• Competitor rules

    Please remember that any mention of competitors, hinting at competitors or offering to provide details of competitors will result in an account suspension. The full rules can be found under the 'Terms and Rules' link in the bottom right corner of your screen. Just don't mention competitors in any way, shape or form and you'll be OK.

Star Swarm - Oxide bench thread

My GPU's hang around 30% but I get terrible FPS at the same time! I was looking forward to this engine too.

JaseUK said:
Is Crossfire supported with Star Swarm as it doesn't seem to be working for me?

I have downloaded the latest 14.1 drivers
Click to expand...

Crossfire doesn't work for Mantle + Star Swarm right now.
 
I wouldn't say the dx crippling is 100% on purpose, oxide appear to have developed an engine/scenario that specifically targets any weak points in the api that on the other hand make mantle look good. I think poorly coded/optimised may be a better description.

Quite ironic too that after all these years of 'cpu can do physics just fine' we find our selves in a situation where a demo geared at mantle shows to get big gains, physics needs to be done on the gpu :p
 
Direct X

== Results ================================================
Test Duration: 360 Seconds
Total Frames: 16290

Average FPS: 45.25
Average Unit Count: 4313
Maximum Unit Count: 5574
Average Batches/MS: 721.00
Maximum Batches/MS: 1758.71
Average Batch Count: 17870
Maximum Batch Count: 107928
===========================================================



Mantle

== Results ================================================
Test Duration: 360 Seconds
Total Frames: 22957

Average FPS: 63.77
Average Unit Count: 4772
Maximum Unit Count: 6194
Average Batches/MS: 1297.73
Maximum Batches/MS: 3290.17
Average Batch Count: 22788
Maximum Batch Count: 148539
===========================================================

Not bad at all :)
 
It is crippling directx. They claimed in a video before release that the way they do motion blur is to use multiple renders, as opposed to a tacked on after effect you see in most games. This does cripple directx, which I am not sure why there are even complaints about. This isn't a benchmark. It's a program they use to test their engine, which they kindly released so we had something to test Mantle.

Obviously we are keen to see what the differences are in this scenario with Directx and Mantle. Does this situation effect current games? Very unlikely. Will this situation effect games in the future? Quite possibly. Unfortunately that the feature they wish to use as an addition to Mantle is motion blur, one of the most turned off features in the history of gaming settings.

I don't know what the results will be without motion blur, but they said that the motion blur is the fps killer in Directx.

Whilst they could be crippling directx elsewhere, I don't believe they are, I mean why would they intentionally hurt their engine which they are trying to ship out to other developers and to make their own games?

They said that they can add features such as motion blur with Mantle, it can be turned off for Directx versions of their games (and will probably be turned off by Mantle users too). So whilst the results are kind of artificially skewed, it's trying to raise a point that Mantle can excel where Directx falters. You won't see such giant boosts in other area's but any place where directx is starting to struggle, Mantle should be able to be handy to either avoid, or limit the issue.
 
Star Swarm crashes for me when running with Extreme settings in DirectX mode, so here's my "High" quality result.

== Hardware Configuration =================================
GPU: AMD Radeon R9 200 Series
CPU: GenuineIntel
Intel(R) Core(TM) i7-4770K CPU @ 3.50GHz
Physical Cores: 4
Logical Cores: 8
===========================================================


== Configuration ==========================================
API: DirectX
Scenario: ScenarioCustom.csv
User Input: Disabled
Resolution: 1920x1080
Fullscreen: True
GameCore Update: 16.6 ms
Bloom Quality: High
PointLight Quality: High
ToneCurve Quality: High
Glare Overdraw: 8
Shading Samples: 64
Shade Quality: Mid
Deferred Contexts: Disabled
Temporal AA Duration: 16
Temporal AA Time Slice: 3
Detailed Frame Info: Off
===========================================================


== Results ================================================
Test Duration: 360 Seconds
Total Frames: 11281

Average FPS: 31.33
Average Unit Count: 4067
Maximum Unit Count: 5532
Average Batches/MS: 793.05
Maximum Batches/MS: 1474.38
Average Batch Count: 24548
Maximum Batch Count: 63093
===========================================================

== Configuration ==========================================
API: Mantle
Scenario: ScenarioCustom.csv
User Input: Disabled
Resolution: 1920x1080
Fullscreen: True
GameCore Update: 16.6 ms
Bloom Quality: High
PointLight Quality: High
ToneCurve Quality: High
Glare Overdraw: 8
Shading Samples: 64
Shade Quality: Mid
Deferred Contexts: Disabled
Temporal AA Duration: 16
Temporal AA Time Slice: 3
Detailed Frame Info: Off
===========================================================


== Results ================================================
Test Duration: 360 Seconds
Total Frames: 12646

Average FPS: 35.13
Average Unit Count: 4395
Maximum Unit Count: 5411
Average Batches/MS: 953.01
Maximum Batches/MS: 2780.55
Average Batch Count: 26085
Maximum Batch Count: 69175
===========================================================
 
NEW BUILD AVAILABLE FOR THIS STRESS TEST!!!!!!!

So Oxide have posted they have a new build for this, apparently fixed some stuff relating to Mantle or something, anyhow just done first DX Parse, CPU at Stock 4770k, GPU at stock 290p

DX PARSE
===========================================================
Oxide Games
Star Swarm Stress Test - ©2013
C:\Users\Si\Documents\Star Swarm\Output_14_02_04_1950.txt
Version 1.00
02/04/2014 19:50
===========================================================

== Hardware Configuration =================================
GPU: AMD Radeon R9 200 Series
CPU: GenuineIntel
Intel(R) Core(TM) i7-4770K CPU @ 3.50GHz
Physical Cores: 4
Logical Cores: 8
Physical Memory: 8474652672
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: 13817

Average FPS: 38.38
Average Unit Count: 4068
Maximum Unit Count: 5434
Average Batches/MS: 664.36
Maximum Batches/MS: 1355.97
Average Batch Count: 19937
Maximum Batch Count: 146631
===========================================================

Mantle Parse
===========================================================
Oxide Games
Star Swarm Stress Test - ©2013
C:\Users\Si\Documents\Star Swarm\Output_14_02_04_1959.txt
Version 1.00
02/04/2014 19:59
===========================================================

== Hardware Configuration =================================
GPU: AMD Radeon R9 200 Series
CPU: GenuineIntel
Intel(R) Core(TM) i7-4770K CPU @ 3.50GHz
Physical Cores: 4
Logical Cores: 8
Physical Memory: 8474652672
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: 19374

Average FPS: 53.81
Average Unit Count: 4530
Maximum Unit Count: 5647
Average Batches/MS: 1055.77
Maximum Batches/MS: 3175.42
Average Batch Count: 22056
Maximum Batch Count: 86632
===========================================================

OC GPU Core run, CPU at Stock 4770k , GPU at 1100/1250 - 290p
Mantle Parse

===========================================================
Oxide Games
Star Swarm Stress Test - ©2013
C:\Users\Si\Documents\Star Swarm\Output_14_02_04_2011.txt
Version 1.00
02/04/2014 20:11
===========================================================

== Hardware Configuration =================================
GPU: AMD Radeon R9 200 Series
CPU: GenuineIntel
Intel(R) Core(TM) i7-4770K CPU @ 3.50GHz
Physical Cores: 4
Logical Cores: 8
Physical Memory: 8474652672
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: 22581

Average FPS: 62.72
Average Unit Count: 4531
Maximum Unit Count: 5810
Average Batches/MS: 1030.37
Maximum Batches/MS: 3536.16
Average Batch Count: 18780
Maximum Batch Count: 81364
===========================================================

OC GPU Core and Mem Run, 4770k at Stock, 290p at 1100/1400

===========================================================
Oxide Games
Star Swarm Stress Test - ©2013
C:\Users\Si\Documents\Star Swarm\Output_14_02_04_2019.txt
Version 1.00
02/04/2014 20:19
===========================================================

== Hardware Configuration =================================
GPU: AMD Radeon R9 200 Series
CPU: GenuineIntel
Intel(R) Core(TM) i7-4770K CPU @ 3.50GHz
Physical Cores: 4
Logical Cores: 8
Physical Memory: 8474652672
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: 22727

Average FPS: 63.13
Average Unit Count: 4531
Maximum Unit Count: 5709
Average Batches/MS: 1104.80
Maximum Batches/MS: 6193.45
Average Batch Count: 20529
Maximum Batch Count: 174270
=========================================================
 
Last edited:
So what are results like in this if you run both (DX/Mantle) with motion blur turned off?
Do they both gain the same amount or does Mantle gain less because it was impacted less?

I realise this is a tech demo, but personally I tend to turn off blur effects, so if it was used in a proper game I'm wondering how much of a difference it might make, unplayable -> playable?
 
Personally I find it hard to imagine what games I'd want to play that use the engine. Although in fairness I'm assuming it's an engine made for space based stuff (or probably water based). Guess it could also do land based, sort of RTS games? So maybe.

Of course I believe the Unreal engine was used for more than just FPS games, so I'm not sure how versatile this engine is.

EDIT:
Going back to the motion blur thing. MY other thought is that if 'we' were benching Metro 2033 (not sure if Last Light has a built in benchmark?) then it would probably be fairer to run the benchmark with PhysX disabled as it hit non-Nvidia users harder than Nvidia users. So for StarSwarm it seems fairer to disable the motion blur IF it affects DX more than it affects Mantle (or vice-versa).
 
Last edited:
humbug said:
Did this not turn out as you hoped? not updating this thread any more Gregster? :p:D
Click to expand...

I don't mind updating it, I just would prefer a consistent bench, which this isn't. One person gets more units than others, hardly seems fair now does it?

Ohhhh and just to add part of the rules for those who may have missed it ;)

Also, all entries MUST have your CPU and your GPU and be titled "Official entry" Any entry will be ignored if it doesn't follow that criteria. That is starting from 3pm Saturday 1st Feb 2014.
Click to expand...

I don't see that on any of the entries. It just makes it easier for me to keep track of what is done with the settings in the OP (thanks DamnedLife) and what is just a random run. I have no bias in any of the bench threads and love to see people pushing their hardware but simple rules need to be set.
 
Last edited:
This bench is an absolute waste of time. For all intensive purposes, whether it's intentionally crippled or not. GPU usage under DX is catastrophically unoptimised.

One of the key basis of Mantle being in many ways more efficient is that it takes weights off the CPU, if this is true then why does heavily overclocking the CPU have an adverse effect. Surely the heavy unit count under DirectX, would benefit greatly from high CPU clocks?

Personally, I want a decent, combined test that's consistent to get results off both APIs. Because using this isn't conclusive enough for any kind of score chart, even with the ini changes.

EDIT: Just seen the new build, maybe that will give me higher than 30% GPU usage lol.
 
Last edited:
The only way this test will be of any use if a) they sort the DX side of things, atleast optomise it for Dx11, apparently its Dx9 or something at the mo?

b) allow you to input a specific set of stats to benchark with, then everyone is on the same even playing field.

Until then its nothing more than a pretty stress test tech demo
 
Had a go at this earlier - ran in DX mode with average of 18fps and reason why it isn't officially - unable to run in mantle mode - keeps crashing ( prob. because the hd 7*** aren't fully optimised but I thought I;d try my luck any-ways lol)
 
Mine... DX:
Average FPS: 15.75
Average Unit Count: 3931
Maximum Unit Count: 5459
Average Batches/MS: 646.70
Maximum Batches/MS: 863.32
Average Batch Count: 42130
Maximum Batch Count: 94982

Mantle:
Average FPS: 29.77
Average Unit Count: 4158
Maximum Unit Count: 5533
Average Batches/MS: 1100.24
Maximum Batches/MS: 4630.35
Average Batch Count: 36790
Maximum Batch Count: 96631


Cant complain with a 100% increase. FX8320 @ 4Ghz and 2GB 7850 @ 1040/1250.
 
You must log in or register to reply here.
Back
Top Bottom