Basemark Announces GPUScore Benchmark

Published by

Click here to post a comment for Basemark Announces GPUScore Benchmark on our message forum
data/avatar/default/avatar25.webp
I'm looking forward to this. 3DMark has grown stale and even more meaningless than before. Am I off base? Pun intended
https://forums.guru3d.com/data/avatars/m/277/277212.jpg
I think they need a benchmark to be able to repeatably replicate the characteristics of MS Flight Simulator. Especially, the enormous amount of video memory used for both meshes and textures. This is actually somewhat easier than might be assumed if they make heavy use of procedural generation techniques. What I mean by that is they shouldn't model all of those buildings and the terrain because then you'll need a multi-DVD-sized download. They need to generate the terrain and buildings in code. The textures also need to be generated and that's easier because they can just replicate a few but they have to generate many GBs of textures just like FS uses. Terrain and buildings can be randomly generated and textured easily. That's been done for many years. They could be smart about it and let you specify how much memory it uses for textures and meshes and let you vary detail levels. I think this could be very interesting.
https://forums.guru3d.com/data/avatars/m/246/246171.jpg
Srsbsns:

I'm looking forward to this. 3DMark has grown stale and even more meaningless than before. Am I off base? Pun intended
I've never liked 3DMark. It's good in the sense that it can let you know what your hardware is capable of (as though what's printed on the box isn't enough), but is otherwise pointless. I don't like any complex synthetic benchmarks, because despite what they claim about simulating real-world workloads, they aren't real-world workloads. I am fine with basic synthetic benchmarks, such as Prime95 or FurMark, for 3 reasons: 1. They push the hardware to extremes, which is great for stability testing, and can run for as long as you need them to. 2. They're great for showing regressions in hardware technology or drivers, since their results are far more affected by the most subtle changes. 3. They're usually free, light-weight, and fast to get going. But I am not ok with those tests as a metric for what product to get.
https://forums.guru3d.com/data/avatars/m/16/16662.jpg
Administrator
Basemark is founded by a few ex Futuremark folks 🙄
https://forums.guru3d.com/data/avatars/m/269/269912.jpg
I enjoy 3DMARK. Whenever I build a new rig, or buy a new gpu, it enables me to see the base score before any overclocking and the result after overclocking. I don't use it to try and top any list or try to be the big dawg on the porch like other users do.