Performance in Synthetic Benchmarks
The Radeon HD 4870 CrossFire platform is not far better than the HD 4850 CrossFire in this test and did not manage to score 14,000 points. Interestingly, the Radeon HD 4870 X2 is worse than the CrossFire tandem here.
There is no difference between the graphics solutions in the SM2.0 tests despite the numerous textures and light sources. The CPU’s performance is obviously a limiting factor. In the more difficult SM3.0/HDR tests the Radeon HD 4870 tandem scores over 8,000 points, improving upon the results of the Radeon HD 4850 CrossFire.
It is the first time that an ATI solution takes first place in the first SM2.0 test. The gap is small, yet indicative of ATI’s ability to compete and beat Nvidia on the latter’s home turf. The second test depends on the graphics subsystem’s computing capabilities where the Radeon HD 4870 CrossFire knows no rival. It easily wins this test.
It is not quite clear why, but the tested CrossFire configuration boasts a performance growth of 30% relative to the Radeon HD 4850 tandem and 100% relative to the single Radeon HD 4870. In other words, CrossFire technology is at its maximum theoretical efficiency here! This impressive result is partly due to the high memory bandwidth of the 4870 CrossFire subsystem (over 200GBps) as well as to its 1600 ALUs and 80 texture processors clocked at 750MHz.
There is no notable advantage of the Radeon HD 4870 CrossFire over the Radeon HD 4850 CrossFire in this test. Fast memory doesn’t constitute an advantage here. The small difference in the results is a surprise because the former configuration has a 125MHz higher GPU frequency.