[Gigabyte P55A-UD4P] Does PCIe bus supply mainten power while suspend?

[wmp]

Does PCIe bus supply maintenance power to client cards while suspended?

The reason why I am asking is that I have a 3ware 9650SE raid controller with BBU and I don't want to put my system to sleep if doing that will draw down the BBU's battery. I imagine that power is maintained to some components so that WOL for example will work, but is power maintained to the bus generally or to the 3ware in particular when the system is in suspend mode? User docs for the 3ware says that power-off will drain the battery, but it does not mention the effects of "suspend".

MB: Gigabyte Technology Co., Ltd. Model P55A-UD4P
Raid card: 3ware 9650SE-4LPML with BBU

OS: Microsoft Windows XP Professional
Version 5.1.2600 Service Pack 3 Build 2600

 

[GhislainG]

Suspend and power-off will drain the battery, but it will get recharged as soon as the system is powered up again. I presume that your system is powered up several hours every day?

 

[wmp]

Yes, it normally spends some on-time every day. However it seems to me that multiple partial-discharge/charge oscillations would wear out the battery faster, would it not? Seems to me that there is power on the board somewhere during suspend, it seems like a shame to not use it to power the raid card BBU

 

[GhislainG]

RAID cards with BBU are designed for servers that are powered on all the time.

 

[wmp]

I guess so. Incidentlly, a 3 Gig files copy that takes almost exactly 2 minutes with write-cache on takes almost exactly 12 minutes with it off. Copying was the entire C:\Program Files (from an SSD) to the 3ware with a 3-drive RAID 5 unit. That's 24.57 Mb/s versus 4.24 Mb/s. HUGE performance difference.

 

[GhislainG]

I agree that write-back caching is great, but controllers with BBU should be used on systems that are powered on all the time. I always use RAID controllers with BBU on servers, but I never do on PCs.

 

[wmp]

OK, I'll enter this in case anybody ever reads this....

I have determined that the 3ware 9650SE software logs every instance of battery charging, even if it is charging for a second or two only. I have also determined that the slightest amount of battery discharge will trigger a charge action log event at the earliest opportunity.

Accordingly, if there is no battery charging after a period of "Suspend", then it is safe to presume that there was no battery discharge occurring during the suspend, and that, by implication, the 3ware 9650SE card had power during the entire suspension.

That is in fact what I found with two test periods of 12-hour suspend: there was no battery discharge evident and therefore it seems, for this MB and card at least, that power is in fact maintained to the card during suspend.