Mac Pro Westmere Photoshop Performance: diglloydSpeed1

See test machine configuration.

The diglloydSpeed1 benchmark measures core Photoshop operations: sharpening, resizing, inverting, etc. It is an in-memory benchmark, so no memory limitations or scratch disk speed come into play.

As the graph shows, OpenGL is a hindrance, even with the bad-ass Radeon 5870 video card in the 2010 3.33GHz 6-core machine. Don’t waste your money on it for regular Photoshop work. Since this test encompasses operations that dominate my use of Photoshop, OpenGL is a clear loser. But your use might differ, so test your own usage scenarios.

Clock speed “rules” for this test, the number of CPU cores help only a little.

OpenGL does not contribute except to degrade performance.

The 2010 Mac Pro 3.33GHz hexacore performs best here, but even the iMac does very nicely. The differences are modest between the machines.

As seen with the diglloydSpeed1 benchmark, Photoshop CS5 gains nothing from a 12-core machine, even degrading in performance.

RAW-file Processing Speed

RAW-file Processing Speed

CPU core usage

See Monitoring system performance with Activity Monitor.

This is typical Photoshop—modest use of CPU cores, with about 60% of the computing power squandered. This graph is from 10 iterations.

CPU core usage with Photoshop CS5, diglloydSpeed1 10 iterations
2010 Mac Pro 6-core @ 3.33GHz