Capacity: | 2 TB (2000 GB) |
---|---|
Variants: | 250 GB 500 GB 1 TB 2 TB |
Overprovisioning: | 185.4 GB / 10.0 % |
Production: | Active |
Released: | Sep 22nd, 2020 |
Price at Launch: | 430 USD |
Part Number: | MZ-V8P2T0 |
Market: | Consumer |
Form Factor: | M.2 2280 (Single-Sided) |
---|---|
Interface: | PCIe 4.0 x4 |
Protocol: | NVMe 1.3 |
Power Draw: |
0.04 W (Idle) 6.1 W (Avg) 7.2 W (Max) |
Manufacturer: | Samsung |
---|---|
Name: | Elpis (S4LV003) |
Architecture: | ARM 32-bit Cortex-R8 |
Core Count: | 5-Core |
Foundry: | Samsung |
Process: | 8 nm |
Flash Channels: | 8 @ 1,400 MT/s |
Chip Enables: | 8 |
Controller Features: | DRAM (enabled) |
Manufacturer: | Samsung |
---|---|
Name: | V-NAND V6 |
Type: | TLC |
Technology: | 128-layer |
Speed: | 1200 MT/s |
Capacity: | 2 chips @ 8 Tbit |
Toggle: | 4.0 |
Topology: | Charge Trap |
Die Size: | 102 mm² (5.0 Gbit/mm²) |
Dies per Chip: | 16 dies @ 512 Gbit |
Planes per Die: | 2 |
Decks per Die: | 1 |
Word Lines: |
136 per NAND String
94.1% Vertical Efficiency |
Read Time (tR): | 45 µs |
Program Time (tProg): | 390 µs |
Block Erase Time (tBERS): | 3.5 ms |
Die Read Speed: | 711 MB/s |
Die Write Speed: | 82 MB/s |
Endurance: (up to) |
3000 P/E Cycles |
Page Size: | 16 KB |
Type: | LPDDR4-1866 |
---|---|
Name: | SAMSUNG |
Capacity: |
2048 MB
(1x 2048 MB) |
Organization: | 16Gx32 |
Sequential Read: | 7,000 MB/s |
---|---|
Sequential Write: | 5,100 MB/s |
Random Read: | 1,000,000 IOPS |
Random Write: | 1,000,000 IOPS |
Endurance: | 1200 TBW |
Warranty: | 5 Years |
MTBF: | 1.5 Million Hours |
Drive Writes Per Day (DWPD): | 0.3 |
SLC Write Cache: |
approx. 236 GB
(230 GB Dynamic + 6 GB Static) |
Speed when Cache Exhausted: | approx. 2000 MB/s |
Cache Folding Speed: | 1500 MB/s |
TRIM: | Yes |
---|---|
SMART: | Yes |
Power Loss Protection: | No |
Encryption: |
|
RGB Lighting: | No |
PS5 Compatible: | Yes |
Drive:Nand bus speed: 800 MT/s up to 1400 MT/s Controller:NAND Flash Bus not confirmed, could be higher or slower. NAND Die:A Dual-plane Die with 2 sub-planes with 8 KiB pages in order to improve performance through paralellism. |