Is High-Performance Memory Necessary for State-of-the-art Chipsets for Intel Processors?

An occasion for the analysis described in this article was given by the announcement of superfast Corsair DDR2-667 memory, which was made simultaneously with the announcement of NVIDIA nForce4 SLI Intel Edition chipset. In fact, it was a joint announcement in a sense: DDR2-667 was officially supported only by the new NVIDIA chipset at that time and nForce4 SLI IE badly needed such memory with minimum timings, so that this support was not just in name only. Indeed, an attempt to use DDR2-667 even with the best timings (4-4-4-12) in our review of the first NVIDIA product for Intel processors just resulted in performance drop. Let's see what the production-line motherboard on this chipset can do with a pair of Corsair CM2X512A-5400UL modules, which claim to support fantastic 3-2-2-8-1T timings at 675 MHz (but at the voltage raised by 0.3V they are guaranteed to operate with nForce4 SLI IE only).

We got the memory modules only after the chipset announcement. But that's even good, as the tests were carried out already with a production-line motherboard and the results obtained are not preliminary. There is no need to introduce the modules — it's a typical pair of Corsair TwinX XMS2. But unlike the 5400C4 that we had used (CAS Latency 4 at 675 MHz), this one is marked 5400UL (Ultra Low at 675 MHz) and the holographic label lacks the SPD timing values (4-4-4-12 for 675 MHz). Of course, no special technologies are used to manufacture these memory modules, impressive characteristics are obtained only because the chips are thoroughly sorted (and of course due to the increased voltage).

The objective of this article is to test the high-performance DDR2-667 in action. We uncovered the potential of dual-channel DDR2-533 not long ago. Expectedly, the mandatory condition for this was the appearance of processors with 1066(266) MHz FSB. Now NVIDIA offers DDR2-667 support (Intel promises it in the nearest future). It would be logical to expect that this memory type will be required only for processors with 1333(333) MHz FSB. Meanwhile, we don't know about Intel plans to manufacture such models. Moreover, there exist only two Pentium 4 Extreme Edition processors with 1066MHz FSB and one Pentium Extreme Edition (all of them cost about $1000). It turns out that DDR2-667 support in the current chipsets is useful only to memory manufacturers, who need to work through the manufacturing process and skim the cream off those who want to buy the most progressive components. The test results of nForce4 SLI IE backed up our assumption, but the evident "irons on the hands" of DDR2-667 modules were high latencies. Let's see what this memory can do with the minimum possible timings.

Performance tests

Testbed configuration:

CPU: 3.73GHz Intel Pentium 4 Extreme Edition, Socket 775
Motherboards:
- MSI P4N Diamond (BIOS 1.0BK) on NVIDIA nForce4 SLI Intel Edition
- Gigabyte 8AENXP-D (BIOS F3) on Intel 925XE
- Gigabyte 8I945P-G (BIOS F1) on Intel 945P
Memory: 2x512 MB PC2-5400 (DDR2-675) DDR2 SDRAM DIMM Corsair
Video card: [PCIEx16] ATI Radeon X800 XT 256 MB
HDD: Seagate Barracuda 7200.7 (SATA), 7200 rpm

Software:

OS and drivers:
- Windows XP Professional SP2
- DirectX 9.0c
- NVIDIA nForce unified drivers package 7.02
- Intel chipset drivers 7.0.0.1019
- ATI Catalyst 5.2
Test applications:
- RightMark Memory Analyzer 3.47
- 7-Zip 4.10b
- WinRAR 3.41
- DivX 5.2.1 Pro codec
- XviD 1.0.2 (29.08.2004) codec
- SPECviewperf 8.01
- Doom 3 (v1.0.1282)
- FarCry (v1.1.3.1337)
- Unreal Tournament 2004 (v3339)

Test results

We've taken the same chipsets as in the nForce4 SLI IE review, but now the NVIDIA product is represented by a production-line (very interesting) motherboard from MSI and the i925XE is still represented by the best model based on this chipset — Gigabyte 8AENXP-D. Besides, most motherboard manufacturers have already prepared their products based on Intel 945-series chipsets, though this series is not yet announced. Today we have an opportunity to test Gigabyte on i945P. This motherboard uses the first (probably, pre-release) BIOS version, so you should be careful to rely on these test results. But still it's interesting to have a look what the new series of Intel chipsets has in store for us (we shall acquaint you with the details in due time).

Corsair modules under review (with reduced latencies) really manage to operate on i925XE and nForce4 SLI IE motherboards with 3-2-2-8(-1T) timings both in DDR2-533 as well as in DDR2-667 modes (the latter mode is certainly available only in the NVIDIA chipset). However, Gigabyte 8I945P-G coped with 667 MHz memory only with 4-3-2-12 timings. But at 533 MHz it actually reached only T_RAS=12 instead of T_RAS=8 set in BIOS (we have already come across such behaviour of memory modules). Of course, such situation heavily undermines our interest in testing i945P, but let's do it anyway.

Let's start with memory performance tests in RightMark Memory Analyzer, a test package developed by our programmers. You may remember that the engineering sample of a motherboard on nForce4 SLI IE was strongly outperformed by i925XE in real and maximum memory read rates. NVIDIA fans may relax now: the production-line motherboard did not confirm this tendency and it even slightly outperformed the i925XE representative. Note that nForce4 demonstrates practically the same performance with 2xDDR2-533 and 2xDDR2-667. It matches the performance of i925XE+2xDDR2-533 and its maximum value is actually limited by the FSB throughput (in fact, it's even a tad higher (?)). In case of single-channel mode, the read rate of nForce4 SLI IE is predictably limited by the memory rate (FSB throughput is higher), you can evaluate the lag from the dual-channel mode on your own. Note that single-channel modes are not very interesting from a practical point of view, but as our article is devoted mostly to testing the theory, we have carried out a full cycle of tests in these modes for the NVIDIA's chipset.

The maximum write rate in all the chipsets [in dual-channel mode] is limited by the CPU architecture; there is nothing interesting here (the limiting factor of the single-channel mode is again the memory module bandwidth). But the real write rate of nForce4 SLI IE turned out slightly higher than in Intel chipsets, so NVIDIA can celebrate victory (the reference board was on a par with i925XE): its performance with DDR2-667 is higher by ~23%, with DDR2-533 — by ~17%.

Last time we had a draw in memory [pseudo random] read latency, but now the NVIDIA chipset shoots forward (approximately by 2—6%), it outperforms i925XE not only in dual-channel modes, but also with single-channel DDR2-667. It's only slightly outperformed with single-channel DDR2-533.

The interesting effect demonstrated during previous tests, when nForce4 SLI IE outperformed the Intel chipset in random memory read latency (being on a par with it in pseudo-random memory read latency), can be now seen only in a single-channel mode with DDR2-533.

Let's draw a preliminary conclusion based on low-level memory tests: NVIDIA nForce4 SLI IE is actually faster than i925XE; DDR2-667 outperforms DDR2-533 a little at the same low timings; we are inclined to take i945P results as preliminary, so we are not going to rely on them in future (what really kills our interest is increased timings for 667MHz memory, which nullify potential advantages of the increased bandwidth). Now, let's have a look at real applications and evaluate the difference between system results for regular users.

The difference between Intel and NVIDIA chipsets in archivers is totally contemptible (about one percent to the advantage of the latter) and is quite comparable with the natural spread in test results. Trying to take this difference seriously, the i925XE goes on a par with nForce4+1xDDR2-533, while a motherboard on NVIDIA's chipset with single-channel DDR2-667 is insignificantly faster.

As archivers didn't show any differences, we can safely skip tests without 3D graphics, because the other applications that we use are less dependent on memory performance. Indeed, as you can see in this case, video encoding results are the same for all contenders. And compiling a PDF file in Adobe Distiller provides a significant chaotic spread in results (due to continuous access to HDD).

SPECviewperf (a benchmark for professional OpenGL application performance) is very useful because its tests react differently to the performance of system components, so that you can almost always notice interesting patterns. Today, three tests turn out critical to the memory bandwidth, which leads to the victory of nForce4 SLI IE with dual-channel DDR2-667. In other cases the chipsets are either on a par (3 tests), or outperform Intel chipsets (including i945P, 2 tests). nForce4 works best in dual-channel modes, 1xDDR2-667 is a tad slower, while single-channel DDR2-533 always lags noticeably behind.

Here is the situation in games: Doom 3 and FarCry in low modes prefer nForce4 SLI IE (a couple percents), three best nForce4 modes are on a par again. nForce4+1xDDR2-533 is on a par with i925XE. But at higher resolutions (and consequently complex graphics quality) both Intel chipsets suddenly shoot forward, retaining their advantage over NVIDIA even under the heaviest conditions. As memory speed and other factors have no effect on the performance, we can only assume that manufacturers of these chipsets (or these motherboard models) optimize PCIEx16 operations on the driver level (or on the level of hidden BIOS parameters) in a different way.

However, Unreal Tournament 2004 does not show this effect, all contenders are very close to each other here, but the general alignment of forces remains.

Conclusions

I repeat that today's conclusions about the advantages of DDR2-667 over DDR2-533 can be applied only to processors with 1066MHz FSB. This issue is not relevant for the majority of simple Pentium 4 models (and obviously the future Pentium D), DDR2-533 with good timings will be more than enough for them.

Going back to the results of today's tests, I want to note the following points:

On the whole, nForce4 SLI IE turns out faster than i925XE, though this difference is almost impossible to make out in practice. The i945P model that we tested has demonstrated relatively low performance, but we are not inclined to take it for the final result: let's wait for the official announcement of the chipset, probably new drivers, and certainly new BIOS versions.
The chipsets are no slower with DR2-667 memory than with DDR2-533 at the same timings, but no faster either (under dual channel conditions) — fractions of a single percent are insignificant. Considering that the overwhelming majority of DDR2-533 memory offers 4-4-4 timings, there is absolutely no sense in striving for DDR2-667 with extra low latencies. But we render homage to Corsair diligence and capacities — its CM2X512A-5400UL modules will long be famous as the fastest memory available.

Sergei Pikalov (peek@ixbt.com)
May 17, 2005.

Write a comment below. No registration needed!