AMD Will Not Endorse SYSmark 2012 Benchmark Discussion

AMD today announced that it will not endorse the SYSmark 2012 Benchmark (SM2012), which is published by BAPCo (Business Applications Performance Corporation). Along with the withdrawal of support, AMD has resigned from the BAPCo organization.

Read the full content of this Article: AMD Will Not Endorse SYSmark 2012 Benchmark

Related Articles:

• First Look: AMD's New A-Series Fusion Processor Review
• AMD Begins Shipping "Llano" Accelerated Processor Unit
• AMD "Zacate" APU First Look: Hands On with the Brazos Platform
• AMD Unveils "Brazos" Platform for Notebooks and Netbooks

 

the article is interesting.

 

Can't say AMD is wrong since synthetics aren't really all that accurate at representing "real world" performance, and can be skewed to favor how different CPU handle their business, but their protest isn't going to mean much if reviewers keep using SYSMark in their reviews.

 

I think you hit the nail on the head, Jerry; AMD would not have made the announcement if their processors performed better in synth. benches.

However, they have a point; 'real-life' benchmarks, especially scripted ones that make use of actual applications e.g. Microsoft Office, are a much better indicator of "will this processor work for me" - not just a bunch of numbers that only tech gurus on these forums know how to interpret.

 

Sysmark is a lost cause. The problem is not that it doesn't use actual applications, it's a matter of how to weight stuff. Obviously, if you give applications that give GPUs more weight, AMD and Nvidia come out ahead while CPU heavy applications favor Intel.

A decade or two ago, benchmarking made sense because the applications practically everyone used were actually bottlenecked by the hardware. Today, even if you somehow made a chip that takes the relatively pitiful Llano CPU and pairs it with the relatively pitiful HD3000 graphics (i.e. the worst of both worlds from AMD and Intel), most people would not notice. Microsoft Office, web browsers, email and almost everything else that is commonly used is not bottlenecked by processing power -- it's limited by I/O and best served by an SSD. To get meaningful differences between the various hardware varieties, one has to pick a specific application or task that is done by only a small subset of all users. Thus, you can't really look at benchmarks that combine many applications as Sysmark does, but need to look at the exact ones that you will actually use.

 

I agree that synthentic benchmarks such as SYSmark aren't really indicative of real-world performance but this still seems a little immature on AMD's part. It is like taking away your ball because you are losing.

 

nah its called marketing game, see who will embrace the change and hope for the best

 

Nvidia and VIA have also dropped SYSMark.

Nvidia, AMD, and VIA quit BAPCO over SYSmark 2012 | SemiAccurate

 

so its working

 

If remember somewhere that when you change the CPU ID to Intel brand, you get an extra bonus in performance. I love Intel and all, but this kind of software is simply bad.

 

nvidia dropped first, amd second, via third. it had to do with intel having almost full control over what went into sysmark, and the fact that sysmark 2012 was going to reduce the impact of gpu performance to almost 0. simply put sysmark was always a benchmark influenced heavily by one manufacturer to show their products in possitive light, i believe even hammer lost to netburst on sysmark for instance. nvidia, amd and via have all simply decided they will not lend their name to the credibility of this benchmark any longer