New 3DMark 11 Video Tantalizes With DX11 Splendor
Futuremark hasn't told anyone when it plans to launch 3DMark 11, but the video and screenshots the company has released thus far point towards a gorgeous application. It's arguably timely; the three-year-old 3DMark Vantage doesn't support DirectX 11. This could probably be fixed via an update, the same way 3DMark 2005 was rebuilt as 3DMark 2006, but Futuremark apparently has enough changes in mind to justify a new program.
These tidings are virtually guaranteed to kick off a fresh round of debate over whether or not 3DMark is a "real" test, or whether it deserves to be considered a benchmark at all. In addition to providing news of the actual video, we've decided to tackle this question. It's an ironic argument—Futuremark was actually founded by several former members of Future Crew. Our fellow dinosaurs might recognize that name—from 1992 - 1994, Future Crew earned recognition as one of the best demo-building* organizations around. Demos from this time period could be considered the ancestors of modern benchmarks. These programs were often coded in Assembler and could be highly tuned for the processors they ran on. In the days before 3D video cards, a well-designed demo could serve as a valid test of a CPU's performance when executing different types of code.
In the two shots above, Tessellation is on the in the first, off in the second. To see the difference, compare the ridges of carving on the totem--they stand out more noticeably on the first.
Fast forward to the modern era, and the argument chiefly focuses on whether or not a benchmark that isn't based on a shipping game can be a real or useful test. Our answer to this question is yes, though it helps to understand exactly what's being tested. On the simplest level, performance in 3DMark does correlate to performance in <Insert Game Of Your Choice Here.> Because it's a well-known test that both AMD and NVIDIA examine with a fine-toothed comb, it's a bellwhether of what we should expect from other games. When games drastically deviate from 3DMark's relative performance ranking (either positively or negatively), the application provides an additional point of comparison.
This doesn't make 3DMark the single-most important 3D test, but it does offer a relatively unchanged series of ranking that can be referred to year after year. Because it's designed to run as independently from the CPU as possible, it's potentially a better performance indicator than another game that puts a heavier load on the CPU.
*The term "demo" means something slightly different in this context. Demos aren't restricted (but free) examples of programs or games—they're projects that a group of coders have built to demonstrate their programming skills. Future Crew became famous for Unreal (1992) and Second Reality (1993). An early 3D benchmark, Final Reality, was designed by members of Future Crew who formed Remedy Entertainment.
These tidings are virtually guaranteed to kick off a fresh round of debate over whether or not 3DMark is a "real" test, or whether it deserves to be considered a benchmark at all. In addition to providing news of the actual video, we've decided to tackle this question. It's an ironic argument—Futuremark was actually founded by several former members of Future Crew. Our fellow dinosaurs might recognize that name—from 1992 - 1994, Future Crew earned recognition as one of the best demo-building* organizations around. Demos from this time period could be considered the ancestors of modern benchmarks. These programs were often coded in Assembler and could be highly tuned for the processors they ran on. In the days before 3D video cards, a well-designed demo could serve as a valid test of a CPU's performance when executing different types of code.
In the two shots above, Tessellation is on the in the first, off in the second. To see the difference, compare the ridges of carving on the totem--they stand out more noticeably on the first.
Fast forward to the modern era, and the argument chiefly focuses on whether or not a benchmark that isn't based on a shipping game can be a real or useful test. Our answer to this question is yes, though it helps to understand exactly what's being tested. On the simplest level, performance in 3DMark does correlate to performance in <Insert Game Of Your Choice Here.> Because it's a well-known test that both AMD and NVIDIA examine with a fine-toothed comb, it's a bellwhether of what we should expect from other games. When games drastically deviate from 3DMark's relative performance ranking (either positively or negatively), the application provides an additional point of comparison.
*The term "demo" means something slightly different in this context. Demos aren't restricted (but free) examples of programs or games—they're projects that a group of coders have built to demonstrate their programming skills. Future Crew became famous for Unreal (1992) and Second Reality (1993). An early 3D benchmark, Final Reality, was designed by members of Future Crew who formed Remedy Entertainment.
