BANANA Benchmark Results Comparison

3950X with PBO enabled:


That would put me at #2 on the leaderboard, not bad!

 

Oh, ignore what I wrote, I'm clearly blind and can't tell 5 from 3
I'm curious as your 1 car result is around 60Mbeam/s however other 5000 series Ryzen CPU's seem to have lot higher, even up to 98Mbeam/s for 1 car, is this because of lower clock speed for higher number of cores or is there something else that affects that?

Also I saw one another with around 70Mbeams/s, but several with 80 to 98Mbeam/s which is something I would except as single core performance of these is really excellent.

Here is update to my result, i7-8086k, 4.5Ghz minimum turbo instead of stock 4.3Ghz (under test it is constant 4.5Ghz, cpu is very poor at going to higher turbo speeds), maximum CPU temp I have seen is 60C as I have minor -50mV undervolting.
Asrock Taichi x370 MB, had to replace Asus as two Asus MB's has gone dead, one Gigabyte, weak VRMs.
32GB of Kingston HyperX, running XMP 3000Mhz with 15T latency as that works better for me than 3200 with 16T.
Still running Windows 8.1, next up would be hopefully Linux.

Also tampered with power limit 4 in ThrottleStop 9.3 to get rid of constant EDP Other throttling that has been cutting clock speeds.


348.402 Mbeams/s, I did several tests and it is constantly there, +/- 2 Mbeams/s.

 

Damn. How much did you get it for? Should start Folding@Home or something now

 

interesting to see that my 3900x is really close to your 3950x wich means upgrading to it wont do that much for me sadly

 

I'm pretty certain my score is bugged, no way a 3950X scores better than a 5950X.

 

A friend of mine has a 5900x and it scored higher than the 5950x too so i guess it underperformed?

 

R9 3900X, this is probably my best score so far. I had 350 and 398 a while ago.

 

My stock 5950x is only hitting ~310. Wonder what is going on with that. Other cpu benchmarks are all where they should be.

Edit: Looks to be something with the Windows 11 scheduler changes. Booting to a Windows 10 install gave me ~560 with no OC or PBO. So just a heads up to people thinking of updating.

 

Ryzen 7 1700 @ stock speed (3.0GHz @ 3.7 boost)
12GB DDR4 2400 MHz

idek if this is any good.

 

Some additional hardware tested.

Ryzen 9 5900X stock clocks, haven't tweaked anything, only checked that there is no overclocking of any kind
Asrock Taichi x570 motherboard
Ripjaws V 3600Mhz CL16 DDR4
Windows 8.1, yes installs and works fine, everything works.

4.55Ghz at beginning and 4.4Ghz at end


650.846Mbeams/s
1 vehicle 94.054Mbeams/s

 

Undersized

 

Wow either they did something with the bench or that 5900x goes like hell if thats the performance gain i get when id upgrade id do it

 

Windows 8.1, I think it is that, because if you look post of camrox92, you can see how Windows 11 gave him really low results.

Also funniest thing is that when I have power saver profile active in windows power options, I get much higher scores than when I have high performance selected, so I'm not sure how reliable benchmark Bananabench is with these CPUs, if you look results that people are getting they are all over the place.

But still, 5900X is insanely huge improvement over earlier gen models as IPC is really that 20% better, at least and there is so much where single core performance is needed.

I'm still looking for GPU, would like to go rtx3080 because triple screens, but not with current prices, so have been testing beam with 1050ti and 5900X, which is really sensible thing to do. Even with low graphics GPU is maxed out, you can drop as many cinderblock walls to top of car as you want though, there is no lag.

Well worth it at least over 8th gen i7, it is like having this game finally unlocked as there is not really need to have more vehicles running what this can run, it becomes too many vehicles and you are just grid locking all the time.

I'm thinking of testing w10 and seeing how banana bench score compares in that.

 

I currently have the 3900x with a 3070 but i sometimes have the feeling that the game gets bottlenecked alot by my 3900X dont really know why though.

 

3900X and i7-8086K are pretty same in 1 core performance and on some situations I faced limits of what 1 core can do.

One thing you should know though is that 5900X runs pretty hot, with one of the better 280mm AIO, I'm seeing 80C regularly and idling is 40C, it does not throttle or anything, but warms up liquid quite well.
I was supposed to build this with Corsair iCUE 150 Pro XT that is 3 fan AIO, but it is still in backlog, so though to see if 280mm would cut it and sure, but it does get noisy after a while.

Win 10 testing, so I did same test as in https://beamng.com/threads/banana-benchmark-results-comparison.8595/page-65#post-1343501 clock speeds were same, but results were different.

High performance power plan
1 vehicle 82.429 Mbeams/s
Max 504.149 Mbeams/s

Power saver power plan (don't ask me how this works, usually clock speed in power saver is very reduced, but in banana bench clocks get up just fine)
1 vehicle 81.953 Mbeams/s
Max 508.871 Mbeams/s

So that is quite far from results I get from Win 8.1, also there is less fps drops in Win 8.1 while playing, it's like if Win 10 SDD operations would halt everything else when they get really heavy, super annoying, but it is what it is.

Also it seems to be that in Win 10 there is not a huge difference between power plans, although sometimes high performance plan produces lot smaller numbers, but also in Win 10 at the beginning clock speeds does not go so high as in Win 8.1 from some reason, not sure where the differences really come from, it would be quite lot to say that Win 8.1 would be 10% or so faster in Bananabench, but maybe more like to how time is counted or then those small clock speed differences.

Anyway if you check my i7-8086K results, you can see score is somewhat high for the CPU and only explanation I find would be Windows 8.1 instead of Win 10, so maybe there is something there, maybe people with more understanding of how Bananabench works can figure out if it is different OS giving more calculations per second or difference in how OS counts.

 

The higher temps on the new 5000 ryzen are quite normal my friend hast it at around 90 degrees with a 280mm AiO i would let mine run on air though since i have quite a beefy bequiet Cpu cooler wich can tame my 3900x quite nice highest i got was 76 degrees while i had 100 percent usage and a hot room around 30

 

My 5900x runs at 72c on 100% load. Using a bequiet dark rock 4

 

My guess is that air isnt actually that bad and the dark rock 4 is also huge wich also helps alot