There was a performance regression with v0.10 (due to changing the C++ compiler). It should be largely fixed with the 0.10 hotfix (v0.10.1).
BeamNG.drive version 0.10.1 CPU: i5 6600k @ 4.5GHz RAM: 16GB 3100MHz CL15 SAME PC: 0. 9.0 - 149 Mbeams/s 0.10.0 - 126.5 Mbeams/s 0.10.1 - 153 Mbeams/s 0.9 vs 0.10.1 ~3% performance uplift 0.10.0 vs 0.10.1 ~21% performance uplift Spoiler: Let's compare it with early 2016 result Thanks to user @yolo-phil and his report from 25th February 2016 https://www.beamng.com/threads/banana-benchmark-results-comparison.8595/page-32#post-326272 First of all we have different clock speeds 4.1 vs 4.5, so let's normalize the score, the difference is ~9.75% in clockspeed (4.5 x 100 / 4.1=109.756...) His score was: 69 Mbeams/s + 9.75% (for CPU clock speed difference) ~ 75.5 So in February of 2016 same CPU with same clockspeed as mine would got around 75.5 Mbeams/s, while now it scores 153 Mbems/s, that give us uplift of over 102% of performance! That's a real milestone, and now as RYZEN multicores CPUs are getting more popular because of very good "perf per dollar" in year or two, average BeamNG player could be able to handle like ~20 cars on CPU with stable playable framerate - that's just wow, nice that desktop CPU technology is actually moving forward thaks to RYZEN. But for such a result we would need also similar graphics optimizations, which atm is the biggest performance limiting factor.
Is there any other people who have Ryzen with Asus Prime B350-Plus motherboard ? I've been running Bananabenches with it for 6 months now and it still does that strange coil while, that I confirmed coming from the motherboard itself, GPU is quiet. So I would need to get some opinions about it, if other people don't hear any coil while, I might have a dud motherboard, that causes unstability issues. I made some overclocks today and the Bananabench wasn't stable at all and it managed to fail the system and the coil whine has gone even louder over the time.
My PC: MSI GTX 1060 6GT OC (+150MHz on memory & core by myself) Intel i5-6400 2,7Ghz 4C/4T (stock clocks) 8GB DDR4 RAM MSI B150M PRO-VDH mobo 128GB ADATA SSD 1TB WesternDigital Blue HDD
The optimizations have been insane lately! I benchmarked my laptop which has a 2.4GHz i3, and it gets surprisingly close to the 2016 result with only 23.4% difference:
ASUS H110MA/M.2 I3-6100 @3.7GHz RX 460 2Gb @ 1.27GHz & 7GHz 8Gb DDR4 2133MHz 128Gb SanDisk SSD (Has BeamNG.drive) 1Tb 7200RPM Toshiba HDD BeamNG.drive V0.10.0.1 63.893 Mbeams/s 7 Realtime / 8 Realtime Major drop offs at 3 and 5
Hello there, here is banana from BeamNG version 0.10 hotfix. Ryzen 1800X not overclocked with 2800 MHz RAM (32GB), Asus prime x370 pro board.
Nice, those couple extra cores gives around 40 JBeams/s over R5 1600, which gets around 180 JBeams/s with 3.8 OC, and 2933 16GB RAM.
Also nice to see how much improvement was made in one year: 1 year ago (April 2016) Now: (September 2017)
weird lower score then my 1700 at 3.7ghz edit: rerun as that was a few versions ago I just got 222.186 MB/s
Probably not because its loading beam itself without any data it cant load it so it wouldnt work but you could try and use a laptop where you havent installed beam and then try to run it would be interresting to know
Wanted to know how my Windows 10 china tablet performs in beamng. I am quit impressed with the results here, I suppose if my GPU could handle the game it would've run smoothly with up to 4 (simple) vehicles . Tablet is CHUWI Hi13 btw.
I have a gigabyte GA-AB350-Gaming-3 I get coil whine as well its pitch get lower throughout the test.
Okay, so the test itself does something, that causes it. I got a confirmation for that some weeks ago, when I accidentally run into someone who had BeamNG and Prime B350 Plus board. He had the same coil whine on the BananaBench as I did, so definitely not a hardware issue.
First post (of mine) in this thread in a good long time, so here is my results with a stock clocked 6950x I'll say it takes the task fairly well