Take the BFTB performance test!

[MarkShot]

Right you are.

[cdbeck]

Ok here are mine:

Intel Core 2 Duo T7250 2.0 Ghz
4 GB Ram
nVidia 8600M 256MB
Windows 7

B250 = 83
B600 = 24

Analysis:
B250 ran at a little over a minute a second, and would probably be pretty workable. The B600 scenario might take a while to run, however. I re-ran the tests and turned the AA Map option off, increased and decreased zoom, changed the priority of the game to Highest, and put my Anti-Virus (Norton's Internet Security 10) in Sleep mode - all with negligible effect (a change from 83 to 85 was the most I saw). I also turned off the "Show surrendered/lost units" options - also negligible effect.

Sigh... if only I had the coin for an upgrade from this laptop to a proper desktop with a newer CPU.   

Strangely, PirateJock and I have similar systems - his CPU is a little faster - our B250 numbers are the same and the increase is in the B600 number. How odd.

[cdbeck]

Just for fun, I ran the tests on my native resolution of 1440x900.

The effect was also negligible - in fact I improved slightly on the B600 test, although dropping in the B250 test.

At 1440x900:
B250 = 78
B600 = 25

The variance in the test makes me wonder if simulation speed is also contingent upon hits/misses (it may slow down when the "explosions" and casualties are processed) different movement orders, or other "hidden" factors.

SoM

[MarkShot]

It is pretty clear that unit count is one of the key drivers of performance and that clock speed is the other key factor.

Granted there may be yet other factors (some of them already noted here). However, by everyone using the same methodology to test and mainly looking at relative relationships as opposed absolutes, I think this provides reasonable estimations of what people can expect.

My next project will be to relate demo performance to retail performance. STAY TUNED

[MarkShot]

See first post with red highlight "MORE NEW STUFF".

Included a means to extrapolate performance from the BFTB demo to the BFTB retail game.

[Toby42]

So you can load BFTB120 into the Demo game and run it?

[MarkShot]

No, it will not be possible to use retail files with the demo or demo files with retail.

However, the provided demo scenario will be the St. Vith tutorial. That scenario clocks in at 120 total units when all the reinforcements have arrived.

So, I created a BFTB120 (retail) save game benchmark.

If you follow the methodology I have documented with the demo, you should be able to create your own BFTB120 (demo) figure. Then, based on the ratios I published here, you should be able to infer what type of performance you are going to see with the retail game on your particular machine for BFTB250 and BFTB600.

I hope that is clear.

[Toby42]

Ok, So I run the demo to the amount of time that you posted (3days 10hrs.) Then run it for 1 minute of real time and see how much simulated time runs. Then match it against your numbers?

Am I close?

[MarkShot]

You then take your demo result and multiply it by the factors I have provided in order to get an idea of how the retail game will run for you. (The other thing which you personally can do is use the HTTR250 benchmark to infer what would be the BFTB250 and BFTB600 results.) We are just talking about basic proportions equations which they covered in school when were kids:

BFTB250 (average scenario size) = [your BFTB120 derived from the demo] * 60%

BFTB600 (largest scenario size) = [your BFTB120 derived from the demo] * 17%

So far, I am the only one to post BFTB120 results. Thus, you would have to go with my ratios. But other posts in this thread have seemed to indicate fairly consistent ratios. For example, BFTB250 has been around 4X of BFTB600.

[Toby42]

ORIGINAL: MarkShot

You then take your demo result and multiply it by the factors I have provided in order to get an idea of how the retail game will run for you. (The other thing which you personally can do is use the HTTR250 benchmark to infer what would be the BFTB250 and BFTB600 results.) We are just talking about basic proportions equations which they covered in school when were kids:

BFTB250 (average scenario size) = [your BFTB120 derived from the demo] * 60%

BFTB600 (largest scenario size) = [your BFTB120 derived from the demo] * 17%

So far, I am the only one to post BFTB120 results. Thus, you would have to go with my ratios. But other posts in this thread have seemed to indicate fairly consistent ratios. For example, BFTB250 has been around 4X of BFTB600.

When the Demo becomes available, I'll run it and post my numbers!

[NewBobolix]

ORIGINAL: MarkShot

See first post with red highlight "MORE NEW STUFF".

Included a means to extrapolate performance from the BFTB demo to the BFTB retail game.

Mark I need your opinion:
-I'll be out for the summer and will be able to play ONLY on an old notebook
-I'll buy a new notebook but not before september (will be a desktop replacement)
-I did the COTA250 test and scored a poor 22 at ">>> Speed"

Will I be able to play the tutorial scenario even several times to learn all the mechanics of the game (as far as I understood should be a "120" scenario) while reading the manuals during the summer ?
I bought COTA in a wrong moment of my life and never really played with also due to the location of the scenarios..I know The Bulge will be veeery different for me...and I have much more time to invest now

[boogada]

NewBobolix: I scored a 26 on the Cota250 test and the game runs just fine! At least the smaller scenarios, including the tutorial. So go ahead. 

[MarkShot]

NewBobolix,

First, I just want to point out the whole reason for these benchmarks was to give people an opportunity to evaluate what is reasonable performance for themselves. I remember when I used to play flight sims. One guy could look at 30 FPS and call it "smooth as silk" and another would say that he couldn't "shoot for sh*t" because his plane was jerky and uncontrollable.

COTA is a good choice to get up to speed for BFTB. The major engine concepts are all there. I did a major portion of my COTA beta work on a 2004 DELL Inspiron 5150 Laptop (Intel Pentium 4 HT Mobile 2.8GHz). I really had a good time with it, but personally avoided the largest scenarios, because it was too slow for my tastes.

Take a look a this table. I've sorted it by number of units. You will see that there are at least 15 or so scenarios to practice on with a lot less units than the COTA250 reference.

And keep in mind that COTA will run in background when minimized. So, if you are playing larger scenarios and feeling impatient especially during the slower night periods, you can set a RUN UNTIL for 2-3 hours and go off and surf the Web.

Enjoy the Summer!

[MarkShot]

One other thing to keep mind is that it seems that the performance of the game engine does not scale linearly. So, you are going to see a bigger jump in performance than just 2X if you cut the number of units in half.

[NewBobolix]

thank you Mark...just a little small thing you did not realize...I will buy immediately BfTB and start directly from there as you just told me that my old notebook will do the job

[NewBobolix]

from your table I can see I will be able to play some scenarios under the 250 unit as well in BFTB [:D]

[MarkShot]

Post removed because it was confusing. I even confused myself ... which is dangerous.

[RayWolfe]

ORIGINAL: MarkShot
I even confused myself

Easily done! [:'(]

[NewBobolix]

ORIGINAL: MarkShot

Post removed because it was confusing. I even confused myself ... which is dangerous.

You really confused me !!!
In the middle of my confusion I finished pressing the BUY button and put 16 random numbers found on a piece of plastic that was quietly stitting in my pocket...
As a result I'm now at 6% of a download of something big [8|]
[8|][8|]

[Larac]

BFTB120: 338
BFTB250: 106
BFTB600: 35

I ran at 1024x768 and default 1920 x 1080 no real speed difference.

Also no Allocation of CPUs or memory just let it run

AMD Phenom II X3 2,80 GHz
3 GB RAM
NVIDIA 9800 GTX
WIN 7 64 Bit

More RAM may help I have another gig to add, will repost with that info just to see.


Lee