Re: Overclocking of QX9650
Many out there is the recently purchased processor was not fast enough. An effective way to satisfy this need for more speed is to overclock your hardware. As a small sample stability, we let the 3DMark06 and run Prime 95 and subsequently tested the highest clock by using a few selected applications. But you have to say before the measurements that the results can not be transferred to each processor of the same type as the quality of chip to chip is different. For the test, we raised the multiplier to 10 and increases the front-side bus to 1,600 MHz. This overclocked, of course, also the memory that runs at 1,600 MHz but still fully in the specifications. The mainboard from Gigabyte did these things with good grace, not for nothing that the company is the board also certifies FSB1600 with equivalent memory. For a shot at more stability provides a small increase in voltage of the processor, which are therefore all tests run correctly.
Re: Overclocking of QX9650
As expected, shows that the CPU plays modern games do not matter too much more. Crysis ranks as the simulated one of us 4 GHz Quad-Core FSB1600 still not as fastest model, leaving the highest settings remain hidden.
A little different the picture in the theoretical tests and including Cinebench R10. The memory bandwidth is increased due to the increased front side bus, of course, a few percent, Cinebench R10 benefited significantly from the pure processor speed, however.
Re: Overclocking of QX9650
Less in a sink with the power consumption is the determination of the maximum temperature. This showed that Prime is not sufficient to temperature of the processor to achieve the maximum - of the display in the system that all processor cores to full capacity are. 100 percent despite Core2MaxPerf that we include in our cooler tests use under would be at this point been the better tool, since it the processor more effectively loaded and thus generates higher temperatures, but it recognizes that the new 45-nm Intel processor yet. Therefore, a combination of different applications, serve to determine the maximum value. To this end, the closed system for an hour in advance 3DMark06 loop heated and the temperature is then the combination of Prime95 and 3DMark06 determined. The room temperature was always around 21 degrees, the temperature read from the motherboard to periods of full load at maximum 46 degrees. The image of the power consumption is also reflected in the temperatures. The new four-core processor from Intel impressed with temperatures similar to those of the two-core processors. With its 3.0GHz QX9650 is the cooling of a lot easier than the 2.4 GHz Q6600 fast. The old QX6850 at 3.0 GHz is almost 20 degrees Celsius over the new model, even if the temperature measurement of individual nuclei is not so much different. The 45-nm technology is playing a whole, but once again the stockings.
Re: Overclocking of QX9650
Of course, we have dealt with the new processor generation, the issue of power consumption. In such amazing results came to light that we have repeated the test a half dozen times to completely ensure that the values are real. As usual, we measure in our system, the power consumption of the entire PC. We will measure the "Voltcraft Energy Check 3000" on the one hand power consumption in the BIOS, in any application - the Idling - in Windows, and under full load. To fully utilize the machine actually requires a program that both appeals to all processor core, and the graphics card calls completely. Since the meter has a recording function which suspicious programs were quickly found. The processor is called Prime 95 on all four cores fully, in parallel, the two graphics tests "Firefly Forest" & "Canyon Flight" from 3DMark06 in a loop, because at this current consumption is not even approached a DirectX 10 application. Besides recording function stores the meter off and the maximum and minimum, it is found that after one hour duration test, the value for the most power. 85 watts difference in power consumption caused the author initially doubt his abilities. If there really such a serious difference between two equally clocked processors in a complete system? But even after the trials three, four, five and eleven was always the same result to the fore. In the endurance test, consisting of Prime 95 and 3DMark06, requires the old Core 2 Duo with the complete system up to the 470 watts, while the needle deflects maxmal up short of the 490 watts. The successor needs in conjunction with the Radeon HD 2900 XT and the remaining components in the same cases on average "only" 380 watts, the maximum deflection is also here a little over 400 watts was measured. Already in the BIOS and no activity in Windows, the latest offshoot won by his power-saving features and technical innovations lower power for themselves.
Re: Overclocking of QX9650
To re-examination, we conducted the test under Windows Vista with no maximum graphics card influence with Prime 95th But after the start was only the tool started and operated with the power meter after a few minutes, the recording function.
The measurements showed the Core 2 Extreme QX6850 values 319-339 watts, while the new offshoot QX9650 with a maximum of 250 to 272 Watt went to work. On average, this still corresponds to a difference of 68 watts. These are explained on the one with the available voltage, which is specified in the BIOS QX6850 with 1.3125 volts, the QX9650 needs at the same point 1.2500 volts. As a comparison processor once the model operates with two cores - the X6800 at 2.93 GHz at 1.3250 volts, in which the system has an average of 241 watts with no load on the graphics card does not despise food. The reasons for the still very big difference is in the details, such as the available voltage, but especially in the new technology. Intel indicates that the use of 45nm Hi-K metal gate silicon technology can ensure that the new processors will need about 30 percent less switching power and a more than 20 percent faster response speed or five times less leakage currents have. In all areas but especially in the notebook segment, in addition to high performance is above all an even higher energy efficiency in its efforts by Intel, which seems to have succeeded with the first impression.
Re: Overclocking of QX9650
To provide a fair and realistic comparison between the counterparties, all tests in a closed Midi-Tower with complete fan assembly have been (a fan back inlet, one at the hard drive cage sideways blowing) carried out in order to call attention to thermal problems in the cars to be. As part of the new processor generation, we have enhanced the complete test system. Benchmarks are now under a newly installed Windows Vista Ultimate instead of the 32-bit version. Not only with the software, the foundation stone was laid for further testing, the hardware has undergone an extensive update. In order to exhaust fully current and future processors, we have decided on the Intel test system for a motherboard with an X38 chipset that is equipped with 2 GB of DDR3 memory. In collaboration with a Radeon HD 2900 XT, this system provides for the coming months, the reference. In the same breath, of course, all drivers updated and the benchmarks on its feet.
Re: Overclocking of QX9650
As processors is next to the new Core 2 Extreme QX9650 based on the Yorkfield core, its predecessor, the QX6850 with core in the G0-stepping, are used. Both processors use 333 MHz front side bus, supports DDR3-1333 so as working memory tasks of the processors. For the comparison to the entry level into the market of the four cores was used with a Q6600 2.40GHz G0 Stepping in with DDR3-1066. Also, as a highly clocked dual-core subsidiary of Intel, will contribute the old warhorse X6800 DDR3-1066 memory with its values. For comparison, we round down from the test with a Core 2 Duo E6600.
Processor
- Intel Core 2 Extreme QX9650 - 3.00 GHz, 12MB L2 cache, FSB1333 (Wolfdale C0)
- Intel Core 2 Extreme QX6850 - 3.00 GHz, 8 MB L2 cache, FSB1333 (Kentsfield G0)
- Intel Core 2 Extreme X6800 - 2.93 GHz, 4 MB L2 cache, FSB1066 (Conroe B1)
- Intel Core 2 Quad Q6600 - 2.40 GHz, 8 MB L2 cache, FSB1066 (Kentsfield G0)
- Intel Core 2 Duo E6600 - 2.40 GHz, 4 MB L2 cache, FSB1066 (Conroe B1)
Motherboard
- Gigabyte X38T-DQ6 (X38) - Rev. 1.0 BIOS: F3h
Memory
- 2x 1024 MB OCZ Platinum DDR3-1333 (CL7-7-7-20, Dual-Channel)
- 2x 1024 MB OCZ Platinum DDR3-1066 (CL6-6-6-16, Dual-Channel)
Video Card
- ATI Radeon HD 2900 XT (742/828), 512 MB
Driver Versions
- ATi Catalyst 7.10
- Intel Chipset Driver 8.3.1.1009 (for X38 chipset)
