Rien ne va plus! The ultimate RamDisk driver test is here.
Drivers tested:QSoft RamDisk Enterprise 5.3.1.10
ImDisk 1.40
M$ RamDisk 5.2.3790.1830
FiraDisk 0.0.1.30
WinVBlock 0.0.1.8
Testsystem: i5-750 @ 2.67GHz, 6GB/PC3-10600, LiveXP (Windows XP Professional [5.1 Build 2600] (x86)), 235MB ISO
QDisk10.gif 27.59KB
56 downloads
ImDisk1.40.gif 27.78KB
58 downloads
RamDisk.gif 28.02KB
48 downloads
FiraDisk.gif 32.4KB
50 downloads
WinVBlock.gif 25.2KB
43 downloadsMost impressive, ImDisk outperforms in almost all categories QSofts RamDisk which was shown to be the fastest in all internet comparisons.
Strange is that M$RamDisk actually performs worst with the bigger queue depth.
And now to the boot-time comparison. I timed again the boot process of a LiveXP from end of ramdisk loading till PENetwork startup.
with M$RamDisk - 14,55 seconds
with FiraDisk - 16,60 seconds
with WinVBlock - 19,50 seconds
Again, especialy the huge difference between WinVBlocks performance and the one of the two other drivers in CrystalDiskMark and the small difference in boot time. Most puzzeling!
And now! Especialy for our italian viewers. Live from Germany! A comparison of the complete boot time. Meassured from pressing of the power button till PENetwork startup. (LiveXP is loaded from a DVD not a CD, which would take even longer.)
with M$RamDisk + Setupldr - 2:20:48 minutes
with FiraDisk + grub4dos - 2:22:66 minutes
with WinVBlock + grub4dos - 2:25:82 minutes
For me, it is M$RamDisk + Setupldr all the way. Mostly due to a reason, not shown here.
Only with this combination of scripts, is it possible to have a PE, that can switch between Uni- and Multiprocessor setup.
With the other two drivers, one has to decide for the one or the other at build time. Multi choice setup is so messed up, i couldn't even figure out, where things started to go wrong.