Logo   Banner   TopRight
DDR2 Shoot-Out: Corsair vs. Kingston
Date: Oct 27, 2004
Author: Marco Chiappetta
Introduction & Specifications


With the launch of their 900 series chipsets, and LGA775 based Pentium 4 processors, Intel introduced the use of DDR2 system memory on the desktop.  DDR1 system memory had been the standard for quite a few years, and DDR2 memory had been used on the occasional video card at that point, but DDR2 system memory had yet to be implemented in a consumer level desktop computer.  DDR1 and DDR2 memory are fundamentally very similar.  In essence, DDR2 memory is simply a refinement of the original, however, there are some differences that set them apart.  For example, DDR2 memory requires much less voltage to achieve similar clock speeds, which means lower power consumption.  DDR2 memory also has on chip termination which yields better EMI characteristics, and DDR2 memory can be clocked significantly higher than DDR1 memory given the right circumstances; three very sought after traits.  For a more comprehensive explanation on the differences between DDR1 and DDR2 memory, take a peek at this PDF, and for a look into Intel's implementation of DDR2 in their 900 series chipsets, check out Dave's coverage right here.

DDR2 system memory does have its disadvantages though, at least for now.  Access latencies are currently much higher with DDR2 memory.  It's not uncommon to find DDR1 memory with a CAS Latency of 2T, a RAS to CAS Delay (tRCD) of 2T, a Row Precharge (tRP) of 2T and an Active to Precharge Delay (tRAS) of 5T (2-2-2-5) at 200MHz (DDR400).  But most of the DDR2 memory available today is officially rated for 4-4-4-12 timings.  DDR2 memory is also more expensive than DDR1 memory with a similar density.  These disadvantages will eventually disappear, however, as lower latency DDR2 memory is on the horizon and prices will inevitably fall as the use of DDR2 system memory becomes more common.  Most of the major players in system memory already have DDR2 memory kits in their product lineup.  And today on HotHardware, we're going to take a look at two of them, Corsair's TWIN2X1024-5400C4PRO and Kingston's KHX5400D2K2/1G...

Specification & Features: Corsair TWIN2X1024-5400C4PRO
More Information: www.corsairmicro.com
The TWIN2X1024-5400C4PRO is a 1024 MByte matched pair of DDR2 SDRAM DIMMs. This part delivers outstanding performance in the latest generation of DDR2-based motherboards. It has been tested extensively in multiple DDR2 motherboards to ensure compatibility and performance at its rated speed. This memory has been verified to operate at 675MHz at the low latencies of 4-4-4-12. This module is also available in a single module part.
·_1024 Megabytes of memory in a pair of 240-pin DDR2 DIMMs
·_Two matched CM2X512-5400C4PRO modules
·_Implemented using 32M x 8 DDR2 SDRAMs
·_100% tested at 675MHz in high performance DDR2 motherboards
·_Legendary Corsair reliability and service
·_Lifetime warranty
Test Specs:
·_Each module pair is tested together in an Intel-based motherboard at 675MHz
·_Tested and packaged in pairs
·_Packaged together immediately following system test
·_Tested at 4-4-4-12 latency settings at 1.9V
·_SPD programmed at 4-4-4-12 values



Like Corsair's other "Pro" memory kits, the memory modules included in the TWIN2X1024-5400C4PRO kit feature 18 activity LEDs along the top, that blink rapidly to show the level of memory activity, much like the sound meters found on some older audio equipment.  Each of the two 512MB modules included, for a total of 1GB, are rated for clock speeds of up to 675MHz (337MHz DDR) with 4-4-4-12 latency settings at 1.9v.  Bundled with the modules are a simple installation manual and a Corsair case badge.  We'd also like to point out the relative size of the TWIN2X1024-5400C4PRO sticks.  The large heat spreaders and activity LEDs make these modules significantly taller than most others.


Specification & Features: Kingston KHX5400D2K2/1G
More Information: www.kingston.com
Kingston's KHX5400D2K2/1G is a kit of two 64M x 64-bit (512MB) DDR2-675 CL4 SDRAM (Synchronous DRAM) memory modules, based on eight 64M x 8-bit DDR2 FBGA components (per module). Total kit capacity is 1GB (1024MB). Each pair has been tested to run at DDR2 675MHz at low latency timing of 4-4-4-10 at 1.85V. The SPDs are programmed to JEDEC standard latency timing of 4-4-4-12 at 1.8V. Each 240-pin DIMM uses gold contact fingers and requires +1.8V. The electrical and mechanical specifications are as follows:
·_Power supply : Vdd: 1.8V ~ 0.1V, Vddq: 1.8V ~ 0.1V
·_Double-data-rate architecture; two data transfers per clock cycle
·_Bidirectional data strobe(DQS)
·_Differential clock inputs(CK and CK)
·_DLL aligns DQ and DQS transition with CK transition
·_Programmable Read latency 4 (clock)
·_Burst Length: 4, 8 (Interleave/nibble sequential)
·_Programmable Burst type (sequential & interleave)
·_Timing Reference: 4-4-4-10 at +1.85V
·_Edge aligned data output, center aligned data input
·_Auto & Self refresh, 7.8us refresh interval (8K/64ms refresh)
·_Serial presence detect with EEPROM
·_High Performance Heat Spreader
·_PCB : Height 1.180" (30.00mm), single sided component
·_Clock Cycle Time (tCK) CL=4 - 3.75ns (min.) / 8ns (max.)
·_Row Cycle Time (tRC) - 60ns (min.)
·_Refresh to Active/Refresh Command Time (tRFC) - 105ns
·_Row Active Time (tRAS) - 45ns (min.) / 70,000ns (max.)
·_Single Power Supply of +1.8V (+/- .1V)
·_Power - 1.512 W (operating per module)
·_UL Rating - 94 V - 0
·_Operating Temperature - 0o C to 55o C
·_Storage Temperature - -55o C to +125o C


Kingston's KHX5400D2K2/1G memory kit looks more "traditional" than Corsair's, and has slightly better specifications.  Kingston's modules also feature standard sized heat spreaders, and thus look just like DDR1 modules, with the exception obviously being the number on gold contacts along their edges (DDR2 and DDR1 modules are not interchangeable.  DDR2 modules feature 240 pins, DDR1 modules feature 184).  These Kingston sticks are rated for 4-4-4-10 latency settings at 675MHz with 1.85v.  The kit includes two 512MB modules, for a total of 1GB, and ship with a simple installation guide.

Compatibility & Our Test System


Compatibility Testing: When testing memory modules, we generally like to try them in as many different motherboards as possible, to get a general idea if any incompatibilies exist.  Unfortunatly, because of the relatively small number of motherboards based on Intel's 900 series of chipsets, we were only able to test these modules on three different boards.  We tested both the Corsair TWIN2X1024-5400C4 PRO and Kingston KHX5400D2k2/1G memory modules in the Foxconn 925A01-8EKRS, the Intel D925XCV, and DFI's LANPARTY 925X-T2, and both kits worked perfectly in all three motherboards.

Corsair TWIN2X1024-5400C4PRO

Kingston KHX5400D2K2/1G

We also fired up the latest version of CPU-Z to see how the SPDs (Serial Presence Detect) were programmed. As you can see the Corsair modules are programmed for 4-5-5-12 operation at 333MHz (DDR667).  Kingston's modules, however, are programmed to run at different timings depending on clock speed.  At 200MHz, Kingston's sticks are programmed for 3-3-3-9 timings at 200MHz (DDR400), 4-4-4-12 at 266MHz (DDR533), or 5-4-4-12 at 266MHz (DDR533).

HotHardware's Test System
Not all are created equal...
Socket T - Pentium 4 560 (3.6GHz)

ADFI LANPARTY 925X-T2 Motherboard
VIA K8T800 Pro Chipset

Corsair TWINX 1024-3200XLPRO
Kingson HyperX PC3500

Radeon X700 XT
On-Board 10/100/1000 Ethernet
On-Board Audio

WD "Raptor" 36GB Hard Drive
10,000 RPM SATA

Windows XP Pro SP2
ATi Catalyst 4.10

How we configured our test systems:  When configuring our test systems for this review, the first thing we did was enter the system BIOS and set each board to their "Optimized" or "High-Performance Defaults".  The hard drives were then formatted, and Windows XP Professional (SP2) was installed. When the installation was complete, we hit the Windows Update site and downloaded all of the available updates, with the exception of the ones related to Windows Messenger.  Then we installed all of the necessary drivers, and removed Windows Messenger from the system altogether.  Auto-Updating and System Restore were also disabled, and we setup a 768MB permanent page file on the same partition as the Windows installation.  Lastly, we set Windows XP's Visual Effects to "best performance", installed all of our benchmarking software, defragged the hard drives and ran all of the tests.  Throughout all of our standard benchmarking, we had the memory voltage set to 1.9v.  When overclocking, or experimenting with lower latencies, we set the memory voltage to 2.0v.

Performance with SPD Settings


We began our testing with SiSoftware's SANDRA, the System ANalyzer, Diagnostic and Reporting Assistant, with the memory configured by their SPDs. SANDRA consists of a set of information and diagnostic utilities that can provide a host of useful information about your hardware and operating system.  We ran SANDRA's Memory Bandwidth test with Corsair's and Kingston's memory modules installed on our DFI LANPARTY 925X-T2...

(Note:  Although we had the DFI motherboard we used for testing was configured to detect memory timings by SPD, it ran both memory kits with 4-4-4-12 timings, unless we manually specified otherwise.)

Performance Comparison with SiSoft SANDRA 2004
Raw Bandwidth

Corsair TWIN2X1024-5400C4PRO

Kingston KHX5400D2K2/1G

SANDRA's memory bandwidth benchmark had the Corsair modules finishing just slightly ahead of Kingston's.  In the integer test, the Corsair modules had a bandwidth score 76MB/s higher than the Kingston modules.  In the Floating Point test, the Corsair sticks posted a score 58MB/s higher than the Kingston sticks.

Performance Comparison with PCMark04
Overall Memory Score

For our next round of benchmarks, we ran the Memory performance module built-into Futuremark's PCMark04. For those interested in more than just the graphs, we've got a quote from Futuremark that explains exactly what this test does and how it works...

"The Memory test suite is a collection of tests that isolate the performance of the memory subsystem. The memory subsystem consists of various devices on the PC. This includes the main memory, the CPU internal cache (known as the L1 cache) and the external cache (known as the L2 cache). As it is difficult to find applications that only stress the memory, we explicitly developed a set of tests geared for this purpose. The tests are written in C++ and assembly. They include: Reading data blocks from memory, Writing data blocks to memory performing copy operations on data blocks, random access to data items and latency testing."

The SiSoft SANDRA results mirrored what we saw with PCMark04.  When using PCMark04's memory performance module, the Corsair TWIN2X1024-5400C4 PRO kit posted a score 95 points higher than the Kingston KHX54002K2/1G kit.

In-Game Performance Comparisons With Wolf: ET
System Memory Affects Frame rates?  You Bet!

We also ran through a batch of time demos with the OpenGL game Wolfenstein: Enemy Territory.  Wolfenstein: ET is a free, standalone multiplayer game that is based on the original Return to Castle Wolfenstein that was released a few years back. It uses a heavily modified version of the Quake III engine, which makes it a very easy-to-use benchmarking tool.  We ran this test at the "Fastest" setting with a low resolution of 640 X 480, using 16-bit color and textures.  Running this test with a high-end graphics card at these minimal settings isolates processor and memory performance without being limited by the graphics subsystem. 

Performance with Wolfenstein: Enemy Territory was very similar, with either of the memory kits installed.  Once again, however, the Corsair TWIN2X1024-5400C4 PRO kit posted the highest score.  But with only 2.1 frames per second separating each kit, this test is essentially a tie.

Performance with Lowest Latency


To get this next batch of results, we manually lowered the latencies of each memory kit to 3-3-3-10, raised the memory voltage to 2.0v, and slowly raised their clock speeds until our test system was no longer stable. Although these kits were rated for 4-4-4-12 operation, they were both able to run with 3-3-3-10 timings at DDR533 speeds without a problem. But they each peaked at a slightly different clock speed with these lower latencies...

Performance Comparison with SiSoft SANDRA 2004
Raw Bandwidth

Corsair TWIN2X1024-5400C4PRO
207MHz FSB / 276MHz (DDR552)

Kingston KHX5400D2K2/1G
208MHz FSB / 277MHz (DDR554)

The Corsair TWIN2X1024-5400C PRO modules were able to run with 3-3-3-10 timings at a front side bus speed of 207MHz.  With the memory running at a 3:4 FSB to Memory clock ratio, a 207MHz FSB equates to a 276MHz memory clock.  The Kingston modules were able to hit a slightly higher clock speed though.  The KHX5400D2K/1G modules hit 208MHz (DDR554) with the tighter 3-3-3-10 timings.  The slight clock speed advantage didn't equate to much extra performance, however, as only a few MB/s separated each kit in this test.

Performance Comparison with PCMark04
Overall Memory Score

We also re-ran the PCMark04 memory performance module with these memory kits clocked at their respective peaks, with 3-3-3-10 timings.  This time around the Kingston modules were able to pull ahead of the Corsair modules by 5 points.  As expected, when compared to their "stock" scores, both memory kits performed much better with lower latencies.  They each gained about 250-300 points in this test when running at 3-3-3-10 at slightly higher frequencies.

In-Game Performance Comparisons With Wolf: ET
System Memory Affects Framerates?  You Bet!

When we re-ran the Wolfenstein: Enemy Territory test, performance once again increased by a few percentage points.  In this test, the Kingston KHX5400D2K/1G kit pulled every so slightly ahead of the Corsair kit, but the .7 frames per second difference falls well within the margin of error in this test.  When compared to the stock results, both kits managed to increase their performance by 6-9 FPS.

Performance at Highest Stable Overclock


For our last batch of tests, we set each kit to run with their SPD timings (4-4-4-12), and raised our test system's FSB as high as possible.  To minimize the affect our Pentium 5 560 3.6GHz CPU had on the overclocking results, we raised its voltage slightly and dropped its multiplier to 14 before ratcheting up the FSB...

Performance Comparison with SiSoft SANDRA 2004
Raw Bandwidth

Corsair TWIN2X1024-5400C4PRO
272MHz FSB/ 362MHz (DDR725)

Kingston KHX5400D2K2/1G
276MHz FSB / 368MHz (DDR736)

The Kingston KHX5400D2K/1G kit overclocked a bit higher than Corsair's.  The TWIN2X1024-5400C4 PRO kit hit a maximum clock speed of 362MHz, while the Kingston modules maxed out at 368MHz.  The performance results, however, show the Corsair modules posting higher bandwidth scores.  We double-checked that both kits were running at the same timings (4-4-4-12), and they were in fact configured the same.  We repeated the tests a few times with each kit, and came up with similar results each time.

Performance Comparison with PCMark04
Overall Memory Score

The overclocked PCMark04 results were in line with what we expected, considering the Kingston modules were clocked slightly higher than the Corsair modules, although only 30 points separated the competitors.  As expected, overclocking had a huge effect on performance, increasing scores by about 1200 points for both kits.

In-Game Performance Comparisons With Wolfenstein: ET
System Memory Affects Framerates?  You Bet!

Overclocking also had a significant effect on frame rates in our custom Wolfenstein: Enemy Territory test.  Here, the Kingston KHX5400D2K/1G kit nudged a bit ahead of the Corsair modules (147.6 vs. 146.2), but once again the performance delta fell well within the margin of error in this test.

Summary & Conclusion


Benchmark Summary: Corsair's TWIN2X1024-5400C PRO modules were slightly faster then the Kingston modules is a few of the benchmarks we ran, but the Kingston KHX5400D2K/1G modules were able to run at a higher clock speed with more aggressive 3-3-3-10 timings, and they also overclocked a bit higher than the Corsair modules. Overall though, both of these DDR2 memory kits performed similarly.  The results were so similar in fact, that if we had two more identical kits from each company to test, the results could likely be reversed.

Corsair TWIN2X1024-5400C PRO:
Corsair's TWIN2X1024-5400C PRO DDR2 memory kit is ideal for overclockers, who also want to show off their system's internals.  This memory performed very well at default settings, and overclocked to an impressive clock speed of 362MHz (DDR725).  The activity LEDs installed on these modules make them unique when compared to other DDR2 memory kits, but they also cost more than their competition.  As of today, using our HotHardware pricing engine, the least expensive price we found for this kit was about $366.  That makes Corsair's TWIN2X1024-5400C PRO DDR2 memory kit about 20% more expensive than similarly spec'd products.  The oversized heat spreaded and activity LEDs somewhat justify the higher price tag, however.  In the end, owners of this memory will no doubt be pleased.  This is some of the best DDR2 memory on the market.  We're giving Corsair's TWIN2X1024-5400C PRO DDR2 memory kit a strong 8.5 on the Heat Meter.


Kingston KHX5400D2K/1G:
Kingston nearly hit a home-run with their KHX5400D2K/1G DDR2 memory kit.  Had these modules performed a bit better in their stock configuration versus Corsair's kit, we'd have nothing to take away from Kingston.  At about $292, the Kingston KHX5400D2K/1G DDR2 memory kit is some of the least expensive high-speed DDR2 memory currently available, and its performance throughout our battery of tests was very good.  And where it did "lose" a test, the margin of defeat was miniscule at best.  These modules hit a higher clock speed than Corsair's when using more aggressive 3-3-3-10 timings, and when configured at 4-4-4-12 timings they overclocked a few MHz higher than Corsair's kit as well.  Overclockers who want to hit extreme front side bus speeds, look no further.  Kingston has a great product in their line-up in the KHX5400D2K/1G DDR2 memory kit.  We're giving it a 9 on the Heat Meter.

Get into HotHardware's PC Hardware Forum Right Now!

Content Property of HotHardware.com