VA6 VIA Apollo Pro 133 - Page 4
For comparison with my BX6-II, the VA6 was test with
components such as PC100 DIMMS and ATA/33 HDDs.
Celeron 300A SL36A
X LGS PC100 SDRAM
Deskstar 10 10.1Gb HDD (ATA/33)
Deskstar 8 8.4 Gb HDD (ATA/33)
40X CDROM Drive (ATA/33)
6X DVDROM Drive (ATA/33)
3D Blaster GeForce SDR (3.53 reference nVidia drivers)
Networks ENI-25P ATM card
940TX 10/100Mbps Ethernet card
Stylus Photo 750 USB
Now do you think installing
a motherboard can EVER be a painless thing? No. Big NO NO!
It is always painful and the hassle alone can ruin your entire
I took apart my system for the second time in the
week, displaced all the necessary stuffs and reloaded it
with the VA6. Up to this point, not too bad. The smaller
board fitted snugly and it actually took me less time to put
it in than taking out the larger BX6-2.
My first round of test was
with a Celeron 300A CPU and PC100 DIMMS. Therefore, next to
go in were my DIMMs, CPU, GeForce, ATM card and Ethernet
card taking up their usual positions. With all the cables
properly attached, I crossed my fingers and booted the
system. Not a hitch!
BIOS and Overclocking!
SoftMenu II and Not III?!?!
Straightaway, I hit the DEL key and was in the BIOS
setup. To my disappointment, splashed on the screen was
SoftMenu II and not SoftMenu III (already found on the BF6
and BE6-II). Damn, had to pass the delight of Mhz-by-Mhz
Forget the Mhz-by-Mhz thrill, I went to check the steppings
I could take the Celeron 300A knowing that Abit will still
have a decent bunch of useful FSBs I could use. I saw these:
66 (1/2), 75 (1/2), 83 (1/2), 100 (1/3), 103 (1/3), 105
(1/3), 110 (1/3), 112 (1/3), 115 (1/3), 120 (1/3), 124
(1/3), 133 (1/4), 140 (1/4) and 150 (1/4).
And as for
multipliers, you will find: 2.0, 2.5, 3.0, 3.5, 4.0, 4.5,
5.0, 5.5, 6.0, 6.5, 7.0, 7.5 and 8.0. So we have 14 FSB settings
and 13 multipliers to play with.
Well still, I was a little
disappointed at what I saw. The SL36A was a chip I tested
through and through, and I was sure I could achieve a
527Mhz, 4.5 x
117Mhz (using a 1/4 divider, 1/3 results in data corruption
in my IBM harddisks), but the options jumped from 112 Mhz
(1/3) to 115 Mhz (1/3) to 120 Mhz (1/3)…
I was aware of
the risks of corrupting the data on my HDD, but I took the
chance and ran 115 (1/3) for a 38.3Mhz and 76.6Mhz PCI and
AGP bus speeds respectively. Hopefully, I won't encounter
harddisk failure and the higher bus speeds will compensate
for the 10Mhz drop in processor speed in the tests to
Interestingly, there were many other new options that I had
not been aware of on my older Abit. Under Chipset Features
Setup, you can change the DRAM timings of the individual
memory bank, with values like SDRAM 10ns, 8ns, Normal,
Medium, Fast and Turbo. Turbo sounded good to me, so I used
Options such as Enabling
and Disable AGP2X and OnChip Sound were some of the newer
things I came across. CPU to PCI Write Buffer? What's that?!
No AGP4X you say?!
Features such as support for
1/2 AGP multiplier and the ability to run the memory
asynchronously with the FSB (either run at FSB speed, or at
a fixed 66Mhz) are unique to the Apollo Pro 133
The 1/2 AGP multiplier is a crucial determinant for
overclocking success when we try to use FSB speeds of 133Mhz
and above. The BX allows for only 1/1 and 2/3 multipliers,
and it gives us a very high 89Mhz AGP buss speed even when
using 2/3. This puts tremendous stress on the component and
therefore, it is usually a rarity that to achieve a stable
system at such high FSB speeds. But now, with the Apollo Pro 133
and the advent of PC133 SDRAM, we should hear of more
For the 300A SL36A, locked at a 4.5 multiplier, 115Mhz was about the best
I can get with PC100 RAM. Windows booted and
reconfigured itself for the new platform.
hitting the desktop, I installed the drivers for the onboard
PCI audio and the VIA 4.13 Service pack with its own AGP
driver and UDMA tools.