Hardware MSI P45 Platinum

Thanks very much! Much obliged
The PWM controller should be the square IC just above and to the right of the top-most RAM slot, at the same level as the motherboard's right-most screw hole.

As is typical when photographing an object with many components of different heights, focus is an issue, but one can just make out the stylised 'i' on the chip (horizontally-oriented in this photo) - meaning Intersil. Exactly which model is unclear, thought it's probably either an ISL6327 or an ISL6336.

It is certainly not possible to examine the motherboard design details (trace routing, cooling via copper fills, decoupling etc), but on the face of it the power circuitry design seems very decent - both in terms of theoretical current as well as thermal handling capability: the DrMOS is a real space-saver, and allows the heatsinks to cool even the normally-neglected high-current driver ICs while eliminating the problem of varying component heights which would have been an issue if the MOSFETs and ICs were all discreet.

Let's hope MSI sorts the bugs out pronto; if so, they could have a pretty decent product on their hands.

Thanks again for the review and pics

 

Ah that one. I get confused between clock generator (between PCI-E slots) and PWM controller

 

I hear what you say but 500FSB+ overclocking seems to occur only on v1.0 of this board mine is version 1.1. There are 2 users on MSI's forums one of whom is a moderator who both can clock past 500FSB but when pressed on the version number of their board I have as yet no answer.

I understand the bios is immature but it is odd that somebody with the same ram, cpu and bios settings on air cooling can get past 500FSB with ease but I am stuck at 450FSB and I have a cheap MSI P35 board where the same components clock to 520FSB.
The only other difference is his board is v1.0.

 

Hehe no worries The PWM controller is always very close to the CPU PWM circuitry and on most motherboards, it's often within the 1 o'clock-to-5-o'clock sector of the CPU socket (12 o'clock being the rear I/O panel). Most common is 4-5 o'clock, because the VID pins of the LGA775 CPU are in that area.

As for 500+ FSB and v1.1 boards, it's possible that the circuit traces on the PCB were revised for better stability etc, but in consequence the signal timings/skew etc have to be revised as well to match... hence we might get back to reaching that level even with the 1.1, but it'll probably need a newer BIOS version or two.

 

Not all the code in your BIOS comes from the mainboard OEM or the BIOS vendor. Being an Intel chipset, there is one critical piece of code called the memory reference code that comes from Intel, so there is usually some delay in the problem being found and when the mainboard manufacturer can actually get the fix into their BIOS. So unfortunately, it isn't totally up to MSI, they depend on other companies and have to live with their schedules too.

 

We reviewed version 1.1

PWM controller is the ISL6336A

 

I have confirmed 3 people now who have clocked past 500FSB ALL with v1.0 of this board and 6 users all with v1.1 who can't get anywhere near it. It is annoying that certain reviews of this board have gone out when they are only review samples as it gives off a false impression.

I know a bios may fix it but am not very confident about this as had a similar issue with an abit board a few years back and it was fixed with another board revision. I never bought another board from them after this and the same thing will happen with MSI as well off to asus methinks.

Just found tweaktowns review where it states its the final version although it is v1.0 and gues what goes past 500FSB.

 

The base code of the BIOS generally would come from the chipset manufacturer, yes - just as in the case of Nvidia chipsets. However, if it was solely a base code problem, then all or at least most implementations of the chipset would suffer the same problems. In this case, a product from MSI based on the chipset in question seems to have issues with reaching 500MHz FSB - but products from Gigabyte and Asus, for instance, do not seem to have as much trouble in this department although based on the same chipset. Hence it would strongly suggest an issue with MSI's circuit board design and/or BIOS optimisations - not to mention a somewhat rushed product testing phase. If they manage to tune the latter to work well with the former, great, but if not then we might not see a fix until another board revision comes out.

Bindibadgi: ISL6336 it is then! Spanking new, supports up to 6 phases under VR11.1 (ISL6327 is similar except that it's a VR11 device).

 

Sorry, didn't mean to imply that it is solely an MRC problem, only that BIOS is a complex beast and not all of it is under the control of the OEM. I'm sure they are doing everything they can to resolve the issue, but if they need to pressure an outside source for help, it just makes the process take longer.

 

Were there any overclocking updates in the end? Just wondered, as you can get this mobo with a Zalman CPU cooler for under £90!