3D Benchmark proposal list for 2021

[Leeghoofd]

First the user wall and some other major bugs and then (hopefully) before end of march the crap algorithm should be a just bad memory. Just hang in there...

[FireKillerGR]

Also interesting to point out that GPUPi for CPU gets 180 Globals out of 462 subs (2x cores), 151 out of 154 (1x core) and 200 out of 908 subs (4x cores).
Same way for x265 4k, 144.3 for 116 subs (1x core), 172.3 points for 348 subs (2x cores) and 193.8 points for 743 subs (4x cores).

Meanwhile for 3d, 161.6 points for 3DFSX Single out of  1963  subs (lower than the 4 times less popular gpupi dual core and two times vs quad).
Same goes for dual gpu 3dfsx, 3dtsx single etc.

Most remarkable of all would be 3d11 with 161.6 points with 3678 subs.

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

And to share actual potential solutions publicly:
a) have a sheet where actual calculations go in place to calculate the points of benchmarks before changes are being implemented.
That way you dont end up with a mega unbalanced point system that has to be adjusted again and again (or at least not that often)
b) try to limit the amount of points both 3D and 2D get.
ie. 5500 points for both (equal).
Once a 3d benchmark gains popularity (= more points) the rest get automatically calibrated so the total amount of points remains the same. Same would go for 2d. 

(ie 150 subs PR gets 130 points, 03 gets 202 etc but once PR gets popular both get 166 points)

That way, we dont have to wait for a benchmark to magically gain in popularity while also don't have to change anything in case it gets a massive point increase (due to popularity).

Edited February 15, 2021 by FireKillerGR

[Rauf]

As far as I know the points distribution will be adressed in the upcoming changes. Some benchmarks have had strange points boosts since last rev, like 03, gpupi for cpu etc. But this will surely be fixed. 

[Alex@ro]

Well while i do know and experienced myself that 3d scores are harder to get than 2d you also have to take into consideration the costs related to this.

With current cards costs it makes absolutely nonsense. I mean 2000$ is already a lot of money for a video card however with current mining situation and lack of stock we have a rtx3090 kingpin edition at 4000$ in Romania right now. Getting a random one does not guarantee you anything to say at least. 

 

 

Since 3dcards have no categories like cpu core count you get no points for a mainstream card for example.

 

On the other side it is funny how supported guys with top high-end RTX3090 unavailable to the plebs are making case for higher 3d points.   Yes, random unbinned samples, average Joe can get the same freqs, blablabla. 

 

Point is , everyone is looking after their asses and assets. In my opinion given the current situation and price, the point distribution between 2d and 3d is quite fair. 

[Rauf]

I suggested something like low end, mid end and high end categories for gpu benchmarks many times during PJ's rule. But there was never any interest in implementing it. Maybe something for the future?

Evc2 should open up many non xoc cards for playing with voltage and/or power limits.

On the other hand. Prices are what they are. Same for cpus, many are priced crazy high, and cpus require way more binning than gpus. If the cost is very high, and reward points wise is low, there is little incentive to bench it. I'm not saying 3d should be worth more points than 2d. It's impossible to decide which bench is harder or require the most effort. I think all benchmarks should be valued the same and awarded the same points. But the overall number of rankings you can compete in for 3d are way lower than 2d, so it would be nice to see a few more 3d benchmarks for points. Or introduce low, mid and high end categories

[Leeghoofd]

Stuff me and Roman talked about yesterday:

- maybe max global point cap for 3D 175 while 2D we cap at 150 (Roman is not very fond of this one)

- reduce global point gap between 1st, 2nd and 3rd like 10-15 points max, now its over 50 points
   easily for 3D scores, while 2D it is in general 20 points.

- maybe we should add a bit more competition points for contests with eg lower end cheaper 950-
  1060 cards or such (Level 2-3) Same for 2D too ofcourse

-there will be more points in the upcoming point system for 4th and so on spot

[v5000a]

Have you guys considered adding a scoring system or a category for 3dmark graphics scores? I find it pretty frustrating that to be competitive in 3dmark benchmarks you need to have an expensive cpu. Ggpupi, heaven and superposition are pretty much the only 3d benchmarks I (and probably many others) can take part in because of how much the cpu benchmark influences scores.

(This is a repost, because the friday rollback seems to have removed the post I made before.)

[yosarianilives]

On 4/27/2021 at 6:53 PM, Joaquin10 said:

As , far away its the same 4 x core that virtual / Disable on Bios , cheat not ?

If HWBOT not nothing with UL Benchmark because 3D Links are confirmaed ¿

Always LOD , and AMD What happens ? x64 Cores uff Intel Go out of scene.

I think something was lost in translation?