110
Upcoming Hardware Launches 2019 (Updated)
(110 User Comments) »Introduction
In this article, which our team will regularly update, we will maintain a growing list of information pertaining to upcoming hardware releases based on leaks and official announcements as we spot them. There will obviously be a ton of rumors on unreleased hardware, and it is our goal to—based on our years of industry experience—exclude the crazy ones. In addition to these upcoming hardware release news, we will regularly adjust the structure of this article to better organize information. Each time an important change is made to this article, it will re-appear on our front page with a "new" banner, and the additions will be documented in the forum comments thread. This article will not leak information we signed an NDA for.Feel free to share your opinions and tips in the forum comments thread and subscribe to the same thread for updates.
Last Update (14th Jun):
- Updated AMD Zen 2 / Ryzen 3000 and AMD Picasso APUs
- Added Intel 9900KS, Intel Tiger Lake, Intel Sapphire Rapids, Intel Granite Lake
- Updated Intel Ice Lake
- Updated AMD Navi / Radeon RX 5700 Series and AMD Arcturus / RDNA 2
- Added NVIDIA SUPER
- Updated Intel Xe Graphics
- Added AMD X570, Intel X499 / X299G, Intel 400 and Intel 495 Chipsets
- Added TSMC 6 nanometer and Intel 7 nanometer processes
- Updated PCI-Express 5.0
- Added PCIe Gen 4.0 SSDs
- Added USB 4.0
- AMD Ryzen 50th Anniversary is launched now
- Removed older launched products: GeForce GTX 1650, GTX 1660
- Removed NVIDIA Volta and Ampere
Processors
AMD Ryzen 50th Anniversary [launched]
- The AMD Ryzen 7 2700X 50th Anniversary Edition has launched on Apr 29th.
- Processor is called "AMD Ryzen 7 2700X Gold Edition"
- Same specs as Ryzen 7 2700X
- Integrated Heatspreader (IHS) comes with a signature of AMD's CEO, Dr. Lisa Su
- Comes bundled with two games (Division 2 & World War Z), Lisa Su signature sticker and AMD T-Shirt
- Priced at $329
AMD Ryzen 3000 APUs (Picasso) [updated]
- Release Date: Q2 or Q3 2019 (probably around Computex)
- Based on the same Zen+ as Ryzen 2000 non-APU processors
- 12 nanometer production process at GlobalFoundries
- Ryzen 5 3400G: 4c/8t, 3.7 GHz base clock, 4.2 GHz max Boost, 6 MB Cache, 65 W, Vega 11 Graphics: 14000 MHz, 704 shaders
- Ryzen 3 3200G: 4c/4t, 3.6 GHz base clock, 4.0 GHz max Boost, 6 MB Cache, 65 W, Vega 8 Graphics: 1250 MHz, 512 shaders
- 3400G has soldered IHS
- DDR4 memory
- PCI-Express 3.0
AMD Zen 2 / Ryzen 3000 (Matisse) [updated]
- Release Date: Jul 7th, September for 3950X
- The client-segment Ryzen 3000 processors have been announced at Computex and E3
- Ryzen 9 3950X: 16c/32t, 3.5 GHz Base, 4.7 GHz Boost, 64 MB L3, 105 W TDP, $749
- Ryzen 9 3900X: 12c/24t, 3.8 GHz Base, 4.6 GHz Boost, 70 MB L3, 105 W TDP, $499
- Ryzen 7 3800X: 8c/16t, 3.9 GHz Base, 4.5 GHz Boost, 36 MB L3, 105 W TDP, $385
- Ryzen 7 3700X: 8c/16t, 3.6 GHz Base, 4.4 GHz Boost, 36 MB L3, 65 W TDP, $329
- Ryzen 5 3600X: 6c/12t, 3.8 GHz Base, 4.4 GHz Boost, 35 MB L3, 95 W TDP, $249
- Ryzen 5 3600: 6c/12t, 3.6 GHz Base, 4.2 GHz Boost, 35 MB L3, 65 W TDP, $199
- $750 Ryzen 9 3950X (16C, 32T) beats Intel's $2,000 i9-9980XE (18C, 36T)
- Renoir APUs expected for 2020
- 7 nm production process at TSMC
- 7 nm brings 2x density, 1/2 power (at same performance), or 1.25x performance (at same power)
- Claims better performance/watt than Intel 10 nm
- Floating point unit doubled to 256-bit
- AMD South Korea confirmed the existence of Ryzen 5 3600X and Ryzen 7 3700X (contest asking people to correctly guess the Cinebench scores of the two chips)
- EPYC SKUs come with up to 64 cores, using four to eight 7 nm CPU dies (each has 8 cores)
- The CPU dies are connected to a central IO die, which is still produced on a 12 nm process, probably GlobalFoundries 12LP
- L3 Cache per CCX doubled (16x 16 MB on 8-chiplet 64-core Rome CPU)
- Adds support for PCI-Express 4.0 (128 lanes total: 96-lanes on EPYC from the CPU + 32 from the Southbridge)
- PCI-Express 4.0 support only available on X570 chipset (older motherboards will fall back to PCIe Gen 3)
- EPYC parts use an 8-channel DDR4 memory interface
- Infinity Fabric can be clocked at memory clock or half memory clock (new)
- Major update to the microarchitecture, including IPC improvements
- New front end with improved branch-predictor, faster instruction prefetch, large L1 and L2 cache
- Codename: Matisse (CPU), Renoir (APU w/ IGP)
- Platform codename: Valhalla
- Hardening against Meltdown/Spectre (through architecture)
- Adds new instructions: Cache Line Write Back (CLWB), Read Processor ID (RDPID), and Write Back and Do Not Invalidate Cache (WBNOINVD)
- 13% IPC gain over Zen+, 16% over Zen 1
- Some applications are 29% faster than Zen 1
- Heatspreader (IHS) is soldered
- Continues to use socket AM4
- Tape-out: end of 2018
- Engineering Sample 2D3212BGMCWH2_37 / 34_N: 12c/24t, 3.4 GHz base, 3.6 GHz boost, 105 W TDP, 32 MB L3 cache
- Socket AM4 package a multi-chip module of 2-3 dies, an I/O controller die and 1 or 2 eight-core "Zen 2" CPU complex dies, each talking to the other over Infinity Fabric.
- AMD CES 2019 demo with 8-core/16-thread Ryzen 3000 AM4 prototype beats Core i9-9900K in multi-threaded Cinebench.
AMD Zen 3
- Release Date: 2020
- Codename: Vermeer (CPU), Dali (APU w/ IGP)
- Server platform codename "Milan"
- New process tech: 7 nm+
- Uses EUV (extreme ultraviolet) silicon fabrication node at TSMC
- 20% increase in transistor density, 10% lower power consumption
- Energy-efficiency prioritized, only minor improvements in IPC over Zen 2
- New CPU core
- Continues to use socket AM4
AMD Zen 4
- Release Date: Unknown
- "In Design" as of Nov 2018
AMD Threadripper 3rd Gen
- Release Date: 2019
- Codename: Castle Peak
- New process tech
- New CPU core
- Stays on socket TR4
Intel Core i9-9900KS [added]
- Release Date: H2 2019
- Based on 14 nm Coffee Lake refresh i9-9900K
- Base frequency 4 GHz
- All-core Boost: 5 GHz
- Improved Turbo algorithm for higher clocks at various workloads
Intel Lakefield Heterogenous processor
- Release Date: unknown
- Uses 10 nanometer production process
- A rendition of ARM big.LITTLE, but with Intel x86 cores
- One large performance core combined with four low-power cores, and power-gating added to the mix
- Big core is Sunny Cove based, small cores are Gracemont based
- Integrated Gen11 iGPU
- Fully integrated chipset and network interfaces
- Package designed for PoP (package over package) setups with DRAM and NAND flash chips over Foveros packaging
- Target devices include tablets and tablet+notebook convertibles
- Project Athena is an industry-wide effort involving a dozen companies to develop the next step in powerful mobile computing, rivaling the Ultrabook development a decade ago
Intel Comet Lake
- Release Date: 2019
- Introduces 10-core CPUs
- Uses 14 nm production process
- Socket LGA1151
- No significant architectural changes expected over "Coffee Lake"
- Adds BFLOAT16 deep learning boost instructions
- Same cache hierarchy, with 256 KB per core L2 cache, and 20 MB shared L3 cache
- Developed to counter AMD's Zen 2
Intel Cannon Lake
- Release Date: Additional processors delayed to 2019
- One mobile SKU launched in May: Core i3-8121U 2.2 GHz, no integrated graphics
- Core M3 8114Y: 1.5 GHz base, 2.2 GHz boost, 4.5 W TDP, Intel UHD iGPU
- 10 nanometer production process
- DDR4L support
- Intel is reportedly having difficulties ramping up 10 nm, which could lead to delays
- Adds AVX512 instructions (so far available only on HEDT platform, since Skylake-X). New instructions: AVX512F, AVX512CD, AVX512DQ, AVX512BW, and AVX512VL. New commands: AVX512_IFMA and AVX512_VBMI
Intel Cascade Lake
- Released: April 3, 2019
- Server/Enterprise version of Whiskey Lake
- Same microarchitecture as Coffee Lake
- Still on 14 nm, using a slightly improved process
- 56 cores spread across two dies, using multi-chip-module
- FC-BGA package (not socketed), 5,903-pins
- Adds support for Optane Persistent Memory
- 12 memory channels per CPU (six per die)
- 88 PCI-Express lanes
- Mitigations for recent Intel vulnerabilities
- Supports AVX-512 and AES-NI
- Deep Learning Boost (Variable Length Neural Network Instructions)
- Intel Speed Select to select specific multipliers for CPU cores on running system, without reboot
Intel Cascade Lake-X
- Release Date: Fall 2019
- Same feature set as Cascade Lake but using LGA2066 platform
Intel Cooper Lake
- Release Date: 2019
- Refresh of Cascade Lake
- 14 nanometer production process
- Uses same socket and platform
- Deep Learning Boost gets additional instructions (compared to Cascade Lake): BFLOAT16
- Higher clock speeds
Intel Sapphire Rapids [added]
- Release Date: 2021
- Successor to Cooper Lake
- 8-channel DDR5
- For enterprise / data center
- PCIe 5.0
- Probably 7 nm process
- Platform name: Eagle Stream
Intel Granite Rapids [added]
- Release Date: 2022
- Successor to Sapphire Rapids
- 8-channel DDR5
- For enterprise / data center
- PCIe 5.0
- Probably 7 nm+ process
- Platform name: Eagle Stream
Intel Ice Lake [updated]
- Release Date: Late 2019, desktop 2020
- Uses 10 nanometer DUV (deep-ultraviolet) process
- Will use "Intel 10th gen" branding, with Core i3, i5 and i7
- First processors are mobile quad-core designs
- Ultra-low-power (ULP) platform was displayed at Computex: Using a multi-chip-module design, with TDP between 8 and 15 W
- Mass production started in Q3 2019
- Uses a brand-new CPU core design codenamed "Sunny Cove"
- Adds AVX512 instructions (so far available only on HEDT platform, since Skylake-X). New instructions: AVX512F, AVX512CD, AVX512DQ, AVX512BW, and AVX512VL. New commands: AVX512_IFMA and AVX512_VBMI
- 20-30 broadening of various number crunching resources, wider execution window, more AGUs
- SHA-NI and Vector-AES instruction sets, up to 75% higher encryption performance vs. "Skylake"
- Supports unganged memory mode
- First implementation will not be for a desktop processor, but a low-power SOC for notebooks (Ice Lake-U)
- Ice Lake-U is 4-core/8-thread based on Sunny Cove w/ Gen 11 GT2 graphics with 64 EUs.
- Ice Lake SP are Xeons, which will launch in 2020.
- Integrated GPU based on new Gen11 architecture, up to 1 TFLOP/s ALU compute performance
- Integrated GPU supports DisplayPort 1.4a and DSC for 5K and 8K monitor support
- Gen11 also features tile-based rendering, one of NVIDIA's secret-sauce features
- Integrated GPU supports VESA adaptive V-sync, all AMD FreeSync-capable monitors should work with this
Intel Willow Cove and Golden Cove Cores
- Release Date: 2020 and 2021
- Succeeds "Sunny Cove"
- Willow Cove improves on-die caches, adds more security features, and takes advantage of 10 nm+ process improvements to increase clock speeds versus Sunny Cove
- Golden Cove will add significant single-thread (IPC) increases over Sunny Cove, add on-die matrix multiplication hardware, improved 5G network-stack HSP performance, and more security features than Willow Cove
Intel Tiger Lake [added]
- Release Date: 2020
- Mobile chip, succeeds "Lakefield"
- CPU cores based on Willow Cove
- Integrated GPU based on Intel Xe architecture (not Gen 11)
- Uses 10 nm+ process
- "Next-gen I/O"
New 28-core HEDT platform derived from LGA3467
- Released as Xeon W-3175X, priced at $3,000, 255 W TDP, 3.1 GHz Base, 4.1 GHz Boost
- Core i9-9990XE: OEM-only, $2,300, 14c/28t, 4 GHz Base, 5.1 GHz Boost, 19.25 MB Cache, 255 W TDP
- Core i9-9980XE: 18c/36t, 3 GHz Base, 4.5 GHz Boost, 24.75 MB Cache, 165 W TDP
- Core i9-9940X: 14c/28t, 3.3 GHz Base, 4.4 GHz Boost, 19.25 MB Cache, 165 W TDP
- Client-segment implementation of the Skylake XCC die
- No STIM
- Up to 28 cores: 16-core, 24-core, 26-core, and 28-core SKUs possible
- HyperThreading available
- Six-channel DDR4 Memory interface, Up to 768 GB non-ECC memory
- Uses new X599 chipset, new motherboards required
- 44 PCIe gen 3.0 lanes
- Two AVX-512 FMA Units
- 14 nm++ process
- Confirmed Motherboard: ASUS ROG Dominus Extreme
Graphics / GPUs
NVIDIA SUPER [added]
- Release Date: July 2019
- Based on existing Turing chips with more shaders enabled and faster/more GDDR6 memory
- RTX 2080 Ti SUPER: Unlocked chip with higher power delivery capability
- RTX 2080 SUPER: TU102, 3072 cores, 8 GB
- RTX 2070 SUPER: TU104, 2560 cores, 8 GB
- RTX 2060 SUPER: TU106, 2176 cores, 8 GB
- RTX 2070 Ti SUPER could be part of this lineup, too
- Memory for all these SKUs will run at 16 Gbps (up from 14 Gbps on current RTX cards)
- Current RTX cards will drop in price, and the SUPER cards will take up their current price points
NVIDIA GeForce GTX 1650 Ti
- Release Date: Probably 2nd half of 2019
- Price: between $170 and $200
- Based on 12 nm Turing TU117 silicon with all shaders unlocked
- 1024 CUDA cores, 32 ROPs, 64 TMUs
- 4 GB GDDR5 128-bit memory at 8000 MHz (same as GTX 1650)
- Lacks Raytracing and DLSS
- No reference design from NVIDIA
NVIDIA GeForce RTX 2070 Ti
- Gigabyte (where "2070 Ti" originated from), now claims that this was just a typo
- Release Date: Unknown
- Based on Turing TU104 (same as RTX 2080)
- Faster than GTX 1070 Ti and GTX 1080
- Pricing $600-650 (ie, half of RTX 2080 Ti)
NVIDIA GeForce RTX 2050
- Release Date: 2019
- Based on new, smaller, chip
- Looks like this was scrapped in favor of GTX 1660 and GTX 1650
AMD Radeon RX 5700 XT / RX 5700 (Navi)
- Release Date: July 7th, 2019
- Successor for "Polaris" (not a super high-end SKU)
- Uses RDNA architecture (improved GCN)
- TSMC, 7 nm
- 2.3x gain in performance per area
- 1.5x gain in performance per watt
- Chip codenames: Navi 16, Navi 12, Navi 10, and Navi 9
- Navi 10: 251 mm²
- No raytracing support in RDNA 1st generation at launch (DXR emulation through shaders will be added through driver updates)
- 2nd generation RDNA (Arcturus) will have hardware raytracing acceleration (2020 - 2021)
- Radeon RX 5700 XT: $449, 2560 shaders, up to 1905 MHz Boost, 1755 MHz Game Clock, 1605 MHz Base Clock, 8 GB GDDR6 256-bit, 14 Gbps, 448 GB/s, 64 ROPs, 160 TMUs, 225 W TDP, Performance slightly faster than RTX 2070
- Radeon RX 5700 XT 50th Anniversary Edition: $499, black and gold color scheme, 1980 MHz Boost, 1830 MHz Game, 1680 MHz Base. Other specs same as RX 5700 XT, 235 W TDP
- Radeon RX 5700: $379, 2304 shaders, up to 1725 MHz Boost, 1625 MHz Game Clock, 1465 MHz Base Clock, 8 GB GDDR6, 256-bit, 14 Gbps, 448 GB/s, 64 ROPs, 144 TMUs, 180 W TDP, faster than RTX 2060
- PCIe 4.0 support
- DisplayPort 1.4 HDR / 8K60 and 4K120 with a single cable
- DSC 1.2a Stream Compression
- HDMI remains at 2.0
- VP9 decode up to 4K90 or 8K24
- H.265 HEVC encode up to 4K60, decode up to 1080p360, 4K90 or 8K24
- H.264 MPEG4 encoding up to 4K90 and 360 FPS at 1080p, decode up to 4K150 and 1080p600
- Dual BIOS
- Device ID: 0x73101002.
AMD MI-NEXT
- Release Date: Unknown
- Successor to Vega 20
AMD Arcturus / RDNA 2 [updated]
- Release Date: 2020
- TSMC, 7 nm+
- Adds hardware raytracing acceleration
Intel XE Discrete Graphics [updated]
- First Intel Discrete GPU since ill-fated Larrabee
- New architecture built from the ground up, and not an upscale of Gen11
- Could be produced at Samsung to leverage their 10 nm tech, while Intel ramps up its own
- Raytracing hardware acceleration support will definitely be included on the data-center GPUs (and probably on the consumer models, too)
- Double-digit TFLOP/s scaling all the way up to 0.1+ PFLOP/s
- Will be used in upcoming Cray Aurora Supercomputer for Argonne National Laboratory in 2021
- Targeting a wide segment of markets, including consumer (client-segment) graphics, enthusiast-segment, and data-center compute
- Uses new graphics control panel that's being introduced during 2019
Intel Discrete GPU / Arctic Sound
- Release Date: 2020
- Intel will hold a world tour in 2019, to build enthusiasm for the new architecture
- Advanced management for power and clocks
- Test chip: 8x8 mm² die area, 1.54B transistors, 14 nm, 50-400 MHz clock, EUs at 2x clock if needed
- Raja Koduri who left AMD in late 2017 is somehow involved
- Confirmed to support VESA Adaptive Sync
Intel Jupiter Sound
- Release Date: 2022
- Discrete GPU
- Produced on 10 nanometer production process
- Successor to Arctic Sound
Chipsets
Intel X399
- Release Date: unknown
- Supports Coffee Lake-X and Cannon Lake-X
Intel X499 / X299G [added]
- Release Date: Late 2019
- Will be used with Intel's new Core X HEDT processors, probably Cascade Lake X
- Apparently renamed from X499 to X299G — At Computex Gigabyte showed motherboards with "X499" print covered by "X299G" stickers
Intel X599
- Release Date: Unknown
- Supports new Intel 28-core HEDT platform
- New motherboards, with new socket: LGA3647
- Based on C629 Server Chipset
- 6-channel memory
- 48 PCIe lanes
Intel Z399
- Release Date: Unknown
- New HEDT chipset
- Sibling chipset to X599
- Socket LGA2066
- For 20- and 22-core CPUs
- Works with existing Skylake-X LCC and HCC chips
Intel 400 Series Chipsets [added]
- Release Date: Unknown
- Chipset lineup for use with 10-core Comet Lake processors
Intel 495 Series Chipsets [added]
- Release Date: Unknown
- Chipset lineup for use with Ice Lake processors
AMD X499 Chipset
- Release Date: Q1 2019
- Supports Threadripper
- Unknown changes
AMD X570 Chipset [updated]
- Release Date: mid-2019, at the same time as Zen 2 (Ryzen 3000)
- Socket AM4
- Backwards compatible with 2nd gen Ryzen processors (1st gen Ryzen "Raven Ridge" and "Summit Ridge" not supported)
- AMD's own design — does not use ASMedia IP (to avoid security vulnerabilities)
- Supports PCI-Express 4.0
- X570 motherboards expected to be significantly more expensive than previous generations
- Up to 12 SATA ports + 2 by AM4 SoC, up to two M.2 NVMe slots with PCIe Gen 4 x4 + 1 slot by AM4 SoC
- Up to eight 10 Gbps USB 3.1 Gen 2 by the chipset
- 44 PCIe lanes total platform config (24 by CPU, 16 by X570 chipset)
- 15 Watt TDP — most motherboards announced so far come with a small fan on the chipset
- Desktop platform codename: "Valhalla"
- Compatible with earlier Ryzen processors
AMD Z490 Chipset
- Unknown release date
- Looks to be cancelled, in favor of Zen 2 / X570
- Increases number of downstream PCIe lanes by adding four PCIe Gen 3 lanes on top of the existing eight PCIe Gen 2 lanes
- Other features identical to X470
Memory
DDR5 System Memory
- Release Date: Late 2019/2020
- JEDEC standard not fully complete yet, expected for summer 2018
- Demo'd in May 2018 by Micron: DDR5-4400
- Samsung 16 Gb DDR5 DRAM developed since February 2018
- Samsung has completed functional testing and validation of a LPDDR5 prototype: 10 nm class, 8 Gbit, final clocks: DDR5-5500 and DDR5-6400
- SK Hynix has 16 Gb DDR5-5200 samples ready, 1.1 V, mass production expected 2020
- 4800 - 6400 Mbps
- Expected to be produced using 7 nm technologies
- 64-bit link at 1.1 V
- Voltage regulators on the DIMM modules
HBM2e Graphics Memory
- Release Date: 2019
- Offers 3.2 Gbps per pin (33% faster than HBM2)
- Samsung Flashbolt: 16 Gb per die, 8-layers stacked, 16 GB per chip with 410 GB/s bandwidth
HBM3 Graphics Memory
- Release Date: Not before 2019
- Double the memory bandwidth per stack (4000 Gbps expected)
- Expected to be produced using 7 nm technologies
Silicon Fabrication Tech
TSMC 5 nanometer
- Release Date: Unknown
- Risk production as of Q2 2019
- Uses TSMC's second implementation of EUV (Extreme Ultra Violet)
- Up to 1.8x the density of 7 nm
- +15% higher clocks
Samsung 5 nanometer
- Release Date: Unknown
- Ready for customer sample production as of Q2 2019
- Uses EUV (Extreme Ultra Violet)
- Up to 25% the density of 7 nm
- 20% lower power consumption
- 10% higher performance
TSMC 6 nanometer [added]
- Release Date: Unknown
- Backwards compatible with 7 nm process — no new design tools needed
- Uses EUV (Extreme Ultra Violet), up to four EUV layers
Samsung 6 nanometer
- Release Date: Unknown
- First product taped out as of Q2 2019
- Uses EUV (Extreme Ultra Violet)
- Special variant for customers
Intel 7 nanometer [added]
- Release Date: Planned for 2021
- Succeeded by 7 nm+ node in 2022, and 7 nm++ in 2023
- Uses EUV (Extreme Ultra Violet)
- 4x reduction in design rules
- Planned to be used on multiple products: CPU, GPU, AI, FPGA, 5G networking
Other
PCIe Gen 4.0 SSDs [added]
- Release Date: H2 2019
- Up to 5 GB/s transfer rates
- Double the maximum bandwidth compared to PCIe Gen 3
- Controllers available from Silicon Motion (SM2267), and Phison (E16)
- Products on display at Computex by Gigabyte (Phison), ADATA (SM2267), Corsair, Galax (Phison E16)
Hynix 4D NAND
- Release Date: H1 2019
- Developed by SK Hynix
- Sampling in Q4 2018
- Reduces chip physical size, while increasing capacity at the same time
- Supports TLC and QLC
- 30% higher write and 25% higher read performance
- 1.2 V
- 1st generation: 96 stacks, 1.2 Gbps per pin, 512 Gbit TLC
- 128 stacks in development, scales up to 512 stacks
Toshiba XL-Flash
- Developed by Toshiba
- Uses existing SLC flash technology to improve latencies
- 1/10th the read latency of TLC
- Good for random IOPS and better QoS at shallow queue depth
- Can combine SLC and TLC/QLC for tiered, cost-optimized storage
- Intel Optane memory competitor
Toshiba 128-layer 3D NAND Flash
- Release Date: 2020 or 2021
- Developed by Toshiba and partner Western Digital
- Called BiCS 5
- Cells are TLC (not QLC)
- Chip density: 512 Gb
- Write performance per channel doubled to 133 MB/s
Intel CXL Interconnect
- New interconnect for high-bandwidth devices like GPUs
- Targeted at enterprise and servers
- Competitor to NVLink, Infinity Fabric and PCI-Express
- Uses PCIe physical layer
- Link layer designed for low latency
- 32 Gbps per lane, per direction (Like PCIe Gen 5.0)
PCI-Express 4.0
- Specification released in late 2017
- 16 GT/s bandwidth per lane, per direction (2x the bandwidth of PCIe 3.0)
- Reduced latency
- Lane margining
- I/O virtualization capabilities
- Will be supported by AMD Zen 2 and Vega 20 (Radeon Instinct MI60 & MI50)
PCI-Express 5.0 [updated]
- Specification released end of May 2019
- Products not expected before 2020
- 32 GT/s bandwidth per lane, per direction (4x the bandwidth of PCIe 3.0)
- 128/130 bit encoding (= 1.5% overhead)
- Implements electrical changes to improve signal integrity and mechanical performance of connectors
- Physical connector is backward compatible
- Signalling is backward compatible with all previous PCIe versions
USB 4.0 [added]
- Release Date: End of 2020
- Spec expected to be finalized in mid-2019
- Conceptually similar to Thunderbolt 3
- Up to 40 Gbps speeds
- Multiple data and display protocols can share bandwidth
- Uses USB Type-C connector
- Backwards compatible with USB 2.0, 3.0 and Thunderbolt 3