tistou77 Posted October 6, 2010 Share Posted October 6, 2010 Hi I test AIDA64 and I think there is a bug with the memory benchmark (write) and also, with Everest's benchmark was unreliable because it was the benchs mono thread, I think is like AIDA64 is this an update is planned for the memory benchmark is OK (multi thread)? thanks Quote Link to comment Share on other sites More sharing options...
Fiery Posted October 6, 2010 Share Posted October 6, 2010 1) We'll check what's up with the memory write benchmark. However, this issue seems to apply to only Gulftown processors. 2) Why would you think AIDA64 memory benchmarks would be "unreliable", just because they're not multi-threaded? Quote Link to comment Share on other sites More sharing options...
tistou77 Posted October 6, 2010 Author Share Posted October 6, 2010 1) ok thank you;) with the benchmark Everest, no problem with the write 2) I read everywhere that the benchmark Everest (AIDA64 now) did not provide actual values (on X58 platform) compared to other software (Stream Benchmark for example) and that it could be because that AIDA64 is only mono thread for the benchmark thanks Quote Link to comment Share on other sites More sharing options...
Fiery Posted October 6, 2010 Share Posted October 6, 2010 2) I read everywhere that the benchmark Everest (AIDA64 now) did not provide actual values (on X58 platform) compared to other software (Stream Benchmark for example) and that it could be because that AIDA64 is only mono thread for the benchmark Quite frankly, that's an exaggeration that it would "not provide actual values". It's true, that our memory benchmarks are single-threaded, and that it's a different approach to measuring memory bandwidth than what certain other software use. However, it is no less than a useful benchmark, and it can be used to compare performance of various systems, or to measure a performance gain of a hardware upgrade, or the performance gain when you overclock or fine-tune your PC. We've already started developing multi-threaded memory benchmarks, but they're lower on our priority list than certain other improvements, e.g. brand new CPU benchmark methods, auto-update feature, or GPGPU benchmarks. We do feel that there's a room for improvement about our memory benchmarks, but we don't reckon the existing benchmarks to be useless or crippled at all. Our technology partners (incl. Intel Corporation) happily use the existing memory benchmarks for various purposes in their labs, and that alone proves that they're indeed reliable and useful Quote Link to comment Share on other sites More sharing options...
tistou77 Posted October 6, 2010 Author Share Posted October 6, 2010 ok thank you for this feedback and for these details to Intel, the memory benchmark AIDA64 and Everest is good then Quote Link to comment Share on other sites More sharing options...
Fiery Posted October 7, 2010 Share Posted October 7, 2010 I've just sent you a private message about the memory write benchmark issue. Quote Link to comment Share on other sites More sharing options...
tistou77 Posted October 7, 2010 Author Share Posted October 7, 2010 it's ok, I sent you an mail Quote Link to comment Share on other sites More sharing options...
Fiery Posted October 8, 2010 Share Posted October 8, 2010 Make sure to upgrade to the latest beta version of AIDA64 Extreme Edition available at: http://www.aida64.com/downloads/aida64extremebuild1114o1iwfjk6ldzip After upgrading to this new version, make sure to restart Windows to finalize the upgrade. Let me know if it helps. Thanks, Fiery Quote Link to comment Share on other sites More sharing options...
tistou77 Posted October 10, 2010 Author Share Posted October 10, 2010 it's ok, now thanks Quote Link to comment Share on other sites More sharing options...
Fiery Posted October 10, 2010 Share Posted October 10, 2010 Thanks for the feedback Quote Link to comment Share on other sites More sharing options...
Recommended Posts
Join the conversation
You can post now and register later. If you have an account, sign in now to post with your account.