Testing methodology, overclocking
Comparison Memory |
|||||
|---|---|---|---|---|---|
ADATA XPG2133 16GB |
Patriot Viper 3 IEM 8GB |
Patriot Viper 3 8GB |
G.Skill TridentX 8GB |
Corsair Vengeance 8GB |
|
| Model | 2133XW8G10 |
PVI38G213C1K |
PV38G213C1K |
F3-2400C10-8GTX |
CMZ8GX3M2A1600C9 |
| Capacity | 16GB (2x8GB) |
8GB (2x4GB) |
8GB (2x4GB) |
8GB (2x4GB) |
8GB (2x4GB) |
| Speed | 2,133MHz DDR3 |
2,133MHz DDR3 |
2,133MHz DDR3 |
2,400MHz DDR3 |
1,600MHz DDR3 |
| Timings | 10-11-11-30-2T |
11-11-11-27-2T |
11-11-11-27-2T |
10-12-12-31-2T |
9-9-9-24-2T |
| Voltage | 1.65V |
1.50V |
1.50V |
1.65V |
1.50V |
| Price* | £180 |
£48 |
£48 |
£85 |
£40 |
| Cost per GB* | £11.25 |
£6 |
£6 |
£10.62 |
£5 |
| *Approximate, correct at time of writing | |||||
Test bench |
|||||||||
|---|---|---|---|---|---|---|---|---|---|
| CPU | Intel Core-i7 3770K Ivy Bridge | ||||||||
| Motherboard | Gigabyte Z77-UD5H | ||||||||
| Storage device | Samsung 830 Series 256GB SSD | ||||||||
| Graphics card | Intel integrated HD 4000 Graphics | ||||||||
| Power supply | Corsair AX750 | ||||||||
| Operating system | Windows 7 Ultimate SP1, 64-bit | ||||||||
Benchmarks |
|---|
| SiSoft SANDRA (download) |
| AIDA64 (download) |
| HEXUS.PiFast (download) |
| 3DMark 11 v1.0.3.0 (homepage) |
| 3DMark Vantage v1.2.0 (homepage) |
| Just Cause 2 (homepage) |
| Batman: Arkham City (homepage) |
| DiRT Showdown (homepage) |
Notes
We've placed the ADATA XPG Xtreme memory alongside four other modules tested on the latest Core i7-3770K chip sat on top of a Gigabyte Z77 motherboard, which has memory ratios all the way up to 32x - 3,200MHz, at a default 100MHz base clock. The XMP settings are applied correctly, meaning the 16GB kit runs at 10-11-11-30-2T timings for the get-go.
You'll see two standout bars in the ensuing graphs. The white bar denotes the ADATA XPG performance when run to specification, that is, dual-channel DDR3-2,133. Just for kicks and explained on the first page, a dark-grey bar indicates the performance when a solitary stick is in the system, which is more a test of the CPU's memory-controller than the quality of the memory.
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. We'd normally increase the voltage from 1.50V to 1.65V, but seeing as this is the standard voltage for the XPG set, we left it alone. The modules would scale down to 9-9-9-24 at the specified 2,133MHz, which is rather impressive, but, as you would expect, wouldn't run any higher.
We've also overclocked them by relaxing the timings to 11-13-13-35-2T and seeing just how much frequency we could eke out of them, all the while staying at the default 1.65V juice. We managed to run 2,200MHz just fine - though the modules actually scored a little lower in three of our most memory-intensive benchmarks - and the pair of sticks also booted into Windows and ran benchmarks at 2,400MHz, which is better than other 2,133MHz-rated kits we've looked at recently. Upping the frequency to 2,666MHz caused POST failure, however.
They're tuned, and tuned well, to perform at the 2,000-2,400MHz sweetspot, according to our testing. Unlike others, they do have some frequency headroom above their stated frequencies, especially if you sacrifice some latency.