|
Bios
and Overclocking
Initially,
I thought the KK266plus would have more options
in the bios than the KK266 but it was the exact
same bios inside. All the options present in the
KK266 were also present in the KK266plus. The
bios date of the KK266plus was also the latest
bios version of the KK266. Nothing is new; the
same memory tweaks were there, overclocking was
also done under Iwill Smart Setting. Since the
KK266 has been known to hit high bus speeds without
problems, expectations on the KK266plus were at
least the same. I managed to hit a bus speed of
164MHz by lowering the multiplier of my AXIA 1.2GHz
C-type to 8, turning down memory timings a little
but still tried to maintain good settings. Running
164MHz bus is so much fun and this time around,
I added a LAN card to the test system so I could
be able to test my peripherals as well. The KK266plus
was simply impressive. Not a single crash in all
the benchmarks even at the overclocked settings.
While I cannot directly compare the stability
of the KK266 to the KK266plus, I feel more at
ease with the KK266plus.
Test
System
Athlon
1333MHz (10X133), 1312 (8X164)
Iwill KK266
Iwill KK266plus
256MB Mushkin Rev3+ High Performance PC133 SDRAM
(Fastest 5-2-2, CAS2 for 133MHz fsb, Faster 6-2-2,
CAS3 for 164MHz fsb)
Elsa Gladiac Ultra Limited Edition (250/460)
Creative Labs SoundBlaster Live! Value
D-Link DFE-538TX 10/100Mbps LAN Card
Maxtor DiamondMax Plus 60 20.4GB 7200RPM ATA100
HDD
Windows
98 SE
DirectX8a
VIA 4.32v
Detonator 12.90
Shot
of the KK266plus running |
Another
shot of the test system |
Benchmarks
Quake
III Arena
Fastest 640X480, Normal 800X600, High Quality
1024X768
MDK2
Trilinear Filtering, Max Texture Detail, Hardware
T&L 640X480X32, 800X600X32, 1024X768X32
3DMark2001
Triple Buffer, 16-bit Z-buffer, Hardware T&L
1024X768X16
Double Buffer, 24-bit Z-buffer, Hardware T&L,
1024X768X32
BAPCo Sysmark2000
SiSoft Sandra TE Memory Benchmark
The
benchmarks are pretty much similar to the previous
reviews. As noted above, the bus speed of 164MHz
was attained by lowering the multiplier of the
1.2GHz C-type AXIA Athlon to 8, and by turning
down the memory timings a bit. Clock speed was
lowered because this particular AXIA would not
overclock well beyond 1.4GHz so instead of running
a high speed at an unstable setting, the bus speed
was raised but the clock speed was maintained
low. Now let's start with 3D Benchmarks.
3D
Benchmarks |
