Ofpmark - test your rig with ofp!

[toadlife 3]

EDIT: This thread is dead - please check out this thread for the new version ofpmark!

Ok. We allways see Gaming systems and video cards reviewed with the same old programs - Unreal Tournament, Quake (*puke*), Comanche. 3Dmark 200*, ect. I thought it would be cool to have benchmark that uses a real game - OFP!

So I createed OFPMark. OFPMark is a mission, which contains 4 system tests (cutscenes). During each tests, my patented, trademarked, top secret, classified, copyrighted, uber benchmarking scripts caluclate how well the system performs and barfs out a score. At the end the results are tallied, averaged, and multiplied by 100 (to give a more rounded score), and *kaboom*, you've got your 'OFPMark'.

You can download OFPMark at my site here:

http://toadlife.net/ofp/downloads/missions/ofpmark_1.zip

Place the (included in the zip file) flashpoint.cfg in your Operation Flashpoint

directory before lauching the game and running the test. Without it, your OFPMark score is not valid. Also, be sure to turn Anti-Aliasing and Anisotropic *OFF* in your graphics control panel before running the test.

Here is my OFPMark score:

Test System:

------------------------------------

AthlonXP2800+ (@2.2Ghz - 177/12.5)

Gainward Geforce4 ti4800SE (@300/600)

Asus A7V333

512MB DDR 2700

Score:

-----------

Test1: 22.4

Test2: 21.3

Test3: 13.9

Test4: 9.6

OFPMark: 1681.26

NOTEIf your game wont run with the included flashpoint.cfg file, open up your flashpoint preferences and set every setting to the highest it can go to.

[killswitch 19]

Athlon XP1900+ (1.6 GHz, 12x133)

Asus V7700 Deluxe (GeForce 2 GTS, 32 MB @220/348, driver 45.23)

Abit KR7A-RAID (KT266A), 512 MB DDR 2100

POS BSLive Value

First run: OFP 1.92b with -nomap (for no good reason but to lower the score even more due to constant disk access)

Test 1: 13.7

Test 2:   9.1

Test 3:   9.0

Test 4:   4.6

OFPMark: 910.65

2nd run: 1.92b without -nomap and preloading of Nogova via the mission editor, then running ofpmark:

Test 1:  14.2

Test 2:  16.7

Test 3:    8.9

Test 4:    5.0

OFPMark: 1118.57

[Speeeedy 0]

Nice really impressive tool. The benchmark mission it self aint really spectacular but the score thing is a real nice idea. Maybe more spectacular scenes/story in the next version?

test system:

-----------

AMD64 XP3200+

Radeon 9800 Pro

512 DDR-RAM PC3200

Asus k8V Deluxe

my score:

1771

[lee_h._oswald 0]

Very nice idea.

Make a big project out of it and send to all hardware "test" sites on the net.

Send it twice to those quake ]|[ 640+480(16bit) 500fps testers!

---

---

AthlonXP 2500+@3000+ (via multiplicator)

2x 512MB DDR333 Samsung

GF4 Ti4600 detonator 44.03

Shuttle XPC SN45G

---

MfG Lee *piss*

[toadlife 3]

Send it twice to those quake ]|[ 640+480(16bit) 500fps testers!

LOL. No shit. I'll of course be trying to improve it. Right now the results can vary by too much (my ofp mark has been anywhere from 1630-1740). I need to make the cutscnenes a bit more controlled, so the marks are more consistent.

BTW..that's a nice OFPMark you have!

Quote[/b] ]2nd run: 1.92b without -nomap and preloading of Nogova via the mission editor, then running ofpmark:

A similar effect can be obtained by loading the mission, and then hitting "restart". I havnt tried it in 1.92 yet.

[R. Gerschwarzenge 0]

Test System:

------------------------------------

AthlonXP2600+ (2088Mhz)

Chaintech GeforceFX (@343/539)

Epox 8RDA3G

512MB DDR 2700

Score:

-----------

Test1: 20.8727

Test2: 32.2581

Test3: 15.4506

Test4: 7.38871

OFPMark: 1899.25

[mr burns 133]

AthlonXP 2500+@1850 (FSB 166)

2x 512MB DDR333 Infineon

Power Color Rad 9600 Pro ez

Asus A7N8X

------------

Test #1 Toadlifes Config and settings (No AA & AF)

Test #2 My Config and Settings (Hisky, GMR, Inq, 4xAA & 4xAF)

-----------------

Strange..isn´t it ? Also the AH-1 in the dynamic Test Engaged the large Tankgroup in the 1st test and flew anouther route without engaging in the 2nd

But Anyway, great thing u made there Toadlife...very cool !

[theavonlady 2]

Hmmmmmmmmm.

Dell Lattitude C840 P4 2Ghz 512MB RAM, Geforce 440 Go 64MB onboard.

Test 1: 10.6034

Test 2: 12.5

Test 3: 7.85887

Test 4: 5.59649

OFP Mark: 913.97

Checked to make certain Anti-Aliasing and Anisotropic were off in Geforce controls.

Graphics were awful when using this Flashpoint.cfg file. Never seen it so bad.

Ran OFP straight from the 1.92 EXE. No mod folders, etc. No other programs were running during the test.

edit: Just for the record, restored my original Flashpoint.cfg file and ran OFPMark. I would up even worse with a score of 649.109.

[bonko the sane 2]

Nice one ToadLife

After seeing Lee's results methinks I need a new CPU...

AMD XP1700+ ; KT4V board ; 512 DDR Ram@266 GF5600FXUltra running 45.33Driver

Quote[/b] ]Graphics were awful when using this Flashpoint.cfg file. Never seen it so bad.

Yup,it looked terrible with no AA and Anisotropic,and the spinning test made me feel sick

[theavonlady 2]

Quote[/b] ]Graphics were awful when using this Flashpoint.cfg file. Never seen it so bad.

Yup,it looked terrible with no AA and Anisotropic,and the spinning test made me feel sick  

Yes but I normally have these disabled yet during the first test with Toadlife's cfg file, I didn't get the quality I normally have (which was restored when using my own cfg file in my 2nd test).

I assume it has to do with the OFP preference settings being lower in Toadlife's cfg than in mine, where these are pretty much maxed out.

[kegetys 2]

Test One   - 28,9913

Test Two   - 32,2581

Test Three - 23,5751

Test Four  - 13,6116

ofpmark 2460,9

with:

Intel Pentium 4C 2,4 at 3,1GHz

GF4 Ti4200 at 300/650MHz

2x512MB at 416MHz DDR (Dual channel)

Windows 2000 pro

[Taconic 0]

I pulled a 1365, but that isn't entirely accurate.

OFP seems to have a particularly memory leak on multiproc systems. I've tested it on two and both fall victim to it, yet none of the uniproc systems that I've tried it on in the exact same configuration have the problem. It's kind of difficult to maintain an acceptible framerate when the system is paging out like there's no tomorrow because Flashpointresistance.exe has decided to gobble up 411MB of memory. There are certain missions that I can't even play because of this. The paging is so bad and constant that it takes well over a minute to tab out and kill the process.

Anyway, specs:

Dual 2.66GHz Xeon

Supermicro X5DAE

512MB PC2100

GF4 Ti 4400

[Skulleye 0]

I got nothing more than a black screen running OFP with your Flashpoint.cfg

System:

AMD Athlon 1000 Mhz

Geforce 4 Ti 4600 128 MB

256 MB SDRAM

Windows XP

I use 1.92 beta.

I noticed your cfg file made the textures be 4096x4096 while I usually use 256x256 on the important textures and 128x128 on the others. And then the game runs decent if not too much is happening at one time

[theavonlady 2]

I got nothing more than a black screen running OFP with your Flashpoint.cfg

Did you see Toadlife's last note:

Quote[/b] ]NOTE If your game wont run with the included flashpoint.cfg file, open up your flashpoint preferences and set every setting to the highest it can go to.

[Dwarden 1125]

26.1105

31.25

21.4283

13.3097

2302.something

not bad with 4x AA, 8xAF and 128x960x32

not to mention it were used GMR, INQ, Blood, Hisky, others Retextured addons (e.g. silesian's and island ones)...

sadly 3rd test will crash at High terrain geometry (too big view distance for that man) ...

btw that 360 degrees turn is very affected by game loading objects, so 1st 90 degrees is slideshow (no idea why) and rest is smooth ...

what about do like 3x360 degree turning

[ruff 102]

test1- 18.9118

test2- 19.2308

test3-13.2456

test4-8.63061

mark

1500.47

specs

amd 2100(1.7ghz)

gf4 ti 4200

512 ddr

current agp setting 2

is that n e good for my comp?

[toadlife 3]

Hey I thought all of my settings were maxxed out. I think maybe I slid the LOD bars the wrong way. Should LOD setting be at 1.0 or 0.05

[theavonlady 2]

Hey I thought all of my settings were maxxed out. I think maybe I slid the LOD bars the wrong way. Should LOD setting be at 1.0 or 0.05

0.05

That's a finer/higher level of detail.

[toadlife 3]

OK thats how I had it.

[MarcusJClifford 1]

Pentium 4 2.8C

[MarcusJClifford 1]

While I commend you for making this (and your other excellnt contributions to the OFP community) I would like to point out a slight issue I encountered.

It is to do with scene 3, I ran the benchmark twice and got 2 very different scenes - differences I encoutered were.

1. The AH-1s use of its gun (cannon) - none in run 1 lots in run 2

2. A BMP surviving and engaging with the AT3 (?) and its main gun (Sabot) - it being destroyed and exploding right infront of the camera [almost full screen] (didnt happen in Run 1, did in 2)

3. More small arms fire in the 2nd run (is that infanty down the hill?)

I know there is very little that can be done to prevent this as it depends on the AIs action, I am just pointing out the fact.

Still a very useful "program" to bench OFP - better than the number you are given anyway - mine varies from 5000 -> 6500.

[toadlife 3]

It is to do with scene 3, I ran the benchmark twice and got 2 very different scenes - differences I encoutered were.

1. The AH-1s use of its gun (cannon) - none in run 1 lots in run 2

2. A BMP surviving and engaging with the AT3 (?) and its main gun (Sabot) - it being destroyed and exploding right infront of the camera [almost full screen] (didnt happen in Run 1, did in 2)

3. More small arms fire in the 2nd run (is that infanty down the hill?)

I know there is very little that can be done to prevent this as it depends on the AIs action, I am just pointing out the fact.

Still a very useful "program" to bench OFP - better than the number you are given anyway - mine varies from 5000 -> 6500.

You're right. The third test's results can vary wildly depending of what the AI does. The second test can also vary, but not quite as much as the third, because there is no fighting.

The first a fourth test are very reliable.

I'm going to completely redo the third tests and rework the first and second just a bit. I also have some ideas for a fifth test, which will be simialr to the fourth 'spin' test, and be very reliable.

[bonko the sane 2]

One thing I did notice in your flashpoint.cfg was the line:<table border="0" align="center" width="95%" cellpadding="0" cellspacing="0"><tr><td>Code Sample </td></tr><tr><td id="CODE">sky=1 What is that may I ask? 'Cos I never had that in mine...

[toadlife 3]

I have no idea.

[toadlife 3]

I pulled a 1365, but that isn't entirely accurate.

OFP seems to have a particularly memory leak on multiproc systems. I've tested it on two and both fall victim to it, yet none of the uniproc systems that I've tried it on in the exact same configuration have the problem. It's kind of difficult to maintain an acceptible framerate when the system is paging out like there's no tomorrow because Flashpointresistance.exe has decided to gobble up 411MB of memory. There are certain missions that I can't even play because of this. The paging is so bad and constant that it takes well over a minute to tab out and kill the process.

Anyway, specs:

Dual 2.66GHz Xeon

Supermicro X5DAE

512MB PC2100

GF4 Ti 4400

Have you tried setting the processor affinity (making the game run on only one processor) for OFP after launching the game?