We proceed with the low-level analysis of the most important characteristics of high-performance DDR2 memory modules using our RightMark Memory Analyzer. Our today's analysis is devoted to the dual channel kit of high-end HyperX-series Kingston modules, intended for the unofficial DDR2-900 mode (PC2-7200). This kit comes in 512 MB, 1 GB, and 2 GB modifications, which differ from each other in memory size only. We already reviewed Kingston modules of this series (KHX6000D2K2/1G DDR2-750 modules). Nevertheless, we didn't test modules in high-speed DDR2-667, DDR2-800, and higher modes at that time yet — mostly because there was no need in such modes (in fact, these modes are not needed even now). In this article we'll analyze performance of these modules (1GB modification) in standard DDR2-667 and DDR2-800 modes as well as in unofficial DDR2-900 mode, which is still mentioned by the manufacturer.

Manufacturer Information

Module manufacturer: Kingston Technology
Manufacturer of module chips: Infineon
Web site of the module manufacturer:
http://www.kingston.com/hyperx/products/khx_ddr2.asp
Web site of the chip manufacturer: here

Module Exterior

Photo of the memory module

Photo of the memory chip

Module Part Number

Module Part Number Expansion

The manufacturer's web site does not publish the DDR2 Part Number expansion. Datasheet on the modules with Part Number KHX7200D2K2/1G specifies that the product is a kit of two 512MB modules, 64M x 64, based on 64M x 8 chips. The modules belong to the unofficial DDR2-900 group and can operate in this mode with CAS# latency = 5. The manufacturer guarantees 100% stable operation in DDR2-900 mode with 5-5-5-15 timings and 2.0 V voltage. But the default mode written in the SPD chip is standard DDR2-667 with 5-5-5-15 timing and the voltage of 1.8 V.

Chip Part Number Expansion

Unlike previous high-speed modules of the HyperX series (KHX6000D2K2/1G), these modules use chips with the original designation of their real manufacturer (Infineon), which allows us to examine their characteristics using the latest nomenclature version of DRAM components available on the manufacturer's web site.

These characteristics show that the modules are based on 256Mbit chips with 32M x8 organization. What's most interesting, these chips are intended for DDR2-667 mode, moreover, for its modification with slower timings 5-5-5 (faster DDR2-667 with 4-4-4 timings is denoted by "3"), which contrasts the official characteristics (DDR2-900, 5-5-5-15). But it may be the reason why the default mode written in SPD is the standard DDR2-667 mode, 5-5-5-15 timings, 1.8V.

SPD chip data

Description of the general SPD standard:

• JEDEC Standard No. 21-C, 4.1.2 - SERIAL PRESENCE DETECT STANDARD, General Standard

Description of the specific SPD standard for DDR2:

• JEDEC Standard No. 21-C, 4.1.2.10 - Appendix X: Specific SPDs for DDR2 SDRAM (Revision 1.2)

SPD contents look quite standard. Three different CAS# latencies are supported — 5, 4, and 3. The first value (CL X = 5) corresponds to DDR2-667 (3.0 ns cycle time) with 5-5-5-15 timings, which matches the official values, specified in the documentation. The second value t_CL (CL X-1 = 4), as usual, corresponds to DDR2-533 with 4-4-4-12 timings. And finally, the third value (CL X-2 = 3) — DDR2-400 with 3-3-3-9 timings. The module features a relatively high minimum refresh to active/refresh command period t_RFC = 105.0 ns, which we come across rather often in high-speed modules. SPD revision, Manufacturer’s JEDEC ID Code, Manufacturing Date and Serial Number of the module are correct. But there is no data on the module's Part Number.

Testbed configuration

• CPU: Intel Pentium 4 Extreme Edition 3.73 GHz (Prescott N0, 2 MB L2)
• Chipset: Intel 975X
• Motherboard: ASUS P5WD2-E Premium, BIOS 0206 dated 12/21/2005
• Memory: 2x512 MB Kingston HyperX DDR2-900

Test Results

Performance tests

In the first series of tests we used the timing scheme, set in BIOS Setup by default (Memory Timings: "by SPD"). The modules were tested in three modes — DDR2-667 with 200 MHz and 266 MHz FSB (memory multipliers - 1.67 and 1.25), DDR2-800 with 200 MHz and 266 MHz FSB (memory multipliers - 2.0 and 1.5), as well as "DDR2-900" with FSB overclocked to 225 MHz and 270 MHz (memory multipliers - 2.0 and 1.67).

As you can see on the table below, BIOS sets correct timings by default for DDR2-667 mode — 5-5-5-15, while in case of DDR2-800 and DDR2-900 modes the motherboard uses more conservative 5-6-6-18 timings by default, set by BIOS "at random", as SPD does not contain data for faster official DDR2-800 mode.

^*16 MB block size

Performance results of the modules look good — maximum memory bandwidth values in official modes amount to approximately 6.8 GB/s and 8.9 GB/s at the FSB clock set to 200 MHz and 266 MHz. This is typical of high-speed modules operating with a processor with 2MB L2 Cache. These modules are also no worse than their high-speed counterparts from the other manufacturers in latencies. As usual, latencies decrease in faster modes (when you go from DDR2-667 to DDR2-800 and DDR2-900) as well as with higher clocked FSB (when you go from 200 MHz to 266/270 MHz FSB). Thus, minimal memory latency is demonstrated in DDR2-900 mode with the 270 MHz FSB, falling between 43.1 ns (pseudo-random walks, hardware prefetch enabled) and 109.5 ns (random walks, hardware prefetch disabled).

Stability tests

Timing values, except for t_CL, were adjusted "on the fly" due to the built-in RMMA feature that allows to change dynamically memory settings supported by the chipset. Memory operating stability was evaluated with an auxiliary utility RightMark Memory Stability Test, included into RMMA.

^*16 MB block size

Minimum timings, we managed to obtain in DDR2-667 mode with the recommended voltage of 2.0 V, look impressive — 3-3-2 (as usual, the last parameter t_RAS does not participate in the overclocking scheme, as it's ignored). A smaller, absolute record — 3-2-2 (timings cannot be reduced any further) with the recommended voltage of 2.1 V was broken only by Corsair XMS2-5400UL modules. The same can be said about DDR2-800 mode — in this case, minimal timings in the modules under review are 4-4-2, again slightly outperformed by Corsair XMS2-5400UL modules with their record of 4-3-2. And finally, minimal timings in the overclocked unofficial DDR2-900 mode are no less attractive — to obtain stable operation of the modules, we only had to increment t_RP by one compared to the previous scheme.

As usual, "extreme" timings only insignificantly increase memory bandwidth, as it's still limited by FSB bandwidth. Maximum effect of this "timings overclocking" is noticeable only in terms of latencies, only in case of random memory access at that — latencies are reduced by 10%.

Bottom line

Kingston HyperX DDR2-900 (PC2-7200) modules under review (1GB dual channel kit) proved themselves as high-end high-speed modules that can operate in official DDR2-667 and DDR2-800 modes as well as in the unofficial DDR2-900 mode, which is still mentioned by the manufacturer (with relatively low voltage of 2.0 V, compared to the nominal value). The modules offer high performance and noticeable "overclocking of timings" in the modes we tested — in the official DDR2-667 and DDR2-800 modes they can operate with 3-3-2 and 4-4-2 timings (which is very close to the records, broken by Corsair XMS2-5400UL modules), while operation in DDR2-900 mode requires increasing timings to "extreme" 4-4-3.

Write a comment below. No registration needed!