Method and structures provide for testing a SAS link in association with participating in training windows to determine success/failure in using a negotiated speed using one or more configured sets of transceiver training options. For each device linked to a master SAS device, each possible set of transceiver training options is configured and one or more SCSI requests are forwarded from the master device to the attached device. The SCSI requests may be non-destructive of data stored on the attached device. Results of the tests may be used to select a preferred set of transceiver training options for communication between the master device and that attached device. The transceiver training options to be varied and tested may include: amplitude, slew rate, de-emphasis, and spread spectrum clocking.