Skip to main content

AMD Won't Endorse SYSmark 2012 Benchmark

Tuesday AMD made it officially well-known that it has no plans to endorse BAPCo's upcoming SYSmark 2012 benchmark. In fact, America's #2 processor giant has resigned from the BAPCo organization altogether, and even hinted that SYSmark 2012 doesn't provide "clear and reliable measurements" and is "misleading." AMD's Nigel Dessau followed up with a little clarification on the company's stance Tuesday in his blog.

"The heart of our complaint is this: the SYSmark benchmark is not only comprised of unrepresentative workloads (workloads that ignore the importance of heterogeneous computing and, frankly, favor our competitor’s designs), but it actually generates misleading results that can lead to very poor purchasing decisions, causing governments worldwide to historically overspend somewhere in the area of approximately $8B!" he said.

Dessau went on to explain that AMD tried to work within the BAPCo consortium in order to get the next-generation benchmark, SYSmark 2012, correct. AMD even stressed that it needed to be open, transparent and processor-neutral. Unfortunately, that didn't happen.

"Our good intentions were met with an outcome that we believe does a disservice to the industry and our customers," he said. "We weren’t able to effect positive change within BAPCo, and the resulting benchmark continues to distort workload performance and offers even less transparency to end users."

That said, AMD pulled itself out of BAPCo and asked that the AMD brand be removed from marketing materials promoting SYSmark 2012. According to the company's official press release, it will only endorse benchmarks based on real-world computing models and software applications, and which provide useful and relevant information.

"AMD believes benchmarks should be constructed to provide unbiased results and be transparent to customers making decisions based on those results," AMD said in an official statement. "Currently, AMD is evaluating other benchmarking alternatives, including encouraging the creation of an industry consortium to establish an open benchmark to measure overall system performance."

"We are committed to working with likeminded companies that want to give consumers and business users an accurate, honest measure of what they can expect from their PCs and mobile devices," Dessau added.

  • fleeb
    AMD, nVidia, and VIA already left BAPCo. What does that say?
    Reply
  • noob2222
    fleebAMD, nVidia, and VIA already left BAPCo. What does that say?sysmark is still based on dual core testing, ie ... crap
    Reply
  • erhardm
    fleebAMD, nVidia, and VIA already left BAPCo. What does that say?Intel CPUs will be the world's champions from now on:)
    Reply
  • crisan_tiberiu
    Still, where is Bulldozer? i have expected it on 20.06 ^^. I went i7 2600k on month ago, could not wait anymore :(
    Reply
  • mactronix
    We are committed to working with like minded companies that want to give consumers and business users an accurate, honest measure of what they can expect from their PCs and mobile devices,"

    Yes thats all very noble but the point is if AMD are involved then Nvidia people will start shouting biased.
    We already have the info needed to make an informed decision its out there on this and various other sites. Common sense says you look across a selection of sites to get a clearer more accurate picture. Some sites are after all leaning one way or the other.
    Mactronix
    Reply
  • mkrijt
    I found some interesting read over here. Ok, It's old but I think it does cover the problems with SYSmark...

    In fact, Fog points out that even benchmarking programs are affected by this, up to a point where benchmark results can differ greatly depending on how a processor identifies itself. Ars found out that by changing the CPUID of a VIA Nano processor to AuthenticAMD you could increase performance in PCMark 2005's memory subsystem test by 10% - changing it to GenuineIntel yields a 47.4% performance improvement!
    Reply
  • crisan_tiberiu
    mkrijtI found some interesting read over here. Ok, It's old but I think it does cover the problems with SYSmark... ermm...i believe that ^^
    Reply
  • KT_WASP
    SYSmark should have been abandoned years ago by the companies that are leaving it now (its basically an Intel marketing tool anyway). What I hate is that tech/review sites still use it, even though they know how skewed it is towards Intel.

    SYSmark = synthetic benchmark that falsely gives Intel the edge every time. Intel controls it and no input/suggestions are ever implemented from the other companies that finally said "enough".
    Why no mention of Nvidia and VIA leaving at the same time as AMD? It was a mass exodus leaving only Intel. The deal is, these companies are finally fed up and are dropping this sham of a benchmark.

    Reply
  • palladin9479
    Yep, the big benchmarking companies have been in bed with the highest payer (usually Intel) for years now. Usually it takes the form of checking CPUID and using that to determine code path and optimizations. You then make the code for your preferred CPU more "efficient" and leave everyone else on a non-optimized path. This isn't only in benchmarking but also in software development. Lets say publisher A is making a game, and decides to "work with" Intel to "get the most" out of their code. Intel provides a SDK with a compiler they say will help optimize their code. The compiler does indeed optimize the code, but only for Intel branded systems, and further the compiler may intentionally compile the program so that it runs a very inefficient non-optimized path if it detects its running on a "non-Intel" system. After all Intel can claim they only support their own products and have no requirement to support competitors products.

    This is why I take ~all~ "benchmarks" with a grain of salt. Different code paths can swing performance 10~40% depending.
    Reply
  • warmon6
    crisan_tiberiuStill, where is Bulldozer? i have expected it on 20.06 ^^. I went i7 2600k on month ago, could not wait anymore
    sometime between next month to late September area is all the info i got. Although this info been out for a little while.
    Reply