You also need to understand subtle differences in how each SSD manufacturer tunes their GC algorithms, in some cases how the same manufacturer tunes them for different markets.
For example, the Micron-based consumer SSDs (e.g. Crucial M500) are notorious for having significant performance drop-offs during GC. Most of the time these SSDs are tuned to do GC during "idle" periods, with the assumption being that consumer use cases are mostly idle anyway. But if GC triggers when the device is active watch out - transfer rates collapse to the floor until it gets all done.
Most "Enterprise" SSDs (e.g., Samsung PM853T) tune their GC to be rather continuous and done in small steps. This causes the SSD to deliver slightly lower peak performance and consumes more power, but also causes the SSD to deliver consistent performance without the big dropouts seen in consumer SSDs. It also handles the continuous usage of Enterprise workloads better - it does not depend upon the SSD ever experiencing idle periods to work.
In fact, sometimes manufacturers will deliver substantially identical SSDs with different GC algorithms into different markets. For example, Samsung delivers SSDs into the Enterprise, prosumer and consumer markets that are almost identical, except that the Enterprise drive in the series has GC tuned for consistency at the expense of peak and the prosumer/consumer version may be tuned for Peak over consistency (along with differences in case materials and over-provisioning levels, of course).
Almost all recent SSDs will have GC. How effective it is and what side-effects it has differ widely. YMMV.
