Test RigsTest Methodologies: As indicated in the introduction each kit will be run at tightest latencies possible at 800MHz, 1000MHz and then overclocked. While I could have run the memory in certain circumstances at a slightly higher speeds as stated in the at some point increasing the speed a few MHz has little no discernable affect on performance. Both kits topped out in stability running CL3 @ 800MHz and then CL4 at 1000MHz, beyond these speeds SPD settings were more stable. While this may be indicative of non-traditional SPD programming I compared the parameters above to SPD settings at 800MHz and then 1000MHz to determine if that were the case.
Maximum performance at 800MHz
Maximum performance at 1000MHz (default speed)
Maximum performance at 1066MHz (Intel FSB speed)
Maximum performance at maximum speed (Max Performance most often = SPD)
In every instance the memory was run at 2.25V ~ 2.75V to compensate for under volting. Testing with a digital multimeter did show some fluctuation on both platforms fortunately the range above compensated for this and was only used if absolutely necessary. To simplify matters and provide as much data as possible each performance chart includes overclocking SPD results as well.
Intel System:
AMD System:
|
CPUs | 1. Intel Conroe 6400 (8x266FSB)Socket-775 2. AMD Opteron 1212 (10x200FSB) |
Mainboards | 1. Asus P5W DH Deluxe/WiFi (BIOS 1901) 2. Asus M2N32-SLI Deluxe (BIOS 903) | |
Memory | 1. Super Talent : DDR2-1000 / PC2-8000, 5-5-5-15 Latencies (CAS, tRCD, tRP, tRAS), Voltage 2.2V 2. Crucial Ballistix : DDR2-1000 / PC2-8000, 5-5-5-15 (CAS, tRCD, tRP, tRAS), Voltage 2.2V |
Graphics | 1. Leadtek 7950GX2 2. BFG 8800GTX |
Power Supply | 1. Silverstone ZEUS ST85ZF 2. PCPower&Cooling Silencer 750EPS12V |
Cooling | 1. Swiftech Apex Ultra 2. Danger Den NVIDIA kit 4101 |
Storage | 1. Maxtor 300GB SATA 2. 2x Seagate Barracuda 80GB SATA Perpendicular |
Optical | 1. Plextor PX755-SA DVD/RW 2. NEC DVD/RW |
Operating System | Windows XP |
To show the difference between running our test subjects on SPD as opposed to manual timings I've provided the thumbnails below. Each series only shows 800MHz ~ 1000MHz and not overclocked which would simply take up too much space. Our set represents Super Talent running on the E6400 / Asus P5W DH (1701 BIOS) under SPD then manual. Manual timings will represent the tightest timings possible and those found on the proceeding performance charts. Thumbnails reflect CPU-speed (FSB), Vcore, temp, motherboard divider, memory-speed, timings and module-ID.
Super Talent / E6400 - P5W DH Intel 975X chipset
800MHz @ SPD ||| 800MHz @ CL3 ||| 1000MHz @ SPD ||| 1000MHz @ CL4
Our next series represent Crucial Ballistix running between 800MHz ~ 100MHz only on the Opteron 1212 / Asus M2N32-SLI (BIOS 903) under SPD then under manual timings. Manual timings are the tightest timings possible and those found on the proceeding performance charts. Thumbnails reflect CPU-speed (FSB), Vcore, temp, motherboard divider, memory-speed, timings and module-ID.
Crucial Ballistix / Opteron 1212 - Asus M2n32-SLI NVIDIA 590 chipset
800MHz @ SPD ||| 800MHz @ CL3 ||| 1000MHz @ SPD ||| 1000MHz @ CL4
Benchamrks:
Once each setting was Orthos stable it was tested running the following benchmarks:
Sisoft S.A.N.D.R.A. XI
Lavalys Everest Ultimate
Maxon Cinebench 9.5
MadOnion/Futuremark 3D Mark 2001SE
Futuremark 3D Mark 2006
Futuremark PC Mark 2005
Superpi 1.5 Xtremesystems edition (1M/2M).
Game 1 : DOOM3 Timedemo (Minimum settings)
Game 2 : Monolith F.E.A.R. Multiplayer demo, built-in benchmark LOW settings.
As we get into the specific result charts more info will be provided. First up Intel tests ->