.csv

User Input:            Disabled

Resolution:            1920x1080

Fullscreen:            True

GameCore Update:        16.6 ms

Bloom Quality:            High

PointLight Quality:        High

ToneCurve Quality:        High

Glare Overdraw:            16

Shading Samples:         64

Shade Quality:            Mid

Deferred Contexts:        Disabled

Temporal AA Duration:        16

Temporal AA Time Slice:        2

Detailed Frame Info:        Off

===========================================================

== Results ================================================

Test Duration:            360 Seconds

Total Frames:            14237

Average FPS:            39.54

Average Unit Count:        4186

Maximum Unit Count:        5463

Average Batches/MS:        662.63

Maximum Batches/MS:        1564.45

Average Batch Count:        17559

Maximum Batch Count:        116759

===========================================================

 

Ran RTS mode and got just under 9 FPS. Either this is designed to run better on AMD, or it really, REALLY hates Intel. The Engine doesn't look any better to me than Sins of a Solar Empire on max, yet it seems to be a lot rougher on the CPU/GPU. Still was a but surprised that it didn't push my CPU above 60% again...

Posted

let us please all use the same Settings as the OP ok? ;)

 

Star Swarm settings

Scenario: Follow

Settings: Extreme

Stress Test: Timed Run (6 min)

 

System Hardware

Intel i7-3370K@3,50Ghz

Nvidia Geforce GTX 680 OC

16GB RAM

 

 

 

iLGMp4h.jpg

 

Hmm...I'm really think about to buy a new GPU... :unsure:

 

Maybe OC your CPU? I'm running a rig with lower numbers on the RAM, and with a "lower" end card than yours, but I got over 10fps higher and the only difference is my OC'd chip.

Posted

Mantel definitely has potential based on the star swarm tests. I to would like to see real game improvement. Currently we only have battlefield 4 and the current beta driver is only optiomised for hte r290 in BF4. With my 7970 GHZ as I said earlier I only gain about 12% with settings on ultra. For a lower spec system that was struggling 12% could see huge but 12% when your average is above 80 fps is neglibible.

 

The netx version of the driver will hopefully optimise mantle in BF4 for all of the supported cards and when this happens we will hopefully see the real world game improvements we are all hoping for.

 

Incidently, is there anyone here that has a R290X along with BF4 and has tried Mantle?

Posted

@bod699 I have the 290x and ran BF4 in Mantle on Ultra. To be honest I noticed nothing special. Coming from a GTX580Ti I would be hard pressed to notice better BF4 performance to begin with.

 

On a side note. One issue that did seem to keep occurring is game crashes while running mantle, I was getting some DirectX conflictions. Forgot to screen cap the

error message but DX conflictions was the jist of it.

Posted

@bod699 I have the 290x and ran BF4 in Mantle on Ultra. To be honest I noticed nothing special. Coming from a GTX580Ti I never had any performance issues with

BF4 to begin with.

 

On a side note. One issue that did seem to keep occurring is game crashes while running mantle, I was getting some DirectX conflictions. Forgot to screen cap the

error message but DX conflictions was the jist of it.

  It's all beta.  And lets not get silly BF4 is run by EA so everything to do with them is void and should be destroyed.

 

 

 

  BTW there was an update to Star Swarm.

Posted

@ShootterSad to hear no real difference in BF4 performance with a r290x as that is the card that is supposedly optimised for mantle with the beta driver.

 

@Genesis, the reason I asked about BF4 is because its the 1 real game that currently has support. With the amount of issues still facing that game any real game seen in it surely cant be trusted atm. Shootter seeing no real difference with the 1 card they have supposedly optimised tells its own story.

 

Star Swarm on the other hand continues to show crazy improvements.. I tried the test again on RTS mode, Extreme setting with my unclocked 7970 GHZ and I7 3970k.

 

DX11 gave 9.59fps and Mantle gave 29.60fps, thtas close to a 200% gain. That was also before the latest update and going by the patch notes it states "Fixed bug in Nitrous that turned threading off in Mantle."

 

Not sure of nitrous or if this will improve my performance more.

Posted

just wait till the release of the game thief by ubisoft. it will be the first game which will give mantle support from the first release.

i think we will than see the hole potential of mantle. also we have to wait for better drivers.

 

  :golfclap:  amd

Posted

Mantle beta drivers are out http://support.amd.com/en-us/kb-articles/Pages/latest-catalyst-windows-beta.aspx

AMD Mantle Beta driver is currently supported on:

  • AMD Radeon™ R9 Series GPUs
  • AMD Radeon™ R7 Series GPUs
  • AMD Radeon™ HD 7000 Series GPUs
  • AMD Radeon™ HD 8000 Series GPUs
  • AMD A10-7000 Series and AMD A8-7000 Series APUs
Thought mantle was only the very latest cards, R9 etc, interesting.
Posted

  There are claims that the DX API was gimped on purpose.

 

Someone test this:

 

Go in to the "assets" folder in the starswarm demo folder and open up Settings_Extreme.ini and then change "deferred contexts=0" to =1

run

enjoy higher FPS

this is what I mean - star swarm HAS been coded to take advantage of deferred contexts, but it has been disabled by default and you have to hack an ini file to turn it back on

e.g. I get 35fps on my AMD machine, turn on mantle get 50fps
on nvidia machine (same CPU) I get 30fps no DC, 55fps with DC, so bigger gain percentage wise and higher average fps on NV

interestingly, BF3 favoured nvidia hardware because of DC, but DICE have removed DC from BF4 because they claim they saw LOWER performance with DC, however every other game using DC gets a big boost for nvidia hardware

Create an account or sign in to comment

You need to be a member in order to leave a comment

Create an account

Sign up for a new account in our community. It's easy!

Register a new account

Sign in

Already have an account? Sign in here.

Sign In Now
×
×
  • Create New...