All about Memory
 Latest Memory articles
All Memory articles

Newsletters


  • Ask your question about IT issues
  • Post

Partners

The Games selection

adventure : Ray Adventure game, South Park style. Pick the way the story goes by picking an answer among those offered.
violent : Interactive Buddy Unwind on your interactive buddy: Do anything you want to him, it will earn you money, and you can buy other stuff to torture him with.
Ads

Sponsored links

Boosting Memory Clock Speed: 23% Performance Gains

Previous Next
8:29 AM - 02/27/2006 by The editorial team

On the other hand, boosting memory clock rate by itself can produce significant performance gains. As before, we used the outstanding Corsair CM2X512-8000UL Corsair memory modules for the memory clock speed and latency analysis below.

The clock rates for DDR2-400 up to DDR2-800 were analyzed in tandem with FSB speeds of 800 and 1066. The reference point for our analysis is DDR2-533 RAM, because this is the type most frequently-installed in commercial PCs. That's because DDR2-400 is neither very prevalent nor very widely used.

Speed analysis with an FSB speed of 800:

Module Clock Speed (CL 5.0-5-5-15) FSB800 (200 MHz)
FSB800 (200 MHz) Read Write Latency
DDR2-800 (400 MHz) 6189 MB/sec 2142 MB/sec 82.9 ns
DDR2-667 (333 MHz) 6105 MB/sec 1985 MB/sec 88.0 ns
DDR2-600 (300 MHz) 6052 MB/sec 1890 MB/sec 88.1 ns
DDR2-533 (266 MHz) 6024 MB/sec 1743 MB/sec 89.5 ns
DDR2-400 (200 MHz) 5464 MB/sec 1389 MB/sec 111.4 ns
FSB800 (200 MHz) Read Write Latency
DDR2-800 (400 MHz) 102.7% 122.9% 108.0%
DDR2-667 (333 MHz) 101.3% 113.9% 101.7%
DDR2-600 (300 MHz) 100.5% 108.4% 101.6%
DDR2-533 (266 MHz) 100.0% 100.0% 100.0%
DDR2-400 (200 MHz) 90.7% 79.7% 80.3%

This table shows clearly what we measured in our labs at THG: if a DDR2-800 memory module is used with an FSB speed of 800, it can boost memory read access by 2.7% and write access by 22.9%. The latency timing in our measurement improves by about 8%. From these results we draw the following conclusion: Those who don't boost FSB speeds won't realize any significant performance improvements simply from faster memory clock speeds.

A speed analysis with an FSB speed of 1066:

Module Clock Speed (CL 5.0-5-5-15) FSB1066 (266 MHz)
FSB800 (200 MHz) Read Write Latency
DDR2-888 (444 MHz) 7819 MB/sec 2554 MB/sec 73.9 ns
DDR2-800 (400 MHz) 7748 MB/sec 2431 MB/sec 75.6 ns
DDR2-711 (355 MHz) 7673 MB/sec 2258 MB/sec 76.3 ns
DDR2-667 (333 MHz) 7310 MB/sec 2110 MB/sec 83.6 ns
DDR2-533 (266 MHz) 7066 MB/sec 1822 MB/sec 86.3 ns
DDR2-400 (200 MHz) 5732 MB/sec 1458 MB/sec 105.2 ns
FSB800 (200 MHz) Read Write Latency
DDR2-888 (444 MHz) 110.7% 140.2% 116.8%
DDR2-800 (400 MHz) 109.7% 133.4% 114.2%
DDR2-711 (355 MHz) 108.6% 123.9% 113.1%
DDR2-667 (333 MHz) 103.5% 115.8% 103.2%
DDR2-533 (266 MHz) 100.0% 100.0% 100.0%
DDR2-400 (200 MHz) 81.1% 80.0% 82.0%

This table reflects our test series and documents some impressive results. With an FSB speed of 1066 (266 MHz), a switch to DDR2-888 finally starts to pay off. Read access times then improve by 10.7% and write access times by 40.2%. At 16.2%, the speed boost for latency is double that for an FSB 800 system. This doesn't demonstrate anything startling, however: the bottleneck in a DDR2 Intel platform has been and remains the front side bus (FSB).

Talkback
Comments are closed on this page.

Sponsored links