UltraDMA66 with a chipset BX?
For that they have not read the report on the actuality of the hard disks, the UltraDMA66 is a new transference way between the hard disk and the computer. Up to the date, the most powerful available way was the UltraDMA33, which was allowing to reach a valuation of 33 Mb / century maximum theoretical transference
This quantity has been more than sufficient during enough time; nevertheless, the new hard disks, especially the models to 7.200 rpm, have "tops" of transference near to 30 MB/s (momentary, the speed comes up hardly it overcomes 10-15 MB/s), being necessary to evolve the IDE one more step. That does the UltraDMA66: it allows up to 66 MB/s of transference (another different thing is that the hard disk is capable of taking advantage of it...).
As we were saying in the report commented above, to use this technology we have several requisites:
- That the hard disk is UltraDMA/66, of course.
- That the cable is 40 pines with 80 drivers instead of 40 pines with 40 drivers.
- That the operating system has support for DMA.
- That the checker IDE supports UltraDMA/66.
The BE6 and the BP6 solve to the perfection the last three requisites. At present, the hard disk checkers are integrated in the chipset of the motherboard; nevertheless, the chipset BX does not support any more than UltraDMA33, despite being most advanced up to the date (with pardon of new Apollo Pro 133 of ROUTE, more modern but for the time being with few acceptance, and of the scale chipset Intel average - fall i810 - very little advisable-). Therefore, we would need a card of expansion with a control chip IDE... or a control chip IDE integrated to the motherboard, which is the solution that ABIT has adopted.

This small chip of the photo, the HPT366 of HighPoint Technology, is the culprit of that a badge BX commits the wonderful heresy of supporting UltraDMA66. And since thanks to this chip we have two new channels IDE, in addition to the two that the chipset BX provides (that of course they have not been eliminated), the whole climbs up to 4 channels IDE, that is to say, up to 8 devices.
A new technology integrated to not standard chip, we will need to install a controler to the being for our operating system. ABIT includes a diskette with the controlers necessary for Windows 95, 98 and NT; the installation process is described so in detail in the manual (independent explanations for Windows 95, 98 and NT, with 32 images of the screen to illustrate the steps to be realized), that at first sight it can seem long or complicated; not further at all of the reality. One can go so far as to do in less than 1 minute, and the only difficulty consists of being able to read the exhaustive instructions... of course, in English.

Finally, ABIT takes charge also of the requisite of the cable: both basic badges include a special cable of 80 drivers, necessary to be useful UltraDMA66. This cable is very peculiar, with different colors in the connectors (black, gray, blue) as where there should get a job each one (master disk, I dial slave, motherboard); and remember that the sockets for this cable are of white color instead of black. Really, these badges are full of color!
The BIOS of the badges includes a pair of minimal modifications that allow to start from the first channel UltraDMA66 (the connector IDE3); of not being like that, the system would start from the first channel normal IDE (IDE1), which is only UltraDMA33. During the tests we have tuned in up to 7 devices to the system: a disc UltraDMA66, two UltraDMA33, other PIO4, two CDs-ROMs and a tape recorder of CD - ROM; apart from the enormous jumble of cables and the need to use a box supertoast to store so much pot, there was no problem; of course, insure of the sufficiently powerful thing have a power supply...
Why do they serve two Celeron (two!!)
Well, we have the first dual motherboard for Celeron; and now, what? Why does that serve? If we put two 400 MHz Celeron, do we obtain 800 MHz? No, I hope. To have two mikes in a computer can increase enough the process potency, although we will never go so far as to double the effective speed of the system, between other things because the rest of devices (hard disk, memory, videocard...) does not double.
What yes we will be able to obtain is:
- To distribute the work load between the diverse mikes, what is very useful in case of realizing multitask (to have more than one program working simultaneously). The typical example would be a servant.
- To increase the potency in certain applications that need a lot of calculation potency, as there could be the CAD programs. This would be the case of the workstations multiprocessor.
Therefore, we can use the BP6 to mount a servant, of low scale but with an exceptional price, or to mount a workstation quite useful multiprocessor with an even more incredible price. It is enough to look at the following prices table:
On the other hand, we will need an operating system that allows us to make use of two mikes simultaneously. If we use Windows 95 ó 98, one of the mikes will be very rested, since these operating systems do not support any more than one mike. To take advantage of this technology we will need "serious" operating systems as Windows NT or Linux. If he likes the easy thing, choose NT... and pay the price (never rather); if you want free efficacy, obtain Linux... and a good manual it to learn to handle.