I'd charge them all in parallel, up to, I dunno, I guess 4.2v if you want to test them fully. I'd use... hmm, probably something like a few surplus server power supplies or old PC PSUs with the 5v rails ganged together. It'll take forever, so, keep an eye on them, touch them ones in a while to test for heat, and manually disconnect the power before they rise above 4.2v. Maybe use a stove/oven/dryer element to limit the current to something the PSUs can handle without crapping their pants. Or if you like spending money, I dunno, maybe this for $100:
15 Volt Variable Output Switching DC Power Supply 25 Amps . 3-15v, up to 25A.
Then I'd string them all up in series, and discharge them across, I dunno, a few junkyard oven elements or a water heater element (perhaps in a convenient water tank). Enough so that it takes probably, oh, an hour or 2. That'll roughly mimic driving. While discharging, every few minutes I'd be hopping cell to cell, testing the voltages and seeing if any of them are suspiciously lower than their peers. If you start to see serious differences in some cells, I'd stop the test and yank the bad ones out and then continue. Honestly, since you'll be doing like, 100, it'll probably be like painting the Golden Gate Bridge. As soon as you finish painting all the way to the far end, it's time to start over at the beginning again.
Should be able to do it on a Saturday for nearly zero budget.
You don't really need to test them under load like you would with lead acids. Symptoms will be a refusal to rise in voltage, a quick drop in voltage, or obviously excessive heat at any point in the process.
