Forum2News_Exporter@mts.rpi.edu (Barry Bailey) (01/18/91)
I too have perceived a performance problem with the stated configuration. I suspect though, the problem is related directly to the amount of RAM installed in our systems. We have 16Mhz with 1M, 2M and 4M (different purchasing deals depending on the phase of the moon), 20Mhz with 2M and 4M, 25 Mhz with 6M. The machines with the same amount of RAM seem to perform at similar speeds, regardless of clock speed. We later brought the 20Mhz 2M system upto 4M and saw a significant increase in apparent speed. When the 25Mhz had 4M of RAM it performed marginally better than the 20Mhz/16Mhz systems with 4M (none-the-less better - a "percentage" increase vs "multiples" increase). Anything above 4M (i.e. 6M) doesn't seem to effect system performance for "normal" applications. 6M does enable you to run more applications and/or larger data-sets at similar speeds (vs SWAPDISK speeds). The systems described are similarly configured in other respects. Although the 25Mhz system came with a 70M 18ms HD. For disk intensive operations this faster disk, combined with increased SMARTDRV RAM, caused a "multiples" increase in perceived performance, over the 16Mhz 40M 23ms HD systems. In summary: Upgrade from 1M or 2M RAM to 4M RAM to realize a cheap "multiples" increase in perceived performance. Upgrade from 4M to 6M, 8M, etc. to realize a cheap "percentage" increase in perceived performance along with a marked increase in capacity. Upgrade the harddisk to a bigger/faster model (along with an upgrade of SMARTDRV RAM) for a "multiples" increase in disk intensive perceived performance. Lacking time and resources I have emphasized "perceived performance" vs "tested performance ratings". hope this helps. barry usere9w9@rpiecs - OR - barry_floyd@mts.rpi.edu