Memory performance
Memory subsystem performance doesn't always track with real-world applications, but it's a good place to start with integrated graphics chipsets that cannibalize a portion of system memory and therefore bandwidth.

With a higher CAS latency than the other systems, it's no surprise to see the GeForce 9300 fall to the back of the field in our memory subsystem tests. Then again, we wouldn't expect going up just one tick in that memory timing setting to produce the much higher total latency we're seeing with the 9300. Indeed, Nvidia expects memory performance to improve with the next GeForce 9300 driver revision. They say the current driver doesn't enable an Advance Path feature that supposedly provides a lower latency data path for memory reads and writes. Looking at the chipset's pokey memory access latency, there's clearly room for a lot of improvement.

Some motherboards don't cope well when all their DIMM slots are populated, so we threw a couple of extra memory modules into the systems for some four-DIMM testing.

Loading the GeForce 9300 up with memory modules doesn't change the picture much. Nvidia's latest MCP still lags behind the G45 Express, which puts it well off the pace set by the 780G and 790GX.

You don't see any four-DIMM results for the GeForce 8300 because the particular Asus board we used for testing has problems running with four memory modules. We've had four-DIMM configs working on one other GeForce 8300 board, though, so the issue doesn't appear to be chipset-related.

The following latency graphs are a little indulgent, so I won't be offended if you skip them. They show access latencies across multiple block and step sizes, painting a fuller picture of memory controller performance with each chipset. I've arranged the graphs in order of highest latency to lowest. Yellow represents L1 cache, light orange is L2, and dark orange is main memory.

These results are from our two-DIMM configs and nicely illustrate the GeForce 9300's higher access latencies.