Cinebench is a terrible CPU benchmark. Geekbench is far better.
People do not run a single benchmark tool when comparing hardware, they use multiple. Each with their own well defined characteristics and make their own conclusions depending on what matters to them.
Cinebench is part of almost every hardware review for a good reason: its characteristics (CPU only, not dependent on cache, memory speeds nor timings) are well understood and provide valuable information on a specific aspect of the hardware.
People do not run a single benchmark tool when comparing hardware, they use multiple.
Most Youtubers base the performance of the CPU on Cinebench. They definitely aren't Anandtech, where CPU is measured with SPEC.
Cinebench is part of almost every hardware review for a good reason: its characteristics (CPU only, not dependent on cache, memory speeds nor timings) are well understood and provide valuable information on a specific aspect of the hardware.
Cinebench is being used as an end all CPU benchmark too often. Just look at the author I replied to.
Again, if one were to use one single benchmark to denote the performance of a CPU, they should use SPEC. Absent of SPEC, they should use Geekbench. Cinebench is pretty far down the list.
I didn't use it as a benchmarking tool. I don't care about the relatively numbers between CPUs or even what performance it represents. I only referenced the throttling behavior. I could as well have used a simple C program that adds floating points together as fast as possible.
Comments
People do not run a single benchmark tool when comparing hardware, they use multiple. Each with their own well defined characteristics and make their own conclusions depending on what matters to them.
Cinebench is part of almost every hardware review for a good reason: its characteristics (CPU only, not dependent on cache, memory speeds nor timings) are well understood and provide valuable information on a specific aspect of the hardware.
Most Youtubers base the performance of the CPU on Cinebench. They definitely aren't Anandtech, where CPU is measured with SPEC.
Cinebench is being used as an end all CPU benchmark too often. Just look at the author I replied to.
Again, if one were to use one single benchmark to denote the performance of a CPU, they should use SPEC. Absent of SPEC, they should use Geekbench. Cinebench is pretty far down the list.
I didn't use it as a benchmarking tool. I don't care about the relatively numbers between CPUs or even what performance it represents. I only referenced the throttling behavior. I could as well have used a simple C program that adds floating points together as fast as possible.