ASUS P5Q Deluxe

Josh Collins is still struggling to figure out the alpha-numeric system of ASUS’, but oh well, we’re up to Q now.

It’s that time again and the Intel time machine is tick-tocking away.

This month we’re seeing the release of a new chipset for Intel’s Express line. This is in the form of the P45 Express, and is set to replace the aging ‘Bearlake’ based P35. That’ll do P35, that’ll do.

Normally the introduction of a new chipset is hailed as the next step forward in a sensible technological progression. P35 was a step up from P965, and X35 was way too long waiting to replace the ancient 975X. Then it seemed Intel’s chipset division got a slap on the bum and told to go faster. So instead of creating something new, they just speed binned something old and called it something new; enter X48.

With this history, we were a bit hesitant about the P45 release, code named ‘Eaglelake’. There had been whispers for some time, just like there is of X58 right now, about performance gains and overall improved functionality. Oh, and we were to get a new southbridge in the form of the ICH10R.

With the P35, X38 and X48 chipsets as recent examples, there’s quite a lot the P45 has to do to distinguish itself firstly as different, worthwhile and useful.

The first box gets ticked by mere physical differences. The P45 ‘Eaglelake’ based core is a 45nm manufacturing process and therefore requires less voltage than its ‘Bearlake’ 90nm predecessors.

Whether it’s worthwhile and useful is harder to judge. Obviously this board is the first of what will undoubtedly be a string of future releases and is also in a rather infant state in comparison to ‘Bearlake’ chipsets, but we do take issue with the state in which ASUS’ P5Q Deluxe featuring P45 is in. It doesn’t so much feel like an infant; it’s more of a premature birth.

Being first to market is everything in the IT industry, and this notion is even more prevalent in the hardware manufacturing sector; but rushing a product out and effectively using the paying public as beta testers, rather than coming out second, third or later still but with a stellar product, is poor form.

One of the most evident examples of the public beta testing concept is the complete lack of a working 2:3 divider. It simply does not work. At all.

For our testing of DDR2 based motherboards, we run a base standard of DDR2-1000 with latencies of 4-4-4-10 and a preference of 1T if possible. For this reason, when testing with a 333MHz FSB we require a 2:3 divider to obtain the DDR2-1000 frequency. But, it just doesn’t work on this board. So we used the 5:8 divider, resulting in DDR2-1066 frequencies.

While the BIOS sports a lavish range of memory latency settings, the BIOS does not have an option for setting the command rate. Unless we’re extremely blind, the darn thing just wasn’t in sight, or at the very least, is not in a user-friendly location. This ultimately meant our memory trundled along at DDR2-1066 4-4-4-10 at a 2T command rate.



The new chipset looked to demonstrate strong memory bandwidth improvements, though this was hard to make a direct comparison due to the inability to run at DDR2-1000 frequencies.

Disappointingly, for this particular sample, we couldn’t budge above a 400MHz FSB at all – even with a QX9770, a stock 400MHz FSB chip. And for a QX9650, it also stopped at the 400MHz mark.

We believe this issue to be specific to either ASUS and/or this model, as we’ve seen screenshots, and spoken to members of Team.AU involved in beta testing the GIGABYTE P45 boards, of them doing excess of 500MHz FSB, even close to 550MHz in some situations.

All in all, we’re left with a rather unusual feeling about this board. We’re glad to see P45 hit the market but then we also would really prefer to see a completely finished product instead of something that feels like it needs some more R&D time.

If you’re keen on P45, chill for a little longer and wait and see what else comes to the market. After all, it’s early days yet.