Testing
I did what any geek would do when faced with a Booster X3 for testing: I bust out my multimeters and went crazy, poking them anywhere and everywhere in the wish to find the One True PSU Truth. In reality, I simply watched rail outputs and from-the-mains power draw in a couple of PC configs, with and without the X3, to see if it was helping.The PC configs were as follows, testing both SLI and Crossfire using a somewhat on-the-limit supply, and also something a bit more meaty:
Hardware
|
||
|---|---|---|
| Power Supply | FSP Group Booster X3 Tagan TG420-U02 i-Xeye Akasa AK-P550FH |
|
| Processor(s) | AMD Athlon 64 FX-62 2.8GHz, dual-core, 1MiB L2 per core |
Intel Core 2 Duo E6700 2.66GHz, dual-core, 4MiB L2 shared |
| Mainboard(s) | ASUS M2N32-SLI Deluxe NVIDIA nForce 590 SLI |
Intel D975XBK Bad Axe Intel i975 |
| Memory | 2 x 512MiB DDR2-800 | 2 x 1GiB DDR2-800 2 x 512MiB DDR2-800 |
| Memory Timings | 4-4-4-12-1T @ 400MHz | 4-4-4-12-2T @ 333MHz |
| BIOS Version | 0017 | 1304 |
| Disk Drive | 160GB Seagate Barracuda 7200.9 SATA | |
| Graphics Card(s) | 2 x BFG Tech GeForce 7900 GTX OC™ (HEXUS.link) NVIDIA SLI |
ATI Radeon X1900 Crossfire Edition (625/725) (HEXUS.link) ATI Radeon X1900 XTX (650/775) ATI Crossfire |
Info
So each setup was run with both Tagan and Akasa in the multi-GPU setup detailed, both with and without the Booster X3. The appropriate proddage let us check rail outputs and load condition from the mains, to see whether the X3 helped in any way.Some overclocking testing was done too, after that, to check whether the X3 running the graphics hardware also allowed for better overclocks. The Booster X3 gives you around 150W (~12.5A) per card, should you need to check if it's up to running your particular configuration.
First things first: the Tagan is actually enough -- even in the FX-62 box as configured -- to run both multi-GPU setups. There's 22A available on +12V and while it's pushing it with Radeon X1900 Crossfire, as long as you don't go mad with extra peripherals it'll do while you ponder an upgrade.
The Akasa is certainly capable, providing split +12V and enough for any single CPU Crossfire or SLI config today, without going nuts with overclocking on both CPU and GPU (which chew +12V a bit harder).
So here's what happened when we let the Booster X3 power the graphics hardware.
Voltage Results
|
||
|---|---|---|
| Measurement | FX-62 SLI System | E6700 Crossfire System |
| +12V on Tagan No Booster X3 |
+11.72 | +11.93 |
| +12V on Tagan Booster X3 |
+12.01 (+0.29V improvement) | +12.07 (+0.21V improvement) |
| +12V on Akasa No Booster X3 |
+12.00 | +12.03 |
| +12V on Akasa Booster X3 |
+12.08 (+0.08V improvement) | +12.08 (+0.05V improvement) |
Both PSUs sit higher than +12.00V, even under load, in the test systems. Load condition was 3DMark06 looped at default settings.
So the Booster X3 does stop +12V drooping under load, but not by much. A modern PSU core is somewhat designed to keep its end up under such conditions. And given that the Tagan TG420 is on the edge with the Crossfire system, simply because of the +12V pull from the GPUs, the Booster X3 would do nicely to enable us to run Crossfire if we had something less, where we wouldn't have been able to before.
So no problems running somewhat high-end multi-GPU graphics, making it worth investigating where the overclocks stop with and without it, to see if it helps there.
Overclocking Results
|
||
|---|---|---|
| Measurement | FX-62 SLI System (Tagan) | E6700 Crossfire System (Akasa) |
| Max CPU overclock No Booster X3 No VGA overclock |
~2920MHz | ~3280MHz |
| Max CPU overclock Booster X3 No VGA overclock |
~3008MHz (+88MHz improvement) | ~3333MHz (+53MHz improvement) |
|
||
|---|---|---|
| Measurement | FX-62 SLI System (Tagan) | E6700 Crossfire System (Akasa) |
| Max VGA overclock No Booster X3 No CPU overclock |
691/833 | 666/818 |
| Max VGA overclock Booster X3 No CPU overclock |
713/833 (+22/0MHz improvement) | 680/838 (+12/20MHz improvement) |
| Max VGA overclock Booster X3 Max CPU overclock |
713/833 (+22/0MHz improvement) | 680/838 (+12/20MHz improvement) |
The results are simple to interpret. When the Booster X3 is taking the graphics card power load, it has the potential to let you push your overclocks higher on CPU (and likely the rest of the system) and your GPUs, even in multi-GPU mode, indicating that it's doing a good job of supplying the output power the graphics hardware is asking for.
The Crossfire system, powered by a modern capable PSU like the Akasa, saw benefit from the Booster X3 in allowing higher overclocks, simply because the system has a dedicated supply for the high-end graphics hardware.
The Booster X3 therefore works as we expected, allowing us more performance from the use of a dedicated tool for that particular job.
