Visual NAND Reconstructor 4.1
Dear Users,
It is a beginning of new year and we’re happy to deliver the new update of VNR!
We have added several powerful features for ECC correction and reread on GPU as well as some key changes in software, including the new plugin!
Here’s the full list of new features and resources:
Multi-GPU ECC correction
The average capacity of flash chips is growing and we continue to transition time consuming functions and algorithms to GPU in order to speed up the whole process of data recovery. Multiple GPUs can be used now for parallel computation of ECC code. Different graphic cards from different brands can be combined now into one computing unit. During our tests we noticed almost linear increase of speed. No SLI or Crossfire is required, all graphic cards just have to be connected to one motherboard and selected in dialog window. The function is available from ReRead dump menu. The diagrams and timetables below show results obtained from combinations of different GPUs.
Tested devices/codes
Integrated graphic card plus dedicated GPU
Two different vendor graphic cards
Setup of multiple identical cards
ECC ReRead on GPU
New option to perform ECC ReRead of memory chip using GPU. Speed increase is very similar to GPU ECC correction compared to CPU. Reference speed comparison for ECC correction is in VNR 4.0 update news
The buffer size for ReRead is set by number of pages. It has to be found experimentally depending on your PC config, we recommend to start from block size of chip.
Pattern Writer plugin
We got many requests from our users to add special pattern writer plugin into VNR so we developed one. It helps to prepare donor devices for further research of ECC, XOR, translation, etc. Flexible options allow to write static hex value, or counter.
Other changes
-
Bad Column Remover element now works on multiple blocks. Select preset, scroll down and the blocks to calculate statistics from, press add button and read the values.
-
Markers table – Page mode
There are many controllers nowadays that require fine adjustment of block translation and sometimes page translation. Specifically for page translation we added “Page mode” to Markers Table element that is working much faster than in previous versions. For such cases block size has to be equal to page size using context menu in structure viewer
-
Read Retry decision making algorithm was improved, especially for Hynix chips
-
Many minor changes and bugs have been fixed
New NAND chips
-
ECDE98CE74 K9ACGD8U0M
-
ECD798CE74 K9ABGD8U0B
-
B5A46432AA Noname
-
ADDA909544 H27U2G8F2C
-
ADCA90D544 H27U2G6F2C
-
ADBA905546 H27S2G6F2DKA-BM
-
ADBA905544 H27S2G6F2C
-
ADAA901544 H27S2G8F2C
-
AD5E282210 Noname
-
98CC906676 XT61M4G6D2TA-B8BE
-
45D5888272 mSD
-
453EA7827E Sandisk 64GB
-
2C6801A6A9 MT29F64G08AKCCB
-
2079207920 NAND01GW3A2BN6
-
2076207620 NAND512W3A2BN6
-
2075207520 NAND256W3A2BN6
-
2073207320 NAND128W3A2BN6
-
2059205920 NAND01GW4A2BN6
-
2056205620 NAND512W4A2BN6
-
2055205520 NAND256W4A2BN6
-
2053205320 NAND128W4A2BN6
-
2049204920 NAND01GR4A2BN6
-
2046204620 NAND512R4A2BN6
-
2045204520 NAND256R4A2BN6
-
2043204320 NAND128R4A2BN6
-
2039203920 NAND01GR3A2BN6
-
2036203620 NAND512R3A2BN6
-
2035203520 NAND256R3A2BN6
-
2033203320 NAND128R3A2BN6
New scramblers (XOR keys)
-
AllWinner_A23(17664b_128p_xoredSA_ECC70b)_CF7ED4.xor
-
AU_noname(16k_512p)_988EE1.xor
-
IS916EN(9216b_258p_xoredSA)_20C3A7.xor
-
IS917(16k_258p)_DF3C58.xor
-
IS917(8832b_256p_xoredSA)_3C5811.xor
-
MW8690(8832b_256p_xoredECC80b)_988EE1.xor
-
PS2251-08-Q(16k_256p)_FA9E80.xor
-
SM3255QAB(8k_192p)_FF09FD.xor
New ECC/BCH
-
AllWinnerA23_17664b_70ecc__16.bch
-
AU_blob_17760(ecc86b)_16_X.bch
-
AU_blob_18592(ecc124b)_16_X.bch
-
AU_monolith_18432(ecc84)_16.bch
-
CBM2099E_18432(ecc100b)_16.bch
-
CBM2099_9216(ecc100b)_8.bch
-
CBM2199_18432(ecc124b)_16.bch
-
DM8261_9216(ecc54b)_16.bch
-
IS917_8783_9(ecc70b+SAecc20b).bch
-
MW8690_8832(80ecc)_8.bch
-
SM2234H_4320(ECC_25)_8.bch
-
SM3271AA_17664(ecc74b)_16.bch
-
SM_Transcend_mSD_8625(ecc51b+SAecc20b).bch
New monolithic chip pinouts
