Star Swarm Benchmark

[NeverBackDown]

someone who understands what Mantle really is! Good job news guru

 

[WYP]

someone who understands what Mantle really is! Good job news guru

Thank you sir, there are so many misconceptions out there on mantle ATM.

Hopefully everything about mantle will become clearer as more people/tech sites benchmark and games are released.

I will try to bench mantle with BF4 and Star swarm when i am next able and publish my results, hopefully i will have time to test with my cpu overclocked and stock in order to see how mantle performs (how much the framerates drop in mantle and dx11 when the cpu is reduced to stock clocks or maybe even lower).

I'm still curious though what Star citizen is doing supporting Mantle and Physx, are they leveraging the cpu benifits of mantle for amd gpus and placing some cpu load on nvidia's physx for nvidia gpus. They could even be trying this APU co-processing with HSA (discussed in my last post) as a substitute for physx (seems more unlikely to me but you never know with SC).

 

[barnsley]

I'm still curious though what Star citizen is doing supporting Mantle and Physx, are they leveraging the cpu benifits of mantle for amd gpus and placing some cpu load on nvidia's physx for nvidia gpus. They could even be trying this APU co-processing with HSA (discussed in my last post) as a substitute for physx (seems more unlikely to me but you never know with SC).

Star Citizen is an excellent marketing tool for pc companies partly as RSI have taken the "this game will max your pc, but in an acceptable way" line. Whoever has the best gpus/cpus combo for the game will win. RSI have been good guys (in my opinion) by not siding with either side. Its very likely they will use HSA at the rate they are going, I'm sure both the green and the red team are giving them access to everything in a bid to out do eachother. Its exciting to say the least ^_^

 

[NeverBackDown]

I think they are too many variables right now to consider anything for SC. Mantle is flexible and can allow many options to be done. Maybe have the cpu only process AI/Physics and everything else on the GPU. It would free up some processing power and allow more accurate physics and better AI.

in the end its up to the dev to make a choice on what they feel needs to be done and get it done right. Games are different so some games need other things that others need less of and etc. Which is why Mantle will help a lot because of its flexible nature.

 

[SieB]

Mantle does not offload cpu tasks to the gpu, it takes just takes away the driver overhead off the CPU and enables the cpu to talk to the gpu more openly, allowing a boat load more draw calls and allows the cpu to essentially do more with less.

Star swarm is a draw call heavy scenario with all the misc calculations (pathfinding, AI etc) and the power required to feed all the gpu tasks is simply too much for most cpus to handle.

though there is a possibility of using an APU (or rather the gpu in an APU) as a co-processor for physics or other tasks. It is described here on

I guess the only way to use this test as any indication of relative performance will be to use a large number of sequential runs. Can't wait for catalyst 14.1 to be released and for myself to get back to belfast so i may benchmark

I thought it was basically the same thing as offloading onto the GPU.
The performance hit on the CPU is now lessened because it communicates better with the GPU, by taking what would usually all be done on the CPU and increasing what the GPU does with it. Essentially offloading some of the load that was on the CPU onto the GPU.

That was my take on it anyway.

As for the benchmark the RTS and Shump benches seem to be the same each time all the other seem dynamic.

 

[barnsley]

I think they are too many variables right now to consider anything for SC. Mantle is flexible and can allow many options to be done. Maybe have the cpu only process AI/Physics and everything else on the GPU. It would free up some processing power and allow more accurate physics and better AI.

They announced a partnership with Nvidia before they announced one with AMD. I presume that they'll do something that does that for both sides, be it using phyx for the Nvidia lot or mantle for the AMD lot. It is hard to tell

 

[NeverBackDown]

They announced a partnership with Nvidia before they announced one with AMD. I presume that they'll do something that does that for both sides, be it using phyx for the Nvidia lot or mantle for the AMD lot. It is hard to tell

Really? I did not know that.
Which reinforces my point that there are too many variables.. Looks like we need to wait and see!

With the amount of money backing them up though then in all honesty they should go as far as out and include every tech available.

 

[barnsley]

With the amount of money backing them up though then in all honesty they should go as far as out and include every tech available.

Oh they are it seems :lol:

-edit-
I SHOULD TAKE THE OPPORTUNITY TO REMIND YOU ALL THAT there is an OC3D star citizen thread. We also have a casual organisation in game.

 

[SPS]

someone who understands what Mantle really is! Good job news guru

TBH I don't think anyone on here really understands what it is!

But then again why would they? There's no documentation in the public domain yet.

 

[Zoot]

TBH I don't think anyone on here really understands what it is!

But then again why would they? There's no documentation in the public domain yet.

Unless there's any AMD or EA/Dice employee's hiding here?

I reckon I'm probably the closest you'll get to an AMD/Intel employee here.

 

[SeekaX]

Unless there's any AMD or EA/Dice employee's hiding here?

I reckon I'm probably the closest you'll get to an AMD/Intel employee here.

pretty sure we had some actual intel employees around here

 

[Mysterae]

They announced a partnership with Nvidia before they announced one with AMD. I presume that they'll do something that does that for both sides, be it using phyx for the Nvidia lot or mantle for the AMD lot. It is hard to tell

If you look at many of their videos you'll see rigs with that familiar green geforce glow. I think Chris Roberts' own rig has a double glow .

Not that it matters much at this stage, I'm sure they won't forget the red side, that would be foolish. As for optimising or what will run better, we'll have to wait.

 

[Pogal_31]

Here is mine guys, Custom Scenario and Extreme settings, will post with Mantle as well when it 'll be available for us

===========================================================
Oxide Games
Star Swarm Benchmark - ©2013
C:\Users\Julien\Documents\Star Swarm\Benchmark_14_01_31_2148.txt
Version 0.95
01/31/2014 21:48
===========================================================

== Hardware Configuration =================================
GPU: AMD Radeon R9 200 Series
CPU: AuthenticAMD
AMD Phenom(tm) II X4 965 Processor
Physical Cores: 4
Logical Cores: 4
Physical Memory: 8580177920
Allocatable Memory: 8796092891136
===========================================================


== 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: 16
Shading Samples: 64
Shade Quality: Mid
Motion Blur Frame Time: 16
Motion Blur InterFrame Time: 2
Detailed Frame Info: Off
===========================================================


== Results ================================================
Test Duration: 120 Seconds
Total Frames: 2411

Average FPS: 20.08
Average Unit Count: 3719
Maximum Unit Count: 5454
Average Batches/MS: 604.41
Maximum Batches/MS: 849.55
Average Batch Count: 31282
Maximum Batch Count: 205609
===========================================================

 

[Zoot]

This is what I get.

===========================================================
Oxide Games
Star Swarm Stress Test - ©2013
C:\Users\Mark\Documents\Star Swarm\Output_14_01_31_2338.txt
Version 1.00
01/31/2014 23:38
===========================================================

== Hardware Configuration =================================
GPU: NVIDIA GeForce GTX 780
CPU: AuthenticAMD
AMD FX(tm)-8350 Eight-Core Processor
Physical Cores: 4
Logical Cores: 8
Physical Memory: 17160171520
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: 10058

Average FPS: 27.93
Average Unit Count: 4084
Maximum Unit Count: 5505
Average Batches/MS: 535.84
Maximum Batches/MS: 1225.45
Average Batch Count: 19290
Maximum Batch Count: 134062
===========================================================

Based on others scores in the thread, I did better than expected.

It's been a very long time indeed since I've seen something drop to single digit FPS. ^_^

 

[BigDaddyKong]

3770K stock settings, 7970ghz stock settings, 8gb Corsair 2133mhz memory.
Crossfire not working.

===========================================================
Oxide Games
Star Swarm Stress Test - ©2013
C:\Users\james\Documents\Star Swarm\Output_14_01_31_2054.txt
Version 1.00
01/31/2014 20:54
===========================================================

== Hardware Configuration =================================
GPU: AMD Radeon HD 7900 Series
CPU: GenuineIntel
Intel(R) Core(TM) i7-3770K CPU @ 3.50GHz
Physical Cores: 4
Logical Cores: 8
Physical Memory: 8530481152
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: 10768

Average FPS: 29.91
Average Unit Count: 4050
Maximum Unit Count: 5606
Average Batches/MS: 530.76
Maximum Batches/MS: 1394.93
Average Batch Count: 20184
Maximum Batch Count: 159676
===========================================================



Second Run
Average FPS: 35.83

Third Run
Average FPS: 33.46

Fourth Run
Average FPS: 46.16

Fifth Run
Average FPS: 28.30

Xfire Enabled without Profile
Run 1
Average FPS: 26.76

Run 2
Average FPS: 29.90

Some interesting observations during these runs. When single card was working, set for Xfire with profile, GPU usage during the heavy scenes was at 32% with CPU usage at 20%. That is the scenes with a kajillion lasers and ships and other stuff. During lighter action scenes, I got 100% GPU usage, and about 24% CPU usage.

When I enabled Xfire without profile, my GPU usage was 54% on both cards, and 25% on the CPU usage on light action scenes. When it came to the heavy stuff, I got 20% on both GPU's and 21% on the CPU.

I think this test is custom made for Mantle with enough light source draw calls, that it looks like its stressing your system, its just bottle necking the draw calls.

Everything monitored with Aida 64 on G19.

 

[cooperman]

Thay updated there test since i tested so heres a DX rerun.

===========================================================
Oxide Games
Star Swarm Stress Test - ©2013
C:\Users\coop\Documents\Star Swarm\Output_14_02_01_1057.txt
Version 1.00
02/01/2014 10:57
===========================================================

== Hardware Configuration =================================
GPU: AMD Radeon HD 7900 Series
CPU: AuthenticAMD
AMD FX(tm)-8320 Eight-Core Processor
Physical Cores: 4
Logical Cores: 8
Physical Memory: 8502214656
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: 9936

Average FPS: 27.60
Average Unit Count: 4062
Maximum Unit Count: 5732
Average Batches/MS: 341.58
Maximum Batches/MS: 827.20
Average Batch Count: 14876
Maximum Batch Count: 89827
===========================================================

 

[SieB]

Here's mine after the update.

===========================================================
Oxide Games
Star Swarm Stress Test - ©2013
C:\Users\Sie\Documents\Star Swarm\Output_14_02_01_1122.txt
Version 1.00
02/01/2014 11:22
===========================================================

== Hardware Configuration =================================
GPU: AMD Radeon R9 200 Series
CPU: GenuineIntel
Intel(R) Core(TM) i5-2500K CPU @ 3.30GHz
Physical Cores: 4
Logical Cores: 4
Physical Memory: 8545779712
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: 10652

Average FPS: 29.59
Average Unit Count: 4240
Maximum Unit Count: 5557
Average Batches/MS: 642.13
Maximum Batches/MS: 1357.75
Average Batch Count: 24030
Maximum Batch Count: 188498
===========================================================

 

[Edgeyforsure]

heres my score, quite impressed

== Hardware Configuration =================================
GPU: NVIDIA GeForce GTX 780
CPU: GenuineIntel
Intel(R) Core(TM) i7-3930K CPU @ 3.20GHz [actual = 4.4ghz]
Physical Cores: 6
Logical Cores: 12
Physical Memory: 17116270592
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: 12405

Average FPS: 34.45
Average Unit Count: 4208
Maximum Unit Count: 5461
Average Batches/MS: 531.11
Maximum Batches/MS: 1545.39
Average Batch Count: 16531
Maximum Batch Count: 91953
===========================================================

 

[Pogal_31]

Mine after the update too, i'm really CPU limited, my Matrix never went beyond 40% of load, seems it's time to upgrade my old Phenom

Oxide Games
Star Swarm Stress Test - ©2013
C:\Users\Julien\Documents\Star Swarm\Output_14_02_01_1534.txt
Version 1.00
02/01/2014 15:34
===========================================================

== Hardware Configuration =================================
GPU: AMD Radeon R9 200 Series
CPU: AuthenticAMD
AMD Phenom(tm) II X4 965 Processor
Physical Cores: 4
Logical Cores: 4
Physical Memory: 8580177920
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: 8253

Average FPS: 22.92
Average Unit Count: 3578
Maximum Unit Count: 5497
Average Batches/MS: 399.82
Maximum Batches/MS: 847.00
Average Batch Count: 20798
Maximum Batch Count: 98999
===========================================================

 

[barnsley]

If you look at many of their videos you'll see rigs with that familiar green geforce glow. I think Chris Roberts' own rig has a double glow .

Aye, The dev rigs are 3930ks with either a 690 (or two in the case of Chris' rig),780 or a 290x. I know there is a couple of 9590s and a few apus now as well, although I'd imagine the the apus are for development with mantle.