Testing methodology
Comparison Memory |
||||||
|---|---|---|---|---|---|---|
G.Skill Ripjaws 2133 |
Corsair XMS-1333 |
Corsair Vengeance 1600 |
Corsair Vengeance 1866 |
Patriot Viper 3 2133 |
||
| Model | F3-17000CL11D-8GBXL |
CMX8GX3M2A1333C9 |
CMZ8GX3M2A1600C9B |
CMZ8GX3M2A1866C9B |
PVI38G213C1K |
|
| Capacity | 8GB (2x4GB) |
8GB (2x8GB) |
8GB (2x4GB) |
8GB (2x4GB) |
8GB (2x4GB) |
|
| Speed | 2,133MHz |
1,333MHz |
1,600MHz |
1,866MHz |
2,133MHz |
|
| Timings | 11-11-11-30-2T |
9-9-9-24-2T |
9-9-9-24-2T |
9-10-9-27-2T |
11-12-12-30-2T |
|
| Voltage | 1.50V |
1.50V |
1.50V |
1.50V |
1.50V |
|
| Price* | £40 |
£30 |
£30 |
£40 |
£45 |
|
| Cost per GB* | £5 |
£3.75 |
£3.75 |
£5 |
£5.63 |
|
| *Approximate, correct at time of writing | ||||||
Test bench |
|||||||||
|---|---|---|---|---|---|---|---|---|---|
| CPU | AMD A10-5800K APU | ||||||||
| Motherboard | Gigabyte F2-A85X-UP4 (F3G) | ||||||||
| Storage device | Samsung 830 Series 256GB SSD | ||||||||
| Graphics card | AMD HD 7660D integrated | ||||||||
| Power supply | Corsair AX750 | ||||||||
| Operating system | Windows 7 Ultimate SP1, 64-bit | ||||||||
Benchmarks |
|---|
| AIDA64 (download) |
| HEXUS.PiFast (download) |
| 3DMark 11 v1.0.3.0 (homepage) |
| 3DMark Vantage v1.2.0 (homepage) |
| Batman: Arkham City (homepage) |
| DiRT Showdown (homepage) |
Notes
Reiterating the points made on the previous page, we've historically tested DDR3 memory on an Intel Core i7-3770K platform. The chip's integrated graphics, HD 4000, don't exhibit a significant appreciation for high-speed RAM. AMD's Trinity A10-5800K, however, just loves faster memory for its HD 7660D IGP, which scales by up to 25 per cent in Batman: Arkham City when run with DDR3-2,133 memory instead of DDR3-1,600.
Overclocking
We run the overclocking test by inputting standard 9-9-9-24-2T timings and seeing just how well the test modules can scale. Voltage was increased from the default 1.50V to 1.60V. Our set managed to clock up to 2,000MHz at these settings - 2,133MHz would work, granted, but the PC would sporadically reboot.
Going the other way and loosening-off timings to 11-13-13-35-2T resulted in a top speed that fell short of 2,400MHz. We believe this has more to do with the frequency ceiling of the APU's memory-controller rather than intrinsic limit of the G.Skill modules themselves. Testing this assertion by plonking them into our regular Core i7-3770K Ivy Bridge build resulted in stable operation at 2,400MHz. Performance scaling on this platform isn't a patch on AMD's, though.
