The Basics:
The images below show the two 1024MB modules still resting inside the opened packaging. The blue heatspreader looks sharp, and the anodized Aluminum should do a good job at transferring heat from the chips to the air. Each module has two stickers affixed to one end... One details a variety of manufacturing data, while the other simply says "EVALUATION". This sticker makes you wonder if there is a special process to selecting memory for review, but with a few flicks of my finger nail those stickers quickly disappeared.
 
The below left image shows the two different sides of the modules, revealing that only the one side carries the pair of stickers. The below right image shows an edge view of the memory, attempting as best possible to show that the heatspreader is well affixed to the memory chips underneath, and that good contact appears to be made with the thermal tape.
 
Kingston does not publish data on what chips are used on their DDR modules, as they can change from lot to lot. All that they will confirm is that any modules used will meet or exceed their base specification.
Testing:
Testing of the Kingston HyperX 2GB PC3200 Dual Channel DDR Kit was conducted on a test system with the specifications listed below:
» AMD Athlon-64 3200+ Venice core processor
» Thermalright SI-120 CPU cooler
» ASUS A8N-E nForce4 Ultra motherboard
» 1x 200GB Seagate 7200.8 SATA hard drive
» 128MB Gigabyte Radeon X600XT PCIe graphics card
» Windows XP Professional, SP2 with all current patches
The first thing I wanted to investigate was the timings reported by the system's BIOS in order to clear up the confusion. With the BIOS set to "Auto", I checked the SPD data and saw that the timings were actually reported as 3-3-3-8. So, I didn't have CL2 or CL2.5, but CL3. This may fit with what I was told unofficially about the KHX3200AK2/2G kit at 2.6V, but would 2.65V make a difference? I knew SPD values wouldn't change at stock speed due to a voltage change, but I tried it anyway. I went to 2.7, 2.8, and 2.9V and nothing happened.
A second system, an Abit IS7-G with a Pentium-4 2.6C, was used as a second check of the timings, and it agreed with the 3-3-3-8. Testing on either system revealed that running with the timings as tight as 2-3-3-6 was acceptable, but 2-3-2-6 didn't seem too reliable at any voltage.
When I mentioned the timings discrepancy issue to Kingston I was informed that the chips are just "confirmed to run at those timings only if your motherboard and CPU can also." Somewhat of a noncommittal specification, but it is what it is, and I was able to exceed it with zero effort.
Testing of the Kingston HyperX 2GB PC3200 Dual Channel DDR Kit was conducted head to head with another 2GB of performance dual channel DDR, Corsair's 2048MB XMS PC3500LL Pro dual channel DDR Kit. The default timings of the Corsair kit are 2-3-2-6 at PC3500 speed, and even when slowed down to PC3200 speed, I found these to be the appropriate timings.
|
|
