VeraCrypt not working correctly on M.2 Samsung 960 Pro
Open source disk encryption with strong security for the Paranoid
Brought to you by:
idrassi
Menu
▾
▴
I think the SSD performance issues can be traced to two different root causes.
The first is the IRP replacement issue which has been acknowledged by the devs. This causes a general reduction of performance, but affects random access much more than sequential access. This should not visibly degrade performance to mechanical HDD levels.
My theory is that there is a second issue when the SSD is completely filled with random bytes during encryption (by selecting wipe mode different from none for system drives or not selecting quick format for other drives). This would make the SSD controller think the whole drive is full and cause a lot of write amplification. The consequence would be extremely slow writes and 100% disk activity like dmitriy a and a few others experienced.
In order to verify my thoery, can someone affected by the slow writes check for write amplification during the benchmark using a hardware monitor such as HWiNFO64?
Even better if one of you could issue the drive a secure erase command, reinstall Windows and encrypt without wiping the drive. Be sure to also turn on the TRIM option on VeraCrypt.
If erasing the drive is not an option you could try filling all the free space on the drive with copies of a big file and deleting them afterwards. After some time TRIM should zero that area and write performance should be restored.
It'd be fantastic if one of you could test this theory. If it is correct this should explain why some people have decent performance with SSDs and others don't.
Good theory. I can tell you that the 100% disk activity also happens inside VC containers. So, no issues outside the containers, but inside them everything is super slow.
I've observed this problem with NTFS+NTFS container on WD_BLACK HDD 8TB connected via USB 3.0. Annoying. 5 MB/s write speed! From SSD...
Has anyone checked exFAT with this case?
Hello,
exFAT does not provide increase of performance
problem come from file containers handling; I/O data must be re formatted for second NTFS access this take small time but decresase dramaticaly SSD I/O rate
on HDD this small time disappear among others latencies
on SSD is affect directly number of operations by second , performance can be divided by 10 or more
may be solution is to plane spécial veracrypt release whithout support of file containers where I/O don't need to be reformated
Best regards