During my tests, my PIII700E booted and worked fine at 933MHz under a default Vcore setting. I managed to overclock this further to 960MHz, achieved through a Vcore jumper setting of +0.5%V (1.73V). However, anything beyond 960MHz produced random crashes in Win2K or intermittent bootups, regardless of Vcore or Vio adjustments. This points to 960MHz as the optimal setting, of which all my subsequent benchmarks were run.
During this time, my CPU registered a steady-state temperature of 42 to 46 deg C.
As a side-note: the performance of my Alpha PFH6035 Cooler seems extremely dependent and sensitive to proper positioning on the CPU. At one stage, I encountered multiple bootup failures and discovered that the BIOS read a CPU temperature of over 60 deg C!! Apparently, I had shifted the Alpha slighty off-position during some cable manipulation (though it still looked perfectly placed), which was cured eventually by re-situating the Alpha cooler properly, ascertained by monitoring the CPU temperature reading in the BIOS.
As usual, the standard set of benchmark tests were run to gauge performance.
Sisoft Sandra 2000 Standard
Sisoft Sandra features a suite of tests to benchmark various aspects of the system (CPU, CPUMMX, FPU, Memory, HDD, etc):
As one can see, the benchmarks show no trace of hiccups in the CPU and FPU benchmarks, with results following clock speed proportionately. However, the memory scores leave a little to be desired. This is partly attributed to the absence of memory interleaving support offered in the I815E chipset (not available in the BIOS). Yet, it could just be that I've been overly spoilt by faster memory speeds achievable on Athlons
For the HDD tests above, the latter diagram illustrates the test run on a FAT32 partition, whilst the one before it uses NTFS. In either case, the Intel I815E based ATA-100 driver seems to underperform, rendering only sub-17K scores. In comparison, other ATA-100 drives (also Maxtor DM45) and 3rd party controllers come in higher, at 19-20K points. Nevertheless, the tests indicate that both NTFS and FAT32 hardly differ in performance under Win2K.
WinBench 99 v1.1
WinBench has also become a popular benchmarking standard to gauge performance in real-life applications, using various popular sub-programs to realize this.
The following table highlights the benchmarks derived.
Unfortunately, I couldn't really benchmark video performance, as WinBench simply refused to perform video graphics tests without first halting all background processes, including system citical ones (i.e. CAPRPCSK.EXE, mspmspsv.exe & svchost.exe) in Win2K. Nonetheless, the CPU and FPU marks are about right, attaining 84.6 and 5060 scores respectively, which are just short of a 1GHz PIII on an I815E motherboard.