|AMD Phenom-II X6-1075T CPU HDT75TFBGRBOX|
|Reviews - Featured Reviews: Processors|
|Written by David Ramsey|
|Tuesday, 21 September 2010|
Page 7 of 12
Everest Ultimate Tests
Lavalys EVEREST Ultimate Edition is an industry leading system diagnostics and benchmarking solution for enthusiasts PC users, based on the award-winning EVEREST Technology. During system optimizations and tweaking it provides essential system and overclock information, advanced hardware monitoring and diagnostics capabilities to check the effects of the applied settings. CPU, FPU and memory benchmarks are available to measure the actual system performance and compare it to previous states or other systems. Furthermore, complete software, operating system and security information makes EVEREST Ultimate Edition a comprehensive system diagnostics tool that offers a total of 100 pages of information about your PC.
All of the benchmarks used in this test— Queen, Photoworxx, ZLib, and AES— rely on basic x86 instructions, and consume very low system memory while also being aware of Hyper-Threading, multi-processors, and multi-core processors. Of all the tests in this review, Everest is the one that best isolates the processor's performance from the rest of the system. While this is useful in that it more directly compares processor performance, readers should remember that virtually no "real world" programs will mirror these results.
Queen and Photoworxx tests are synthetic benchmarks that operate the function many times and over-exaggerate by several magnitudeswhat the real-world performance would be like. The Queen benchmark focuses on the branch prediction capabilities and misprediction penalties of the CPU. It does this by finding possible solutions to the classic queen problem on a chessboard. At the same clock speed theoretically the processor with the shorter pipeline and smaller misprediction penalties will attain higher benchmark scores.
Here we see the $164 AMD 965 Black Edition handily beating the $195 Intel Core i5-750, while the $245 AMD 1075T and $299 AMD 1090T run pretty evenly with Intel's $280 i7-860 and $285 i7-930. The overclocked 1075T beats everything except the i7-980X, which runs away from the pack with results 35% better than the next closest competitor.
Like the Queen benchmark, the Photoworxx tests for penalties against pipeline architecture. The synthetic Photoworxx benchmark stresses the integer arithmetic and multiplication execution units of the CPU and also the memory subsystem. Due to the fact that this test performs high memory read/write traffic, it cannot effectively scale in situations where more than two processing threads are used. The EVEREST Photoworxx benchmark performs the following tasks on a very large RGB image:
Here, we see some interesting results: the lower-end processors tend to do as well as or better than most of their higher-end brethren. This is especially obvious in the AMD camp where the 965 Black Edition performs better than the overclocked 1075T. The worst performance is turned in by the Intel i7-860, and the best, as usual, by the 980X. The reason for these seemingly odd results is that the higher end processors tend to have longer pipelines and more advanced out-of-order execution and branch prediction capabilities. These features can result in substantial performance improvements much of the time, but the longer pipelines come with a penalty: speculative execution of code initiated by a branch predictor is all wasted if the branch goes the other way, so not only did the processor waste thousands or millions of clock cycles executing code it didn't need to, it must also re-load its pipeline with the instructions it was supposed to execute. Developers using Intel compilers should read the several white papers Intel has written about how to structure code to minimize this problem.
The Zip Library test measures combined CPU and memory subsystem performance through the public ZLib compression library. ZLib is designed as a free lossless data compression library for use on virtually any computer hardware and operating system. The ZLib data format is itself portable across platforms and has a footprint independent of input data that can be reduced at some cost in compression. The AES integer benchmark measures CPU performance using AES data encryption. It utilizes Vincent Rijmen, Antoon Bosselaers and Paulo Barreto's public domain C code in ECB mode and consumes 48 MB of memory. Both of these tests are much more applicable to the "real world" than the previous tests.
The results of the Zip compression test scale almost linearly with processor clock speed on the quad-core Intel side, but there's some threading going on as can be seen with the Core i7-980X results, where two extra cores and and extra 530MHz or so virtually double the performance of the 2.8GHz it-860 and i7-930. Extra cores help on the AMD side, too, with the stock-clocked 3.0GHz 1075T handily outperforming the 3.4GHz quad-core 965 Black Edition.
The AES encryption test is utterly dominated by the Core i7-980X, whose performance is 6.7x better than its closest competitor. The reason is Intel's Advanced Encryption Standard New Instructions (AES-NI), which dramatically accelerate AES code. AES-NI aside, the AMD processors uniformly perform better than the Intel processors without this feature.