Cray (NASDAQ:CRAY), one of the oldest supercomputer companies around, reported earnings Monday night, and the results far surpassed analyst expectations. Quarterly revenue landed at $38.5 million, down from $53.4 million a year ago, but well above the projected $30 million. On the bottom line, there was a $0.32 loss per share, while Wall Street expected a $0.46 loss; the year-ago loss was $1.08 per share.

The thing to remember is that Cray's earnings and revenues are very lumpy, because the products the company sells are very expensive, and it's a small fish in today's technology pond. If you shift the timing of one or two of these large orders to the next quarter, this quarter will look horrible -- and vice versa. As a result, management can't commit to a very firm annual forecast, either, though it hopes to see flat or improved revenues and earnings over fiscal 2005.

Likewise, the research and development line would normally worry me, because that number shrank by more than 50% year over year. However, management has committed to doubling R&D spending in support of a Department of Defense project to develop a "petaflop" computer, so that lowers that red flag for me. To put that project into perspective, one petaflop is about three times faster than today's biggest and baddest supercomputer, and nearly nine times the performance of No. 2 on that list.

If you look over the top 10 fastest computers in the world, you'll note the dominance of IBM (NYSE:IBM). It also drives home the intensely competitive nature of Cray's field, including much larger companies like IBM, NEC (NASDAQ:NIPNY), Dell (NASDAQ:DELL), and Sun Microsystems (NASDAQ:SUNW). Widen the scope to the top 500 systems, and Hewlett-Packard (NYSE:HPQ) makes a strong entrance as well.

After the recently completed reverse stock split, Cray is out of penny-stock land. Government contracts are a nice way to build revenues, but it takes a patient investor to make money from this stock. And with those lumpy sales, remember to wear a seat belt, because the ride will be very bumpy.

For further Foolishness:

You'll find Dell recommended in both Stock Advisor and Inside Value. A free trial will show you why our advisors -- investors like you -- favor this stock.

Fool contributor Anders Bylund holds no position in any of the companies discussed here. You can check out Anders' holdings if you like. Foolish disclosure doesn't require anywhere near a petaflop of performance.