The obvious foil to put the DiamondMax VL17's performance in perspective is Maxtor's own DiamondMax 40. The first thing we should point out is the identical transfer rate
posted by the two drives, indicating that both use the same basic platter assembly. Also note the difference, or lack thereof, in access times
. While specs indicate that the VL20 should lag the DM40 by at least 0.5 milliseconds, the actual difference is less than 0.2ms. Thus, the only really apparent difference between the VL20 and the DM40 (aside from platter count and resultant capacity, of course) is buffer size: the VL20's buffer is one-fourth the size of the DM40.
In a previous test, the impact of increasing cache size by 100% was shown to be minimal. So how does the VL20 stack up when compared to the DM40? As it turns out, the value-priced drive lags behind its full-blown big-brother across the board. The VL20 trails the DM40 by margins of 11%-14% in the ZD Disk Winmarks under Windows 95, and by 16%-18% in NT. ThreadMark results place the VL20 20% behind the DM40 in Win95. NT, however, exhibits a difference of only 4%.
Like the DiamondMax 40, the VL17 features remarkably quiet operation. It's certainly one of the quietest drives around. It's also the coolest drive we've yet tested. It's quite likely that these attributes result from the VL20 being the first 2-platter (most tested units feature at least 3 platters) drive we've tested at StorageReview.com
Conclusion? Well, the VL20 is certainly doesn't match the performance of the DM40. Who would expect this to be the case, however? Maxtor would be in somewhat of a quandary if this were the case. What is interesting, however, is that the VL20 significantly lags behind another seemingly identical drive lacking only the other drive's larger buffer. The way we see it, readers can draw one of three different conclusions from these results:
- Despite presented evidence to the contrary, buffer size has a significant impact, especially when the difference is fourfold as is the case here. Or
- Despite what Maxtor implies, the DiamondMax 40 features significantly better electronics/firmware that allows it to maintain an edge. Cynics will like to point out that it probably involves more effort and/or cost to "cripple" the VL20. Or
- The VL20's firmware is optimized for 2 megs of buffer rather than 512k and thus suffers more of a performance loss than it should. This last conclusion, incidentally, matches our own conclusion in a previous article: Increasing a drive's buffer size results in little performance increase without firmware optimization. Inversely, decreasing a drive's buffer size without tailoring the drive's electronics around it results in a disproportionately large drop in performance.
Whatever the case may be, the VL20 is not likely a drive that performance-oriented users would consider. But then again, 5400rpm drives as a whole probably aren't often considered by power users. They are, however, drives that aim to be incorporated into low-cost systems, whether they're being assembled by an OEM or by a savvy end-user. Here, depending on the enduring price structure, the VL20 may have succeeded in its goals. The true test, of course, will be how this value drive compares to its value-priced competition. Reviews of which, naturally, are coming soon