Well, this time we were lucky to get a memory that "violate" the rules of JEDEC. A-Data branded memory. Maybe you have not how familiar with the word. not like Kingston or Corsair. But that A-Data is in her native Taiwan, is the most famous brand and a leading memory. At the time this article was made already many outstanding memory. Starting from the type of DDR or SDR memory is registered, even in Taiwan from A-Data DDRII had been circulated.
Memory that will go into the test this time is a memory PC 4000 or DDR 500. This memory is called A-Data Vitesta DDR 500. Please note only DDR 500 which are named by the A-Data Vitesta. The first time we saw him, the impression of "frightening" is very visible in this memory. With red colored PCB and in the dressing by heat spreder which is also colored red to make it look classy and memory shows that memory is intended for high-end market.
Large memory capacity A-Data that we can this time is 256 MB X 2. We get 2 separate memory chip is not in one package. We got information that the chipset used in this memory is Hynix 4.3 ns running with 3-4-4-8 timings in the default condition. This chip is also commonly used in other memory manufacturers who labeled DDR 500.
Because we can not wait to try to do some tests to find out how fast this memory to work well in default and overclocked conditions. Proseor we use this time is AMD Barton. Why do we use AMD?, because we want to know the memory timings are derived from this memory when running on AMD platforms. And we also wanted to prove whether people said is true if memory is above the PC 3200 is very difficult and almost impossible to walk the tight timings as well as CL CL2 2.5. We also want to prove is there any difference when using the memory timings were tight with bertiming loose with the same FSB.
test platform used is
Processor: Barton 2500 +
Motherboard: DFI Ultra Infinity bios DFI-Street 4 / 29
VGA: ATI Radeon 9000pro 355/318
HDD: 80 GB
PSU: Thermaltake 480 watt black housing
OS: Windows XP Professional SP1
And as a comparison we also tested the corsairs memory TwinX PC 3200 LLPT rev 1.2. Why do we choose to do comparisons against the memory is not equivalent to A-Data PC4000. Because as we said above we want to know whether there are differences in the timing tight with loose timings at the same FSB. As we know the average memory running at FSB 200 MHz (PC 3200) in default circumstances have strict timing.
Test that we use this time is the benchmark test bandwidth and 3D rendering. Below are various programs that we use to test the A-Data memory:
- PC Mark 2004
- SiSoft Sandra 2004
Immediately, we put A-Data memory is in our DFI mobo. Once ignited, the brain-tweaking BIOSpun done. After we saw was correct according to the estimate and described labeled on the heatspreader that the memory is running with 3-4-4-8 timings at the default condition.
In this test done with a clock speed different. With the aim to find out seberepa large increase in the A-Data DDR 500 which occurred in the conditions of default timing. As a comparison test of Corsair memory in the default timing conditions. To clockspeed is set.
4 clock speed that we run in the test are:
- Barton 2500 + (11 * 166) @ 1833 MHz
- Barton 3200 + (10 * 220) @ 2200 MHz
- Barton 3 400 + (10 * 240) @ 2400 MHz
- Barton 3 600 + (10 * 250) @ 2500 MHz
Test 1: PC Mark 2004
From the above results we can see that the timing is very influential on benchmark results. Although run by the FSB and memory clock speed of the same A-Data with default timings 3-4-4-8 unable to pursue Corsair TwinX 3200 LLPT running with default timings 2-3-2-6.
Test 2: SiSoft Sandra
This benchmark we are doing to find out how much bandwidth is obtained from A-Data if FSBnya raised. In this test we also make comparisons with the Corsair TwinX 3200 LLPT.