Long-term performance analysis of Intel Mainstream SSDs
Intel Option #2: Write Small
Quoteth Intel: “An alternative method (faster) is to use a tool to perform a SECURE ERASE command on the drive. This command will release all of the user LBA locations internally in the drive and result in all of the NAND locations being reset to an erased state. This is equivalent to resetting the drive to the factory shipped condition, and will provide the optimum performance.”
The ability to wipe the LBA table was the obvious solution to our dilemma. Since the LBA remap table is where all of the write combination vectors were stored, wiping it out would theoretically bring our Intel drive back to virgin status. I surfed over to the Center for Magnetic Recording Research and grabbed HDDErase 4.0. This is an excellent utility for securely wiping an ATA device by issuing it the elusive SECURE_ERASE command, which on most hard disk drives will wipe all disk areas, including those previously marked bad by the drives own firmware. There are a few notable catches, as the utility must be run in pure DOS mode, and it can only access drives appearing under legacy IDE BIOS ports, meaning you must manually disable AHCI for the utility to work. After getting past these hurdles we came across a brick wall: HDDErase 4.0 is not compatible with the X25-M!
HDDErase 4.0 choked on the Intel SSD regardless of the options chosen.
I was about to call it quits when I learned that Intel was supplying reviewers with a copy of an older (and more importantly, compatible) version of HDDErase. We tried it and it worked… fast! It took less than a minute to execute the erase command. This is primarily because telling flash memory to ‘erase’ and ‘write’ are two entirely different processes. A single flash block erase command wipes out 512KB at a whack and takes only ~2ms to complete.
Successful wipe using the older HDDErase 3.3.
*Edit* - Due to the large number of requests for HDDErase 3.3 after this article went live, we have made it available here.
Back to Square One
Once restoring our X25-M back to its former glory (whew!), we ran a quick ATTO pass:
ATTO bench of our freshly slicked X25-M (note: AHCI was disabled for HDDErase).
Another Paragon pass showing the X25-M at full write speed (~80 MB/sec).
Next we imaged our OS partition back to the drive and promptly ran another ATTO pass:
ATTO bench of AHCI-enabled X25-M with OS present.
All we did here was write the OS files (~20GB) back to the drive. This was not a full blown installation process or other activity that would normally cause excessive fragmentation. What you see is the result of internal fragmentation caused primarily by write combining. The drops in read speed are also caused by this process. Even with AHCI significantly boosting the reading of fragmented areas, the X25-M can still fall prey to its own fragmentation. The point to take home is that merely placing an OS or any other group of files onto the Intel drive is enough to bring it below its specifications.
Another point to consider is that ATTO performs its benchmark on a live partition by writing a large test file to its root directory. In the above test, that file *should* have gone to a never before used portion of the drive, but the write combining / wear leveling have acted to spread the performance drop as equally as it spreads the wear. Similar results were obtained running an additional ATTO pass on a partition we created at the end of the drive.