iXBT Labs - Computer Hardware in Detail

Platform

Video

Multimedia

Mobile

Other

Memory Module Analysis. Part 14: Corsair XMS2-5400UL Modules




We proceed with the low-level analysis of the most important characteristics of high-performance DDR2 memory modules using our RightMark Memory Analyzer test package, started by our previous article devoted to Kingston DDR2-750. In this article we shall review Corsair modules of the traditional XMS2 series, intended for 333 MHz (DDR2-667, to be more exact, DDR2-675, as the manufacturer calls them). On the face of it, they are no different from the XMS2-5400C4 modules reviewed long ago, but this impression is misleading. The point is that the given modules can operate with the lowest 3-3-2-8 timings - not at 266 MHz, as you may expect, but in DDR2-667 mode.

Manufacturer Information

Module manufacturer: Corsair Memory
Manufacturer of module chips: unknown
Web site of the module manufacturer: http://www.corsairmemory.com/corsair/xms2.html

Module Exterior

Photo of the memory module




Module Part Number




The manufacturer's web site does not publish the DDR2 Part Number expansion of XMS2-series memory modules. The datasheet runs that the product under review is a set of two DDR2-667 512 MB modules, based on 64M x8 chips. 100% stable operation is guaranteed in DDR2-675 mode with Intel chipsets and 3-3-2-8 timings, which requires significantly higher voltage than normal — 2.1V. SPD timings are much higher — 4-4-4-15.

SPD chip data

Description of the general SPD standard:

Description of the specific SPD standard for DDR2:

Parameter Byte Value Expansion
Fundamental Memory Type 2 08h DDR2 SDRAM
Number of Row Addresses on this assembly 3 0Eh 14 (RA0-RA13)
Number of Column Addresses on this assembly 4 0Ah 10 (CA0-CA9)
Number of DIMM Banks 5 60h 1 physical bank
Data Width of this assembly 6 40h 64 bit
Voltage Interface Level of this assembly 8 05h SSTL 1.8V
SDRAM Cycle time (tCK) at maximum supported CAS# latency (CL X) 9 30h 3.00 ns (333.3 MHz)
DIMM configuration type 11 00h Non-ECC
Refresh Rate/Type 12 82h 7.8125 ms — 0.5x reduced self-refresh
Primary SDRAM Width (organization type) of the memory module chips 13 08h x8
Error Checking SDRAM Width (organization type) of the memory chips in the ECC module 14 00h Not defined
Burst Lengths Supported (BL) 16 0Ch BL = 4, 8
Number of Banks on SDRAM Device 17 04h 4
CAS Latency (CL) 18 10h CL = 4
Minimum clock cycle (tCK) at reduced CAS# latency (CL X-1) 23 00h Not defined
Minimum clock cycle (tCK) at reduced CAS# latency (CL X-2) 25 00h Not defined
Minimum Row Precharge Time (tRP) 27 30h 12.0 ns
4, CL = 4
Minimum Row Active to Row Active delay (tRRD) 28 1Eh 7.5 ns
2.5, CL = 4
Minimum RAS to CAS delay (tRCD) 29 30h 12.0 ns
4, CL = 4
Minimum Active to Precharge Time (tRAS) 30 2Dh 45.0 ns
15, CL = 4
Module Bank Density 31 80h 512 MB
Write recovery time (tWR) 36 3Ch 15.0 ns
5, CL = 4
Internal write to read command delay (tWTR) 37 1Eh 7.5 ns
2.5, CL = 4
Internal read to precharge command delay (tRTP) 38 1Eh 7.5 ns
2.5, CL = 4
SDRAM Device Minimum Active to Active/Auto Refresh Time (tRC) 41, 40 37h, 00h 55.0 ns
18, CL = 4
SDRAM Device Minimum Auto-Refresh to Active/Auto-Refresh Command Period (tRFC) 42, 40 69h, 00h 105.0 ns
35, CL = 4
Maximum device cycle time (tCKmax) 43 80h 8.0 ns
SPD Revision 62 12h Revision 1.2
Checksum for Bytes 0-62 63 15h 21 (true)
Manufacturer’s JEDEC ID Code 64-71 7Fh, 7Fh,
9Eh
Corsair
Module Part Number 73-90 CM2X512A-5400UL
Module Manufacturing Date 93-94 00h, 00h Not defined
Module Serial Number 95-98 00h, 00h,
00h, 00h
Not defined

SPD contents are not quite typical of a DDR2-667 module. There is only one CAS# latency value supported — 4, it corresponds to the 3.0 ns cycle time (333.3 MHz, DDR2-667). The timings scheme for this (only) case is 4-4-4-15, as specified by the manufacturer in the documentation. Another distinctive feature is a slightly increased minimum refresh to active/refresh command period (tRFC = 105 ns). But nevertheless, it can be found in other high-performance DDR2 modules as well. Besides, there is no data on a manufacturing date and a serial number of these modules — it actually looks strange for a high-end brand memory module.

Testbed Configurations and Software

Testbed 1

  • CPU: Intel Pentium 4 3.6 GHz (Prescott E0, 1 MB L2)
  • Chipset: Intel 945G
  • Motherboard: Foxconn 945G7MA, BIOS dated 04/12/2005
  • Memory: 2x512 MB Corsair XMS2-5400UL (DDR2-667 mode)
  • Video: Leadtek PX350 TDH, NVIDIA PCX5900
  • HDD: WD Raptor WD360, SATA, 10000 rpm, 36Gb
  • Drivers: NVIDIA Forceware 77.72, Intel Chipset Utility 7.0.0.1025, DirectX 9.0c

Testbed 2

  • CPU: Intel Pentium 4 3.6 GHz (Prescott E0, 1 MB L2)
  • Chipset: Intel 945P
  • Motherboard: Foxconn 945P7AA, BIOS dated 04/01/2005
  • Memory: 2x512 MB Corsair XMS2-5400UL (DDR2-667 mode)
  • Video: Leadtek PX350 TDH, NVIDIA PCX5900
  • HDD: WD Raptor WD360, SATA, 10000 rpm, 36Gb
  • Drivers: NVIDIA Forceware 77.72, Intel Chipset Utility 7.0.0.1025, DirectX 9.0c

Testbed 3

  • CPU: Intel Pentium 4 3.6 GHz (Prescott E0, 1 MB L2)
  • Chipset: Intel 955X
  • Motherboard: ASUS P5WD2 Premium, BIOS 0205 dated 04/22/2005
  • Memory: 2x512 MB Corsair XMS2-5400UL (DDR2-667 mode)
  • Video: Leadtek PX350 TDH, NVIDIA PCX5900
  • HDD: WD Raptor WD360, SATA, 10000 rpm, 36Gb
  • Drivers: NVIDIA Forceware 77.72, Intel Chipset Utility 7.0.0.1025, DirectX 9.0c

Testbed 4

  • CPU: Intel Pentium 4 3.6 GHz (Prescott E0, 1 MB L2)
  • Chipset: Intel 955X
  • Motherboard: Gigabyte 8I955X Royal, BIOS F7 dated 07/13/2005
  • Memory: 2x512 MB Corsair XMS2-5400UL (DDR2-667 mode)
  • Video: Leadtek PX350 TDH, NVIDIA PCX5900
  • HDD: WD Raptor WD360, SATA, 10000 rpm, 36Gb
  • Drivers: NVIDIA Forceware 77.72, Intel Chipset Utility 7.0.0.1025, DirectX 9.0c

Testbed 5

  • CPU: Intel Pentium 4 3.6 GHz (Prescott E0, 1 MB L2)
  • Chipset: Intel 955X
  • Motherboard: ASUS P5WD2 Premium, BIOS 0205 dated 04/22/2005
  • Memory: 2x512 MB Corsair XMS2-5400UL (DDR2-800 mode)
  • Video: Leadtek PX350 TDH, NVIDIA PCX5900
  • HDD: WD Raptor WD360, SATA, 10000 rpm, 36Gb
  • Drivers: NVIDIA Forceware 77.72, Intel Chipset Utility 7.0.0.1025, DirectX 9.0c

Testbed 6

  • CPU: Intel Pentium 4 3.6 GHz (Prescott E0, 1 MB L2)
  • Chipset: Intel 955X
  • Motherboard: Gigabyte 8I955X Royal, BIOS F7 dated 07/13/2005
  • Memory: 2x512 MB Corsair XMS2-5400UL (DDR2-800 mode)
  • Video: Leadtek PX350 TDH, NVIDIA PCX5900
  • HDD: WD Raptor WD360, SATA, 10000 rpm, 36Gb
  • Drivers: NVIDIA Forceware 77.72, Intel Chipset Utility 7.0.0.1025, DirectX 9.0c

Test Results

Performance tests

Pay your attention: we used 200 MHz FSB in all cases — in nominal DDR2-667 mode (Testbeds 1-4) as well as in overclocked DDR2-800 mode (Testbeds 5-6). The reason is as follows: none of the motherboards is capable of operating at 266 MHz FSB in these modes (DDR2-667/800), not only with these Corsair memory modules. Here emerges the reasonable question as to why test the memory in DDR2-800 mode? The answer to it is the fact that the modules under review are quite robust — due to significantly increased voltage (2.1V) as well as excellent timing characteristics of the memory chips (unfortunately we didn't manage to make out any meaningful designation to determine their manufacturer and technical characteristics).

In the first series of tests we used the timing scheme, set in BIOS Setup by default (Memory Timings: "by SPD"). As you can see, the motherboards that took part in our tests (Testbeds 1-4) obviously didn't agree on the default values — despite the correctly written SPD data exactly for this case (DDR2-667). Most of them set the 4-4-4 (-12/15) scheme, while ASUS P5WD2 Premium (Testbed 3) preferred more conservative 5-5-5-15 timings. Paradoxical, but true: this motherboard uses lower 4-4-4-15 timings for these modules in DDR2-800 mode, while Gigabyte 8I955X Royal on the same Intel 955X chipset, on the contrary, sets the slowest 5-6-6-18 scheme.

Parameter* Testbed 1 Testbed 2 Testbed 3 Testbed 4 Testbed 5** Testbed 6**
Timings
4-4-4-12
4-4-4-12
5-5-5-15
4-4-4-15
4-4-4-15
5-6-6-18
Average memory read bandwidth, MB/sec
4760
4764
5724
5550
5719
5707
Average memory write bandwidth, MB/sec
2075
2074
2410
2103
2218
2352
Max. memory read bandwidth, MB/sec
6161
6180
6543
6471
6546
6489
Max. memory write bandwidth, MB/sec
4266
4266
4314
4287
4314
4287
Minimum Pseudo-Random Access Latency, ns
60.1
59.9
49.8
54.1
48.9
48.9
Maximum Pseudo-Random Access Latency, ns
69.6
69.7
58.5
62.8
57.7
57.8
Minimum Random Access Latency***, ns
117.9
117.8
98.8
108.5
102.6
97.6
Maximum Random Access Latency***, ns
141.6
141.7
123.7
133.0
126.5
121.7
Minimum Pseudo-Random Access Latency, ns
(without hardware prefetch)
92.3
92.1
75.5
82.4
72.6
73.8
Maximum Pseudo-Random Access Latency, ns
(without hardware prefetch)
120.1
119.0
102.8
108.4
100.7
103.0
Minimum Random Access Latency***, ns
(without hardware prefetch)
118.5
118.3
98.9
109.0
102.8
97.9
Maximum Random Access Latency***, ns
(without hardware prefetch)
144.6
144.7
127.3
136.1
130.8
126.1

*the best results are marked in bold (all other things being equal)
**DDR2-800 mode
***16 MB block size

So, the best result among DDR2-667 tests is demonstrated (strange as it my seem - higher timings) by ASUS P5WD2 Premium (Testbed 3). It's followed by the second motherboard on Intel 955X — Gigabyte 8I955X Royal. Foxconn products on i945G/P chipsets (which are as slow compared to i955X as 915 chipsets compared to 925 chipsets) bring up the rear. In fact, Foxconn products have never been notable for high memory efficiency.

DDR2-800 mode is completely justified: the modules operate very well in this mode, demonstrating nearly identical data exchange rate with the CPU (it's not surprising — the FSB throughput in this case is just 6.4 GB/s, while the real DDR2-800 potential is twice as high) at somewhat (just a little) lower latencies. There is no absolute leader in this case — ASUS P5WD2 Premium (Testbed 5) still leads in the majority of aspects, while the best random access latency (both with and without hardware prefetch) is demonstrated by Gigabyte 8I955X Royal (Testbed 6).

Stability tests

Timing values, except for tCL, 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.

Minimal timings allowed by the memory modules under review depend on a given motherboard (to be more exact, on what maximum voltage the motherboard can apply to the memory modules) as well as on a memory operating mode (quite naturally). Foxconn products can apply only 1.9-1.95V to memory modules, thus they do not allow to squeeze maximum from the modules. But at the same time, the result is not that bad — minimal timings are 4-2-2 (we again omit tRAS, as it's ignored by these modules as well as by the majority of other modules). Nevertheless, ASUS and Gigabyte motherboards on the 955 chipset, capable of applying the necessary voltage (2.1V), allow the lowest timings (they cannot be any lower) — 3-2-2 — lower than 3-3-2(-8), guaranteed by the manufacturer for stable operation. Thus, Corsair modules break another timings record — the previous record, 3-2-2 in DDR2-533 mode, had been achieved by XMS2 PRO-5400C4.

Minimal timings in DDR2-800 mode are also very good — 4-3-2 on both motherboards from ASUS and Gigabyte (remember that nominal timings for "typical" DDR2 modules are the following: 3-3-3-9 for DDR2-400, 4-4-4-12 for DDR2-533, 5-5-5-15 for DDR2-667, they must be still higher for DDR2-800 — about 6-6-6-18).

Parameter* Testbed 1 Testbed 2 Testbed 3 Testbed 4 Testbed 5** Testbed 6**
Timings
4-2-2
(1.95V)
4-2-2
(1.9V)
3-2-2
(2.1V)
3-2-2
(2.1V)
4-3-2
(2.1V)
4-3-2
(2.1V)
Average memory read bandwidth, MB/sec
4813
4819
5719
5620
5780
5874
Average memory write bandwidth, MB/sec
2278
2286
2421
2421
2480
2426
Max. memory read bandwidth, MB/sec
6253
6264
6556
6492
6556
6522
Max. memory write bandwidth, MB/sec
4266
4266
4314
4287
4314
4287
Minimum Pseudo-Random Access Latency, ns
59.5
59.5
49.0
52.3
48.4
43.9
Maximum Pseudo-Random Access Latency, ns
70.2
69.1
58.6
61.0
57.2
51.0
Minimum Random Access Latency***, ns
112.0
112.2
98.7
102.0
96.8
83.4
Maximum Random Access Latency***, ns
135.6
135.6
123.1
126.3
121.2
119.1
Minimum Pseudo-Random Access Latency, ns
(without hardware prefetch)
91.5
91.6
75.8
80.6
72.8
68.5
Maximum Pseudo-Random Access Latency, ns
(without hardware prefetch)
116.7
117.7
102.5
107.2
100.4
87.8
Minimum Random Access Latency***, ns
(without hardware prefetch)
112.5
112.8
99.1
102.4
97.1
84.1
Maximum Random Access Latency***, ns
(without hardware prefetch)
137.4
138.6
126.9
130.0
125.0
105.2

*the best results are marked in bold (all other things being equal)
**DDR2-800 mode
***16 MB block size

What concerns performance of these modules under "extreme" conditions — ASUS P5WD2 Premium (Testbed 3) in DDR2-667 mode retains its leadership. But in case of DDR2-800 tests, we have a different leader almost in all aspects — Gigabyte 8I955X Royal (Testbed 6), which was outperformed in the previous series of tests, probably due to higher nominal timings 5-6-6-18.

Bottom line

The Corsair XMS2-5400UL modules are well compatible with the motherboards reviewed on Intel 945 and 955 chipsets, they offer good performance as well as huge overclocking potential both in terms of frequency and timings: the modules are stable in their standard DDR2-667 mode with 3-2-2 timings (the absolute record!) as well as in DDR2-800 mode with 4-3-2 timings (which is also an absolute record so far, as we have never carried out our tests in DDR2-800 mode before).

Corsair XMS2-5400UL modules are kindly provided
by Corsair Memory representative office in Russia

Dmitri Besedin (dmitri_b@ixbt.com)
September 8, 2005.



Write a comment below. No registration needed!


Article navigation:



blog comments powered by Disqus

  Most Popular Reviews More    RSS  

AMD Phenom II X4 955, Phenom II X4 960T, Phenom II X6 1075T, and Intel Pentium G2120, Core i3-3220, Core i5-3330 Processors

Comparing old, cheap solutions from AMD with new, budget offerings from Intel.
February 1, 2013 · Processor Roundups

Inno3D GeForce GTX 670 iChill, Inno3D GeForce GTX 660 Ti Graphics Cards

A couple of mid-range adapters with original cooling systems.
January 30, 2013 · Video cards: NVIDIA GPUs

Creative Sound Blaster X-Fi Surround 5.1

An external X-Fi solution in tests.
September 9, 2008 · Sound Cards

AMD FX-8350 Processor

The first worthwhile Piledriver CPU.
September 11, 2012 · Processors: AMD

Consumed Power, Energy Consumption: Ivy Bridge vs. Sandy Bridge

Trying out the new method.
September 18, 2012 · Processors: Intel
  Latest Reviews More    RSS  

i3DSpeed, September 2013

Retested all graphics cards with the new drivers.
Oct 18, 2013 · 3Digests

i3DSpeed, August 2013

Added new benchmarks: BioShock Infinite and Metro: Last Light.
Sep 06, 2013 · 3Digests

i3DSpeed, July 2013

Added the test results of NVIDIA GeForce GTX 760 and AMD Radeon HD 7730.
Aug 05, 2013 · 3Digests

Gainward GeForce GTX 650 Ti BOOST 2GB Golden Sample Graphics Card

An excellent hybrid of GeForce GTX 650 Ti and GeForce GTX 660.
Jun 24, 2013 · Video cards: NVIDIA GPUs

i3DSpeed, May 2013

Added the test results of NVIDIA GeForce GTX 770/780.
Jun 03, 2013 · 3Digests
  Latest News More    RSS  

Platform  ·  Video  ·  Multimedia  ·  Mobile  ·  Other  ||  About us & Privacy policy  ·  Twitter  ·  Facebook


Copyright © Byrds Research & Publishing, Ltd., 1997–2011. All rights reserved.