Multitasking
Let's now turn our attention to multitasking, where more than one application is run at one time.| DivX 6.8.3 + LAME encoding, enhanced multithreading on | ||||
|---|---|---|---|---|
| AMD Athlon X2 7750 BE | AMD Athlon X2 7850 BE | Intel Pentium Dual-Core E5200 | AMD Athlon X2 6000+ | AMD Athlon II X2 250 |
| 410 | 398 | 390 | 444 | 359 |
Here's the graph you've seen before, detailing the DivX-encoding results. Nothing new here, as the X2 250 leads the way.
| DivX encoding with QuickTime 1080p clip running | ||||
|---|---|---|---|---|
| AMD Athlon X2 7750 BE | AMD Athlon X2 7850 BE | Intel Pentium Dual-Core E5200 | AMD Athlon X2 6000+ | AMD Athlon II X2 250 |
| 572 | 550 | 554 | 634 | 476 |
But now we've added a 1080p QuickTime clip into the fray, being rendered at the same time. What we're seeing here is how much the DivX encoding time is being compromised by the additional computational load presented by running the high-resolution clip on a looping basis.
All dual-core chips suffer, as shown by the increase, but the X2 250's hurt least of all, comparatively speaking.
| DivX encoding with ETQW 1680x1050 High-quality running | ||||
|---|---|---|---|---|
| AMD Athlon X2 7750 BE | AMD Athlon X2 7850 BE | Intel Pentium Dual-Core E5200 | AMD Athlon X2 6000+ | AMD Athlon II X2 250 |
| 682 | 670 | 662 | 732 | 643 |
Here's the same test but with the high-resolution Enemy Territory: Quake Wars running, looped, in the background. The DivX times go up to a greater degree this time, and all mid-range CPUs suffer more or less the same, with encoding taking over 50 per cent longer in most cases.
| ETQW 1680x1050 High-quality with DivX encoding running | ||||
|---|---|---|---|---|
| AMD Athlon X2 7750 BE | AMD Athlon X2 7850 BE | Intel Pentium Dual-Core E5200 | AMD Athlon X2 6000+ | AMD Athlon II X2 250 |
| 35.12 | 36 | 42.93 | 39.83 | 42.87 |
And, now, the corresponding average FPS for the game. Contradicting the numbers, the two Athlon 7xxx series chips, with L3 cache, appear to be just very slightly smoother than the X2 250. The empirical evidence would suggest that the extra L3 cache helps.