Matrix Games Forums

Thanks!

Please see below. Interesting results with a pretty unremarkable CPU- low average pulse time, but long total time and pretty significant longest pulse that is similar to sittingDuck.

I have the lowest average pulse time on the sheet with a pretty unimpressive CPU. My highest pulse was 19671 and the total time was 5:35 with an average of 1932.3

Something changed significantly in .12 that completely reset the benchmarks. Skip the pulse times and check the time compression ratio. Note that no one gets about 1:1 now and it hangs around .3-.5

i9-13980HX, 4090 laptop, 64GB RAM. Also tested on an i9-14900HX, 4090 laptop, 64GB RAM. Same results. Both Win11.

Still something mucked up with the benchmark scenario. Time acceleration is running at .4 to .5. And pulse time are all over the place.


Still seeing pausing in larger scenarios and I think its reflected in the benchmark scenario.

So using my frustration with performance on my test/benchmark scenario posted in the below link as my motivation...

https://www.matrixgames.com/forums/view ... 4#p5136054

I decided to just spend a night completely rebuilding it. Went back almost two-years to find a version that ran as I expected. Re-added everything added over the last two years and triple-checked it. The performance on .14 goes from about 7:1 time acceleration @ double-flame using the above scenario to 82:1 @ double-flame using the rebuilt scenario on the same laptop. Yes, thats a 13X improvement. I can only think something got corrupted somewhere down the line.

But it still begs the question about the standard benchmarking scenario weird results.

About to upgrade CPU (and mobo and RAM) so thought this a good opportunity to run the benchmark.

Before the upgrade:

5800X
Highest: 9198
Lowest:219
Average:3433

Post upgrade to 7800X3D:

5700x3d

Consider to buy either i9 14900K or an AMD Ryzen 9 7950X3D. I actually think the Intel is better for this game. (I'm playing a lot hugh scenarios), But the AMD seems pretty good too. So AMD has much less current draw....

Rain08 wrote: Fri Apr 07, 2023 8:34 pm I have to say, it really feels like the only missing thing here are AMD's X3D CPUs. So far, the 12th and 13th gen Intel CPUs are topping the results here. I really wonder if CMO would benefit from the additional cache or is still dependent on just the usual CPU attributes like the frequency/IPC/core count.

Yes, that would be very interesting to know...

There was no one Ryzen Threadripper yet, albeit plenty X3D..

My over 4 years old 3900X is not so bad....

DmitriyBlade wrote: Mon Nov 07, 2022 9:30 am 1) Download scenario https://drive.google.com/file/d/1BtugYF ... share_link
2) Start(Restart) game fresh
3) Press benchmark(Left top corner)
4) Choose scenario, Iterations 1, Fidelity Very Coarse(5-sec pulse)
5) Press start
6) Wait until start button turn green, around 30 seconds after finsished
7) Post result here, also add result to google sheets
https://docs.google.com/spreadsheets/d/ ... sp=sharing


Test runs quickly, around 2-5 minutes
Would love to see 5950x, 5800X3D, 7950x,12900K, 13900K users here

I have Build 1328.18

Where do I place the downloaded scenario?

Where is the Benchmark tab?

TIA

You may extract where you want, you will browse to that location. A good place is this folder:


To access the benchmark.


Note that if your computer is 7 yo it will require time.

I know because mine is older

Yep 7yo ROG laptop
also tried performance mode and times not much better

New laptop in Balanced mode

And in the best performance mode?


There is an option in Command to switch automatically to high performance mode.

Well, in fact for some reason there are two