rpi-update fails when overclocking

[lingon]

After the SD-driver bugfixes I have started to test once more with overclocking. On a 512 MB Raspberry Pi Revision : 000e rpi-update failed with this error message:
...
Cloning into '//root/.rpi-firmware'...
remote: Counting objects: 3203, done.
remote: Compressing objects: 100% (2675/2675), done.
error: RPC failed; result=56, HTTP code = 200B | 2.13 MiB/s
fatal: The remote end hung up unexpectedly
fatal: early EOF
fatal: index-pack failed
!!! Failed to download new firmware files

I run rpi-update twice and got very different error messages, but I don't have the first error messages saved. Setting overclocking to none and rebooting allowed to update from
Linux raspberrypi 3.10.24+ #614 PREEMPT Thu Dec 19 20:38:42 GMT 2013 armv6l GNU/Linux
to version
Linux raspberrypi 3.10.25+ #616 PREEMPT Mon Dec 23 18:13:02 GMT 2013 armv6l GNU/Linux
The problem occurred when using turbo overclocking. I have previously seen on this Raspberry Pi problems with running
rpi-update and turbo overclocking, so the SD-bugfixes have not unfortunately solved this particular problem.
I did not see any error message with dmesg after the rpi-update command failed. The SD card is a SanDisk 16 GB Extreme Pro 45 MB/s UHS 1 card.

Has anyone else seen SD-cards problems with turbo overclocking after the recent SD-bug fixes?

What could be done to further debug this issue?

[MaxK1]

Not all pi's overclock equally well. I'd try something more modest and work up from there. But
usually a failure is more abrupt (kernel crash)

[klricks]

OC is not guaranteed to work. You fix OC issues by not overclocking.

[lingon]

The SD-driver has been improved a lot lately, thanks to the hard work of the developers. Previously I have too been bitten by crashed filesystems on three different Raspberry Pi machines. With the updated SD-driver I have seen no issues so far.
I have previously run the overclocking stability test at http://elinux.org/RPiconfig#Overclock_stability_test without seeing problems, so the Raspberry Pi machines I have can pass that test without reporting error messages when using turbo overclocking.

I am perfectly aware of that overclocking can cause problem with the SD-card. I wanted to see if there are still problems with SD-cards when overclocking and using a kernel with the SD-driver fixes. Unfortunately there still is a problem, at least with the card and Raspberry Pi that I'm using when using turbo overclocking.

It would be interesting to know if other users have SD-card problems with the improved SD-driver to understand the problem better.