ATA100 + 7.200 rpm vs. ATA66 + 5.400 rpm
You try yield
Very well, the whole series of them; let's start by the "synthetic materials", that is to say, those that they evaluate only to the hard disk, and not the general yield of the system.
The first one is the acquaintance Sandra 2000. It is not the most intensive, but it gives us a series of interesting information:
With the cache memory of Windows deactivated ("Bypass Windows Cache"), to evaluate only the yield due to the disc. Since it can turn, there is no color: 7.200 rpm of the IBM, together with his 2 MB of cache memory, give a drubbing to the poor person Quantum, even when the access way limits itself up to the veteran UltraDMA33.
And why does not it change scarcely between UltraDMA100 and UltraDMA66? Because this test points out also that internally the IBM disc is unable to overcome 35 MB/s, therefore it has in abundance with 66 MB/s of the UltraDMA66.
The second test is much more serious: Adaptec ThreadMark: 64 partial tests that last in whole more half an hour and that handle like 1 information GB. In them, we observe the clear assertion of the previous thing:
The Quantum cannot compete, although at least it leaves enough free time to the microprocessor... perhaps because to 733 MHz Pentium III he does not suppose work to take charge of this "wretched" 7,7 MB/s (that are not bad, eye, a few years ago seemed unattainable).
To finish this phase of the tests, we have the reading test of the HD Tach. This test has the advantage of which it presents of graphic form the behavior of the disc to us, in addition to calculating the maximum, minimal, average transference (average) and the use of the CPU. We have selected the most representative graphs:
As for tests "of yield of the finished system", we have decided to leave ourselves of graphs and compare the time that takes every disc (with his way normal UltraDMA) in completing two tasks: the first one, the test Winstone Business 99:
And the second one, the analysis of 404 MB of files LOG of our servant (for the calculation of the statistics of only 1 month, neither more nor less). It is a test of "massive consecutive reading", let's say:
We have several, enough clear as soon as the results were seen: if he is going to buy a new hard disk, it is worth it that it is 7.200 rpm.
It is more: it is worth it be which be the support of UltraDMA of his motherboard. If BX has a badge with chipset or ZX, for example, his 33 MB/s is more than sufficient; a checker UltraDMA100 can buy if he wants, but it will not notice a lot of difference. Of course, in any case use a cable of 80 drivers: they cost little and they are of major quality, being advisable even for checkers UltraDMA33 (although in this case they are not definitely necessary).
Of course, the capacity depends what I needed, and of nothing more. A hard disk of 60 GB will be rapid or slow, but not for 60 has GB, but for his speed of draft, his cache memory and in any case for his access way. Ah, nowadays ENORMOUS discs can buy to themselves nearly money, but bear in mind that it can be dangerous "to put all the eggs in the same basket"... if the fatal errors, already he knows.
If it already has a disc with UltraDMA and is not a fanatic of the yield, do not change it. The slightly ancient discs that UltraDMA does not support (as many of 2 are GB or less) can brake a little the yield, but if it already has UltraDMA and does not import for him that Quake3 late in loading one more second, has why to be updated... not also and if it has a pair of seconds more, it is possible that it serves his "prehistoric" disc to him without UltraDMA.
And, ALWAYS, always, always DO SAFETY COPIES. This time we have lost a Windows "in target", a few tests of yield and a few graphs and texts, but the next time might be his novel of award Pulitzer, his directions agenda, the accounting of his clients... and for that 7.200 serve of nothing rpm not even the UltraDMA100.