X58A-UD3R takes 30sec to wake from S3

[Gary123]

I just realized I made a dumb mistake. I'd had trouble squeezing my GTX 260 into the top PCIe slot, so I put it in the slot just below, not realizing that's x8 rather than x16. So I need to move the card, and perhaps that'll fix the S3-resume delay. I'll let you know.

[Dark Mantis]

Yes it could well be to do with that. Please let us know if you have sorted it now. 

[Gary123]

I'd had trouble squeezing my GTX 260 into the top PCIe slot, so I put it in the slot just below, not realizing that's x8 rather than x16. So I need to move the card, and perhaps that'll fix the S3-resume delay.

I moved my GTX 260 to the top slot. While I was at it, I also moved my SSD boot drive from Channel 2 master to Channel 0 master.

It did make a difference: now it takes 40 seconds to resume from S3, instead of 30 seconds!

StandbyResumeResumeDevices is still about the same (636ms), but StandbyResumeBiosInitTime has increased to 37382ms.

[Dark Mantis]

The extra time it takes is probably because you moved the Boot drive and now instead of it being where it was originally it has to locate it first.

[Gary123]

The extra time it takes is probably because you moved the Boot drive and now instead of it being where it was originally it has to locate it first.

I've rebooted a few times and resume-from-S3 still takes 40 seconds now. Would I need to re-flash the BIOS in order for it to remember the new location of the boot drive? (The boot-drive priority list in the BIOS properly lists the SSD first.)

[Dark Mantis]

Not re-flash the BIOS but clear the CMOS. Make sure you do a complete clear not just press the button.

[Gary123]

Not re-flash the BIOS but clear the CMOS. Make sure you do a complete clear not just press the button.

I'll try that, but just out of curiosity, why wouldn't it save the new SSD location to CMOS when I do a save-and-exit in the BIOS setup? (The BIOS does list the SSD in the Channel 0 Master position.)

[Dark Mantis]

To be perfectly honest I am not 100% sure so I was taking the safe option.

[Gary123]

To be perfectly honest I am not 100% sure so I was taking the safe option.

Ok, thanks. It turns out I'd have to remove the graphics card again to get to the CMOS battery, so I'm going to avoid that unless there's a strong reason to think it would solve the delay problem.

[Dark Mantis]

No sure reason just trying to cover all the possible bases.

[Gary123]

Gigabyte tech support wants me to try it with just one memory module. I don't imagine that'll make a difference, but I guess I'll do the experiment.

[Dark Mantis]

It's the same as on the forum they will go through a list of things to try and narrow down the possibilities. So you need to do as they ask.

[Gary123]

Info on changing to AHCI mode after windows install. http://www.sevenforums.com/tutorials/61869-ahci-enable-windows-7-vista.html

I found that the specified registry  key is already set to 0, and the Device Manager already showed two Intel ICH10 SATA Controllers. So I enabled ICH SATA AHCI in the BIOS, and the BIOS subsequently detected SATA devices (rather than IDE devices), but then the bootstrap hung after "Verifying DMI Pool Data" and never got to the Windows startup screen.

I switched the BIOS back to ICH IDE, and was able to boot normally again.

[Dark Mantis]

What you would have to do is clear the DMI memory pool. To do that you have to use QFlash again and then on the main screen disable Keep DMI Data

[Gary123]

What you would have to do is clear the DMI memory pool. To do that you have to use QFlash again and then on the main screen disable Keep DMI Data

Instead of re-flashing, I just cleared CMOS (via the button), and then I was able to get AHCI to work. Plus that totally fixed the S3-delay problem. It now resumes from S3 in 3 seconds instead of 40! (I had tried the clear-CMOS button earlier, and it didn't resolve the S3-delay problem, so apparently it was AHCI mode--not just clearing the CMOS--that fixed the S3-delay.)

Of course, there really shouldn't have been such a long S3-resume delay even in the default IDE mode. So that's something that Gigabyte may want to investigate (I'll tell their tech support about the outcome). But I'm all set now.

Many thanks to you and to Bill for your suggestions.