17 - Synthetic performance CPU & Memory
DryStone CPU test
We make use of a multi-threaded Dhrystone test, which basically is a suite of arithmetic and string manipulating programs. Since the whole program should be really small, it fits into the processor cache. It can be used to measure two aspects, both the processor's speed as well as the optimizing capabilities of the compiler. The resulting number is the number of executions of the program suite per second.
Okay, first off, quickly a couple of synthetic tests. I already told you that I want to show you the overclocking experience rather than dry test runs and numbers. I have to include a few though, so we'll get that out of the way first.
What you need to focus on are: 790i 3.2 GHz/1600 versus Intel X38 QX9770/1600. Why you ask? Because the 790i 3.2 GHz/1600 is in fact the 790 mainboard with a 260 USD Q6600 processor at a 1600 Mhz FSB @ 3.2 GHz. Now that's precisely the same clock-speed as the 1300 USD QX9770 processor running on the 200 USD X38 platform.
Where you see 1333 or 1600 - that's the FSB yet also memory clock, we have it synced 1:1 which results in great memory bandwidth along the board. Setting is at 4:5 asynchronous at 2000 MHz, made squat difference. We'll stick to 1:1
Furthermore is the AMD Phenom 9700 Quad core processor (B2 revision with TBL bug patch enabled). And then some other configurations I tried on the Q6600 processor with the 790i mainboard. As you can see on a 1600 MHz FSB I was able to scale the processor to 3600 MHz perfectly stable.
In fact, any & every benchmark you'll see today, will have results included based on that clockspeed.
Queen CPU test
This simple integer benchmark focuses on the branch prediction capabilities and the misprediction penalties of the CPU. It finds the solutions for the classic "Queens problem" on a 10 by 10 sized chessboard. At the same clock speed theoretically the processor with the shorter pipeline and smaller misprediction penalties will attain higher benchmark scores. For example -- with HyperThreading disabled -- the Intel Northwood core processors get higher scores than the Intel Prescott core based ones due to the 20-step vs 31-step long pipeline. However, with enabled HyperThreading the picture is controversial, because due to architectural bottlenecks the Northwood core runs out of internal resources and slows down. Similarly, at the same clock speed AMD K8 class processors will be faster than AMD K7 ones due to the improved branch prediction capabilities of the K8 architecture.
CPU Queen test uses only the basic x86 instructions, it consumes less than 1 MB system memory and it is HyperThreading, multi-processor (SMP) and multi-core aware and thus is a multithreading CPU Benchmark with MMX, SSE2 and SSSE3 optimizations.
An excellent test to get some synthetic performance numbers. Once clocked at similar speeds the 790i Q6600 setup is actually a tad faster than the QX9770 on Intel's X38 platform.
ZLib CPU test
This integer benchmark measures combined CPU and memory subsystem performance through the public ZLib compression library Version 1.2.2
CPU ZLib test uses only the basic x86 instructions, and it is HyperThreading, multi-processor (SMP) and multi-core (CMP) aware. Again a very good test to measure multi-core performance. We again see very similar behavior.
Memory test
With DDR3 I just had to include some sort of results for you. All the way to the right our current high-end graphics card test platform based nForce 680i SLI and 1142 MHz DDR2 memory with a Core 2 Extreme 6800 processor. Two steps to the right I have emulated the same setup by disabling two cores of the quad core processor, and lowering the FSB towards 2.9 GHz. I did leave the DDR3 at the Jedec specified clock frequency though. Observe the difference.
Anyway, that's it for boring synthetics. Let's fire up a dozen of games an see how things scale.
