Test #1: A1222 vs X5000 For the first test I was hoping to get some good numbers from ScsiSpeed comparing read/write performance at a low-level directly from the SATA drivers -- the latest available versions of p1022sata.device and p50x0sata.device respectively. Unfortunately ScsiSpeed hung bth machines after just two of the initial tests. But, hey, two datapoints are better than zero! ScsiSpeed Test A1222 SATA X5000 SATA Factor ====================== ========== ========== ====== Read using 512B buffer 5575500 11564185 2.07x Read using 4K buffer 32492134 66781184 2.06x bytes/sec bytes/sec Test #2: SFS vs NGFS For the second test I wanted to compare SFS/02 with NGFS; I used the same partition size, same physical disk type (a Samsung 1TB SSD), latest available versions of SFS and NGFS. For stability reasons I have turned off SFS's diskcache library (see previous emails). While the results below are mixed, on an anecdotal basis, I/O performance 'feels' much better with NGFS, especially when starting games which load a great many files. I certainly intend to use NGFS going forward on my A1222. DiskSpeed Test                A1222 SFS2   A1222 NGFS   Factor ===========================    ==========    ==========    ====== Dir testing: file create         2016    1980     0.98x Dir testing: file open             8715    21458     2.46x Dir testing: directory scan      62634    45559     0.73x Dir testing: file delete           2848    3266   1.15x files/sec files/sec Dir testing: seek/read         3458    273772    79.17x seeks/sec seeks/sec 512B buffer: create file          8370112    22317248 2.67x 512B buffer: write to file       19196160    55551744 2.89x 512B buffer: read from file      22564608   131891136 5.85x 4K buffer: create file           30540288    28107156 0.92x 4K buffer: write to file         75011072    240483328 3.21x 4K buffer: read from file        67158016    360107520 5.36x 32K buffer: create file          54198272    27578368 0.51x 32K buffer: write to file       133873664  227770368 1.70x 32K buffer: read from file      121323520    296333312 2.44x 256K buffer: create file         75792384  25395200  0.34x 256k buffer: write to file      184352768    291274752 1.58x 256K buffer: read from file     156762112   304021504 1.94x bytes/sec bytes/sec Test #3: SFS vs SFS For the third test I again used the same partition size, same physical disk type (a Samsung 1TB SSD), same version of SFS from A-EON. The difference, though, is that diskcaching is not enabled on the Tabor, so this isn't a straight comparison test. But it is for the normal user since I suspect that when released, SFS will come with a warning such that A1222 users do not enable diskcache.library.kmod in their Kicklayout. So for people who have both an A1222 and a X5000, and have disks formatted with SFS/02, I still this is a valid datapoint. If nothing else the seek/read values show just how useful caching is for SFS. DiskSpeed Test                A1222 SFS2   X5000 SFS2    Factor ===========================    ==========    ========== ====== Dir testing: file create         2251     6461    2.9x Dir testing: file open             8703    38417    4.4x Dir testing: directory scan      65833    581909    8.8x Dir testing: file delete           3709      14552    3.9x files/sec files/sec Dir testing: seek/read         3458  216069    62.5x seeks/sec seeks/sec 512B buffer: create file          8238592  20407168    2.5x 512B buffer: write to file       18582336   80512896    4.3x 512B buffer: read from file      22575168  165442432    7.3x 4K buffer: create file           30068224   58327552    1.9x 4K buffer: write to file         69452800  136417792    2.0x 4K buffer: read from file        63177728  93888000    1.5x 32K buffer: create file          48644096   87261184    1.8x 32K buffer: write to file       114806784    105844736    0.9x 32K buffer: read from file      111808512  479514624    4.3x 256K buffer: create file         69828608  114655232    1.6x 256k buffer: write to file      183468032  209616896    1.1x 256K buffer: read from file     162398208  184647680    1.1x bytes/sec bytes/sec Test #4: NGFS vs NGFS For the last test I again used the same partition size, same physical disk type (a Samsung 1TB SSD), latest NGFS from the beta servers. So this is a good straight-up Tabor vs Cyrus performance test. DiskSpeed Test                A1222 NGFS   X5000 NGFS Factor ===========================    ==========    ==========    ====== Dir testing: file create         1980   3466  1.75x Dir testing: file open            21458   50490   2.35x Dir testing: directory scan 45559   94721   2.08x Dir testing: file delete      3266   13348 4.09x files/sec files/sec Dir testing: seek/read      273772    570427 2.08x seeks/sec seeks/sec 512B buffer: create file         22317248 87755456 3.93x 512B buffer: write to file      55551744 114320256 2.06x 512B buffer: read from file    131891136 284454272 2.16x 4K buffer: create file          28107156 264827904 9.42x 4K buffer: write to file      240483328 155956736 0.65x 4K buffer: read from file      360107520 315719168 0.88x 32K buffer: create file          27578368 217362432 7.88x 32K buffer: write to file       227770368 142217216 0.62x 32K buffer: read from file      296333312 116805632 0.39x 256K buffer: create file        25395200  4227072 0.17x 256k buffer: write to file      291274752 311558144 1.07x 256K buffer: read from file     304021504 102694912 0.34x bytes/sec bytes/sec